Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
The death of the Bayer filter? Maybe not.
41 views
Skip to first unread message
Rich
Apr 3, 2012, 11:29:41 PM4/3/12
to
I don't hate the Bayer. I think it's a clever way to deal with a problem
that some find objectionable. But its days could be numbered.
Don't know if resolution is improved yet, but it's nice to know that it
only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
$500 DSLRs that meet the spec.
http://www.dpreview.com/articles/6126592906/first-impressions-using-the-
fujifilm-x-pro1/6
that some find objectionable. But its days could be numbered.
Don't know if resolution is improved yet, but it's nice to know that it
only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
$500 DSLRs that meet the spec.
http://www.dpreview.com/articles/6126592906/first-impressions-using-the-
fujifilm-x-pro1/6
Mxsmanic
Apr 4, 2012, 1:30:39 AM4/4/12
to
Rich writes:
> I don't hate the Bayer. I think it's a clever way to deal with a problem
> that some find objectionable. But its days could be numbered.
> Don't know if resolution is improved yet, but it's nice to know that it
> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> $500 DSLRs that meet the spec.
A Bayer filter and an anti-aliasing filter are two different things.
> I don't hate the Bayer. I think it's a clever way to deal with a problem
> that some find objectionable. But its days could be numbered.
> Don't know if resolution is improved yet, but it's nice to know that it
> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> $500 DSLRs that meet the spec.
Me
Apr 4, 2012, 2:04:57 AM4/4/12
to
theoretically gets rid of *colour moire, and possibly reduces luminance
moire (by more random placement of green sensels than a repeating bayer
pattern), so there's less need for a low pass filter.
I guess that demosaicking raw files is rather complicated, the only raw
converter supporting the camera for now is Silkypix - which I've tried
in the past and I thought it really sucked, but YMMV.
It's a good idea, but with higher and higher resolution with
conventional bayer filters, when at those higher resolutions the low
pass filter needs to be less aggressive (on a whole image basis) to
produce the same anti-aliasing effect (on a per pixel basis as a lower
resolution sensor), any "need" for such an innovative solution may have
already passed.
*It can't work perfectly, only reduce the incidence perhaps in repeating
patterns - as any point of white light hitting the sensor at a size less
than one sensel, is going to be recorded as R, G, or B - not white.
Bruce
Apr 4, 2012, 5:14:30 AM4/4/12
to
made metal bodied camera. This is exactly what you have been asking
for yet, true to type, you cannot resist dissing it!
One of the reasons that you can buy cheap DSLRs with something like
comparable performance is because they are made of plastic. Perhaps
now you can see why?
As for the death of the Bayer pattern, it isn't. This is no Foveon
and there is still a pattern. It may not be the exact Bayer pattern
but the principle is very much the same. However, what we are seeing
is the slow death of the AA filter, which is very welcome.
The X-Pro1 does not seem to be afflicted with the multiple teething
problems of its X100 and X10 companions. However, it seems to have
sloooow AF. I haven't had the chance to give one an extended trial so
I will reserve a final judgement until I do, but based on the reviews
I have read the AF speed could well be the X-Pro1's Achilles' heel.
I am not attracted by the lens selection. I would like to see
equivalents of my favourite rangefinder lenses, which would mean a
16mm f/2.8 and a 24mm f/1.4 in addition to the 60mm f/2.4 that
Fujifilm got about right. Instead, there is an 18mm f/2 (???) and a
23mm f/2, with presumably the same optics as the X100 lens.
There will be two zooms, a 12-20mm f/4 and a 70-200mm f/4, both image
stabilised. I don't think anyone in their right mind would buy a
70-200mm lens for a 'rangefinder' camera with an APS-C sensor.
So I am not convinced by the X-Pro1. The Leica M10 will probably cost
four times as much, but it will be the complete digital rangefinder
camera with lenses in focal lengths that people want. There is a
little over a month to go before the M10 is announced, possibly
accompanied by Leica's take on a mirrorless system camera.
Mxsmanic
Apr 4, 2012, 7:56:56 AM4/4/12
to
Me writes:
> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> theoretically gets rid of *colour moire, and possibly reduces luminance
> moire (by more random placement of green sensels than a repeating bayer
> pattern), so there's less need for a low pass filter.
It would also reduce color resolution. If it's actually being done, it sounds
> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> theoretically gets rid of *colour moire, and possibly reduces luminance
> moire (by more random placement of green sensels than a repeating bayer
> pattern), so there's less need for a low pass filter.
like a gimmick.
The whole issue of aliasing is past history; I don't know why people still
worry about it. When you have a large number of pixels, you don't need to
worry about aliasing. So the ultimate solution for aliasing is more pixels.
There will always be aliasing of details that are sufficiently small, if the
lens can resolve them, but the aliasing will not be noticeable or troublesome.
Even in the old days, when sensors had far fewer pixels, I never really
noticed problems with aliasing. I'd rather take the risk of aliasing than put
up with the blur of anti-aliasing.
Andrew Haley
Apr 4, 2012, 9:31:58 AM4/4/12
to
Me <us...@example.net> wrote:
> On 4/04/2012 5:30 p.m., Mxsmanic wrote:
>> Rich writes:
>>
>>> I don't hate the Bayer. I think it's a clever way to deal with a problem
>>> that some find objectionable. But its days could be numbered.
>>> Don't know if resolution is improved yet, but it's nice to know that it
>>> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
>>> $500 DSLRs that meet the spec.
>>
>> A Bayer filter and an anti-aliasing filter are two different things.
> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> theoretically gets rid of *colour moire, and possibly reduces luminance
> moire (by more random placement of green sensels than a repeating bayer
> pattern), so there's less need for a low pass filter.
It's nothing like pseudo-random.
> On 4/04/2012 5:30 p.m., Mxsmanic wrote:
>> Rich writes:
>>
>>> I don't hate the Bayer. I think it's a clever way to deal with a problem
>>> that some find objectionable. But its days could be numbered.
>>> Don't know if resolution is improved yet, but it's nice to know that it
>>> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
>>> $500 DSLRs that meet the spec.
>>
>> A Bayer filter and an anti-aliasing filter are two different things.
> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> theoretically gets rid of *colour moire, and possibly reduces luminance
> moire (by more random placement of green sensels than a repeating bayer
> pattern), so there's less need for a low pass filter.
http://www.dpreview.com/previews/fujifilmxpro1/3 shows that it's a 6x6
array instead of a 2x2 array. This 6x6 array is rotationally
symmetrical, and it repeats across the sensor. It has the interesting
property that there are fewer blue and red pixels than a Bayer sensor
has: 8R:20G:8B instead of 9:18:9. It probably won't make much
difference.
Andrew.
RichA
Apr 4, 2012, 10:40:47 AM4/4/12
to
converters and it actually produced sharper images than some, like
Capture One. This goes back a few years. But, its interface, at
least with Pentax bundles, was terrible.
RichA
Apr 4, 2012, 10:57:18 AM4/4/12
to
On Apr 4, 5:14 am, Bruce <docnews2...@gmail.com> wrote:
for one with 2 lenses. I'd rather get a traditional DSLR, like the
upcoming D400 with the new 16-85mm f4.0 zoom.
> One of the reasons that you can buy cheap DSLRs with something like
> comparable performance is because they are made of plastic. Perhaps
> now you can see why?
I was referring more to function. See: AF quality and speed.
> As for the death of the Bayer pattern, it isn't. This is no Foveon
> and there is still a pattern. It may not be the exact Bayer pattern
> but the principle is very much the same. However, what we are seeing
> is the slow death of the AA filter, which is very welcome.
True. I had a Fuji S5 so I know about "different" sensors and what
they produce. The jury is still out, but this one looks pretty good.
> The X-Pro1 does not seem to be afflicted with the multiple teething
> problems of its X100 and X10 companions. However, it seems to have
> sloooow AF. I haven't had the chance to give one an extended trial so
> I will reserve a final judgement until I do, but based on the reviews
> I have read the AF speed could well be the X-Pro1's Achilles' heel.
Same as Pentax.
> I am not attracted by the lens selection. I would like to see
> equivalents of my favourite rangefinder lenses, which would mean a
> 16mm f/2.8 and a 24mm f/1.4 in addition to the 60mm f/2.4 that
> Fujifilm got about right. Instead, there is an 18mm f/2 (???) and a
> 23mm f/2, with presumably the same optics as the X100 lens.
>
> There will be two zooms, a 12-20mm f/4 and a 70-200mm f/4, both image
> stabilised. I don't think anyone in their right mind would buy a
> 70-200mm lens for a 'rangefinder' camera with an APS-C sensor.
>
> So I am not convinced by the X-Pro1. The Leica M10 will probably cost
> four times as much, but it will be the complete digital rangefinder
> camera with lenses in focal lengths that people want. There is a
> little over a month to go before the M10 is announced, possibly
> accompanied by Leica's take on a mirrorless system camera.
I thought the practical limit for rangefinder focusing was 135mm for
35mm? Personally, I like how they are breaking from tradition with
different focal lengths than what we are used to getting from Canon
and Nikon. They aren't the first to do that. A 12-20mm f4.0 is
interesting, not quite 2x and not that fast, so maybe it's to keep the
size reasonable? Also, it will be sharper across the board than an
f2.8., all things being equal. As for the 70-200mm, it probably makes
more sense here with a reasonable sized body than on some of the other
more diminutive mirrorless offerings from other companies.
> Rich <n...@nowhere.com> wrote:
> >I don't hate the Bayer. I think it's a clever way to deal with a problem
> >that some find objectionable. But its days could be numbered.
> >Don't know if resolution is improved yet, but it's nice to know that it
> >only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> >$500 DSLRs that meet the spec.
>
> <http://www.dpreview.com/articles/6126592906/first-impressions-using-t...>
> >I don't hate the Bayer. I think it's a clever way to deal with a problem
> >that some find objectionable. But its days could be numbered.
> >Don't know if resolution is improved yet, but it's nice to know that it
> >only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> >$500 DSLRs that meet the spec.
>
>
> You should be celebrating the X-Pro1 because it is a traditionally
> made metal bodied camera. This is exactly what you have been asking
> for yet, true to type, you cannot resist dissing it!
Are you kidding? I'd take one in a minute. But maybe not at $3000+
> You should be celebrating the X-Pro1 because it is a traditionally
> made metal bodied camera. This is exactly what you have been asking
> for yet, true to type, you cannot resist dissing it!
for one with 2 lenses. I'd rather get a traditional DSLR, like the
upcoming D400 with the new 16-85mm f4.0 zoom.
> One of the reasons that you can buy cheap DSLRs with something like
> comparable performance is because they are made of plastic. Perhaps
> now you can see why?
> As for the death of the Bayer pattern, it isn't. This is no Foveon
> and there is still a pattern. It may not be the exact Bayer pattern
> but the principle is very much the same. However, what we are seeing
> is the slow death of the AA filter, which is very welcome.
they produce. The jury is still out, but this one looks pretty good.
> The X-Pro1 does not seem to be afflicted with the multiple teething
> problems of its X100 and X10 companions. However, it seems to have
> sloooow AF. I haven't had the chance to give one an extended trial so
> I will reserve a final judgement until I do, but based on the reviews
> I have read the AF speed could well be the X-Pro1's Achilles' heel.
> I am not attracted by the lens selection. I would like to see
> equivalents of my favourite rangefinder lenses, which would mean a
> 16mm f/2.8 and a 24mm f/1.4 in addition to the 60mm f/2.4 that
> Fujifilm got about right. Instead, there is an 18mm f/2 (???) and a
> 23mm f/2, with presumably the same optics as the X100 lens.
>
> There will be two zooms, a 12-20mm f/4 and a 70-200mm f/4, both image
> stabilised. I don't think anyone in their right mind would buy a
> 70-200mm lens for a 'rangefinder' camera with an APS-C sensor.
>
> So I am not convinced by the X-Pro1. The Leica M10 will probably cost
> four times as much, but it will be the complete digital rangefinder
> camera with lenses in focal lengths that people want. There is a
> little over a month to go before the M10 is announced, possibly
> accompanied by Leica's take on a mirrorless system camera.
35mm? Personally, I like how they are breaking from tradition with
different focal lengths than what we are used to getting from Canon
and Nikon. They aren't the first to do that. A 12-20mm f4.0 is
interesting, not quite 2x and not that fast, so maybe it's to keep the
size reasonable? Also, it will be sharper across the board than an
f2.8., all things being equal. As for the 70-200mm, it probably makes
more sense here with a reasonable sized body than on some of the other
more diminutive mirrorless offerings from other companies.
David Dyer-Bennet
Apr 4, 2012, 11:00:51 AM4/4/12
to
Mxsmanic <mxsm...@gmail.com> writes:
> Me writes:
>
>> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
>> theoretically gets rid of *colour moire, and possibly reduces luminance
>> moire (by more random placement of green sensels than a repeating bayer
>> pattern), so there's less need for a low pass filter.
>
> It would also reduce color resolution. If it's actually being done, it sounds
> like a gimmick.
No, it wouldn't. It has the same proportions of pixels in each color,
> Me writes:
>
>> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
>> theoretically gets rid of *colour moire, and possibly reduces luminance
>> moire (by more random placement of green sensels than a repeating bayer
>> pattern), so there's less need for a low pass filter.
>
> It would also reduce color resolution. If it's actually being done, it sounds
> like a gimmick.
the placement is simply less regular -- more like film grain, less like
a tic-tac-toe board.
> The whole issue of aliasing is past history; I don't know why people still
> worry about it. When you have a large number of pixels, you don't need to
> worry about aliasing. So the ultimate solution for aliasing is more pixels.
> There will always be aliasing of details that are sufficiently small, if the
> lens can resolve them, but the aliasing will not be noticeable or
> troublesome.
> Even in the old days, when sensors had far fewer pixels, I never really
> noticed problems with aliasing. I'd rather take the risk of aliasing than put
> up with the blur of anti-aliasing.
field conditions, you're MUCH better off with the AA filter.
--
David Dyer-Bennet, dd...@dd-b.net; http://dd-b.net/
Snapshots: http://dd-b.net/dd-b/SnapshotAlbum/data/
Photos: http://dd-b.net/photography/gallery/
Dragaera: http://dragaera.info
RichA
Apr 4, 2012, 10:58:26 AM4/4/12
to
On Apr 4, 9:31 am, Andrew Haley <andre...@littlepinkcloud.invalid>
wrote:
P&S? Anyone know how that turned out?
wrote:
> Me <u...@example.net> wrote:
> > On 4/04/2012 5:30 p.m., Mxsmanic wrote:
> >> Rich writes:
>
> >>> I don't hate the Bayer. I think it's a clever way to deal with a problem
> >>> that some find objectionable. But its days could be numbered.
> >>> Don't know if resolution is improved yet, but it's nice to know that it
> >>> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> >>> $500 DSLRs that meet the spec.
>
> >> A Bayer filter and an anti-aliasing filter are two different things.
> > IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> > theoretically gets rid of *colour moire, and possibly reduces luminance
> > moire (by more random placement of green sensels than a repeating bayer
> > pattern), so there's less need for a low pass filter.
>
> It's nothing like pseudo-random.http://www.dpreview.com/previews/fujifilmxpro1/3shows that it's a 6x6
> > On 4/04/2012 5:30 p.m., Mxsmanic wrote:
> >> Rich writes:
>
> >>> I don't hate the Bayer. I think it's a clever way to deal with a problem
> >>> that some find objectionable. But its days could be numbered.
> >>> Don't know if resolution is improved yet, but it's nice to know that it
> >>> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> >>> $500 DSLRs that meet the spec.
>
> >> A Bayer filter and an anti-aliasing filter are two different things.
> > IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> > theoretically gets rid of *colour moire, and possibly reduces luminance
> > moire (by more random placement of green sensels than a repeating bayer
> > pattern), so there's less need for a low pass filter.
>
> array instead of a 2x2 array. This 6x6 array is rotationally
> symmetrical, and it repeats across the sensor. It has the interesting
> property that there are fewer blue and red pixels than a Bayer sensor
> has: 8R:20G:8B instead of 9:18:9. It probably won't make much
> difference.
>
> Andrew.
Didn't Sony produce a sensor with yellow pixels at one point in a
> symmetrical, and it repeats across the sensor. It has the interesting
> property that there are fewer blue and red pixels than a Bayer sensor
> has: 8R:20G:8B instead of 9:18:9. It probably won't make much
> difference.
>
> Andrew.
P&S? Anyone know how that turned out?
Bruce
Apr 4, 2012, 11:29:46 AM4/4/12
to
deficient - and that's in spite of having 50% of the pixels against
25% for each of red and blue.
Bruce
Apr 4, 2012, 2:58:58 PM4/4/12
to
RichA <rande...@gmail.com> wrote:
>I thought the practical limit for rangefinder focusing was 135mm for
>35mm? Personally, I like how they are breaking from tradition with
>different focal lengths than what we are used to getting from Canon
>and Nikon. They aren't the first to do that. A 12-20mm f4.0 is
>interesting, not quite 2x and not that fast, so maybe it's to keep the
>size reasonable? Also, it will be sharper across the board than an
>f2.8., all things being equal. As for the 70-200mm, it probably makes
>more sense here with a reasonable sized body than on some of the other
>more diminutive mirrorless offerings from other companies.
How are you getting along with your new GH2, Rich?
>I thought the practical limit for rangefinder focusing was 135mm for
>35mm? Personally, I like how they are breaking from tradition with
>different focal lengths than what we are used to getting from Canon
>and Nikon. They aren't the first to do that. A 12-20mm f4.0 is
>interesting, not quite 2x and not that fast, so maybe it's to keep the
>size reasonable? Also, it will be sharper across the board than an
>f2.8., all things being equal. As for the 70-200mm, it probably makes
>more sense here with a reasonable sized body than on some of the other
>more diminutive mirrorless offerings from other companies.
Me
Apr 4, 2012, 3:31:18 PM4/4/12
to
Fuji's website describing the array pattern as semi-random. It looks
like it's a lot more pseudo and a lot less random than I'd expected.
Me
Apr 4, 2012, 3:50:52 PM4/4/12
to
Sharp make TV panels with yellow pixels. This seems to be >99% BS.
Sometimes competition/marketing ends up driving complex and even elegant
solutions to problems which never existed.
RichA
Apr 4, 2012, 4:04:35 PM4/4/12
to
On Apr 4, 11:29 am, Bruce <docnews2...@gmail.com> wrote:
> Andrew Haley <andre...@littlepinkcloud.invalid> wrote:
> Andrew Haley <andre...@littlepinkcloud.invalid> wrote:
> >Me <u...@example.net> wrote:
> >> On 4/04/2012 5:30 p.m., Mxsmanic wrote:
> >>> Rich writes:
>
> >>>> I don't hate the Bayer. I think it's a clever way to deal with a problem
> >>>> that some find objectionable. But its days could be numbered.
> >>>> Don't know if resolution is improved yet, but it's nice to know that it
> >>>> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> >>>> $500 DSLRs that meet the spec.
>
> >>> A Bayer filter and an anti-aliasing filter are two different things.
> >> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> >> theoretically gets rid of *colour moire, and possibly reduces luminance
> >> moire (by more random placement of green sensels than a repeating bayer
> >> pattern), so there's less need for a low pass filter.
>
> >It's nothing like pseudo-random.
> >http://www.dpreview.com/previews/fujifilmxpro1/3shows that it's a 6x6
> >> On 4/04/2012 5:30 p.m., Mxsmanic wrote:
> >>> Rich writes:
>
> >>>> I don't hate the Bayer. I think it's a clever way to deal with a problem
> >>>> that some find objectionable. But its days could be numbered.
> >>>> Don't know if resolution is improved yet, but it's nice to know that it
> >>>> only costs $1700(!) to get a non-flawed Fuji when mainstream makers produce
> >>>> $500 DSLRs that meet the spec.
>
> >>> A Bayer filter and an anti-aliasing filter are two different things.
> >> IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> >> theoretically gets rid of *colour moire, and possibly reduces luminance
> >> moire (by more random placement of green sensels than a repeating bayer
> >> pattern), so there's less need for a low pass filter.
>
> >It's nothing like pseudo-random.
> >array instead of a 2x2 array. This 6x6 array is rotationally
> >symmetrical, and it repeats across the sensor. It has the interesting
> >property that there are fewer blue and red pixels than a Bayer sensor
> >has: 8R:20G:8B instead of 9:18:9. It probably won't make much
> >difference.
>
> More green is welcome, because that is where the Bayer pattern is
> deficient - and that's in spite of having 50% of the pixels against
> 25% for each of red and blue.
True. You look at the black-body charts for sensors and green is
> >symmetrical, and it repeats across the sensor. It has the interesting
> >property that there are fewer blue and red pixels than a Bayer sensor
> >has: 8R:20G:8B instead of 9:18:9. It probably won't make much
> >difference.
>
> More green is welcome, because that is where the Bayer pattern is
> deficient - and that's in spite of having 50% of the pixels against
> 25% for each of red and blue.
lacking.
RichA
Apr 4, 2012, 4:04:11 PM4/4/12
to
On Apr 4, 2:58 pm, Bruce <docnews2...@gmail.com> wrote:
I traded it back and ordered the E-5M. The 14-140mm had zoom-
creep. But honestly, I wanted the Olympus in the first place.
creep. But honestly, I wanted the Olympus in the first place.
David Dyer-Bennet
Apr 4, 2012, 5:53:16 PM4/4/12
to
diagram of the actual filter arrangement, yes.
Chris Malcolm
Apr 4, 2012, 7:02:18 PM4/4/12
to
images in a lab. Aliasing was a very serious problem! But accuracy was
sufficiently important that aliasing was dealt with by trying to
recognise the artefacts and the probable underlying reality (such as a
staircased stright edge) rather than trying to avoid aliasing by some
kind of blurring.
--
Chris Malcolm
Mxsmanic
Apr 5, 2012, 1:01:52 AM4/5/12
to
David Dyer-Bennet writes:
> No, it wouldn't. It has the same proportions of pixels in each color,
> the placement is simply less regular -- more like film grain, less like
> a tic-tac-toe board.
Less regular means more anomalies in the recording of color. Yes, that's like
> No, it wouldn't. It has the same proportions of pixels in each color,
> the placement is simply less regular -- more like film grain, less like
> a tic-tac-toe board.
grain ... but is grain what you want?
> I've seen recent examples of pictures containing horrible aliasing.
preferable to blurring every single photo just on the off chance that there
might be a problem with aliasing.
> You mostly haven't had that choice. And if you're working fast and in
> field conditions, you're MUCH better off with the AA filter.
that I can make any necessary adjustments in post. Changes made in the camera
cannot be undone, and if they ruin a photo, you're out of luck.
That's why so many DSLRs actually produce poor video, incidentally.
Mxsmanic
Apr 5, 2012, 1:02:59 AM4/5/12
to
Andrew Haley writes:
> It's nothing like pseudo-random.
> http://www.dpreview.com/previews/fujifilmxpro1/3 shows that it's a 6x6
> array instead of a 2x2 array. This 6x6 array is rotationally
> symmetrical, and it repeats across the sensor. It has the interesting
> property that there are fewer blue and red pixels than a Bayer sensor
> has: 8R:20G:8B instead of 9:18:9. It probably won't make much
> difference.
I prefer more blue and red, not less. I'd like an equal number of green, red,
> It's nothing like pseudo-random.
> http://www.dpreview.com/previews/fujifilmxpro1/3 shows that it's a 6x6
> array instead of a 2x2 array. This 6x6 array is rotationally
> symmetrical, and it repeats across the sensor. It has the interesting
> property that there are fewer blue and red pixels than a Bayer sensor
> has: 8R:20G:8B instead of 9:18:9. It probably won't make much
> difference.
and blue pixels, in sufficiently high number to make favoritism for green a
moot point.
Mxsmanic
Apr 5, 2012, 1:03:54 AM4/5/12
to
RichA writes:
> Didn't Sony produce a sensor with yellow pixels at one point in a
> P&S? Anyone know how that turned out?
I know there are video cameras that have done this, but I don't know about
> Didn't Sony produce a sensor with yellow pixels at one point in a
> P&S? Anyone know how that turned out?
still cameras.
Using cyan, yellow, and magenta instead of red, blue, and green increases
light sensitivity by a factor of 2, but at the expensive of color resolution
and accuracy.
Mxsmanic
Apr 5, 2012, 1:05:15 AM4/5/12
to
Bruce writes:
> More green is welcome, because that is where the Bayer pattern is
> deficient - and that's in spite of having 50% of the pixels against
> 25% for each of red and blue.
How is having twice as many green pixels "deficient" in green?
> More green is welcome, because that is where the Bayer pattern is
> deficient - and that's in spite of having 50% of the pixels against
> 25% for each of red and blue.
Photos already have problems with shimmering reds and blues because there are
too few pixels of both. And remember that lossy compression often reduces blue
and red resolution even further.
David Dyer-Bennet
Apr 5, 2012, 11:53:33 AM4/5/12
to
Mxsmanic <mxsm...@gmail.com> writes:
> David Dyer-Bennet writes:
>
>> No, it wouldn't. It has the same proportions of pixels in each color,
>> the placement is simply less regular -- more like film grain, less like
>> a tic-tac-toe board.
>
> Less regular means more anomalies in the recording of color. Yes, that's like
> grain ... but is grain what you want?
Not necessarily visible anomalies, though.
> David Dyer-Bennet writes:
>
>> No, it wouldn't. It has the same proportions of pixels in each color,
>> the placement is simply less regular -- more like film grain, less like
>> a tic-tac-toe board.
>
> Less regular means more anomalies in the recording of color. Yes, that's like
> grain ... but is grain what you want?
>> I've seen recent examples of pictures containing horrible aliasing.
>
> If there is horrible aliasing, it can be adjusted in post. That's way
> preferable to blurring every single photo just on the off chance that there
> might be a problem with aliasing.
adjust in post. The good solution is to work slowly and carefully so
you catch it immediately; some minor adjustment to the shot will
generally eliminate it (this from multiple people working regularly with
medium-format backs lacking AA filters).
>> You mostly haven't had that choice. And if you're working fast and in
>> field conditions, you're MUCH better off with the AA filter.
>
> Nope, I want the output of the camera to be as unmanipulated as possible, so
> that I can make any necessary adjustments in post. Changes made in the camera
> cannot be undone, and if they ruin a photo, you're out of luck.
tradeoff for essentially any fast-moving photography excecpt perhaps
certain kinds of wildlife.
> That's why so many DSLRs actually produce poor video, incidentally.
David Dyer-Bennet
Apr 5, 2012, 11:54:23 AM4/5/12
to
behavior of the human eye very well, so it'll look less good than a less
"fair" design does (to human eyes).
David Dyer-Bennet
Apr 5, 2012, 11:54:58 AM4/5/12
to
early DSLRs or later consumer models.
Mxsmanic
Apr 5, 2012, 6:48:11 PM4/5/12
to
David Dyer-Bennet writes:
> Strange how well the professionals like the output.
Not necessarily after they've seen side-by-side comparisons, as I have.
> Strange how well the professionals like the output.
DSLRs tend to compress and compromise enormously on video, since they aren't
really designed for that. The result is often images that are software than
images from dedicated video cameras, even though DSLRs usually have sensors
with far more pixels.
I think some pros like having limited depth of field, which is hard to obtain
with small video sensors. But the requirements of still photography conflict
with those of video, so you cannot do both well at the same time.
Mxsmanic
Apr 5, 2012, 6:49:31 PM4/5/12
to
David Dyer-Bennet writes:
> You're probably not going to get it; because it doesn't match the
> behavior of the human eye very well, so it'll look less good than a less
> "fair" design does (to human eyes).
It will look just as well. After all, the real world has the three colors
> You're probably not going to get it; because it doesn't match the
> behavior of the human eye very well, so it'll look less good than a less
> "fair" design does (to human eyes).
balanced, too.
A Bayer filter is compensation for an inadequate number of pixels to begin
with. If you have enough pixels--or three sensors--you don't need a Bayer
pattern.
TheRealSteve
Apr 5, 2012, 9:34:42 PM4/5/12
to
On Thu, 05 Apr 2012 07:50:52 +1200, Me <us...@example.net> wrote:
>IIRC that was white pixels - but then again nothing would surprise me.
>Sharp make TV panels with yellow pixels. This seems to be >99% BS.
>Sometimes competition/marketing ends up driving complex and even elegant
>solutions to problems which never existed.
TV watching something with turquiose blues or bright yellows. I did it
with one of those planet earth shows with bright tropical fish against
a similar Samsung, Toshiba, Pioneer and a few others. You'll be
surprised at the difference.
It would make a good monitor for photo processing since it seems to be
able to display a wider gamut of colors.
Steve
Me
Apr 5, 2012, 10:45:28 PM4/5/12
to
bottom of range, CCFL 1080p) Sony 40", I can't tell the difference. I'm
not going to set them up side by side to test - that would be kind of
silly - the "general impression" is plenty good enough - and there's far
more quality variation between source material than there is between any
two TV sets.
The general impression is that the LED backlight in the Sharp is waaaaay
too bright, default "AV modes" were pretty hideous, and the auto ambient
light sensor worked better in the Sony. But these can tweaked, so it's
an OK TV set.
The colour that bugs me most is blown reds, but that's not the fault of
the TV.
nospam
Apr 7, 2012, 5:07:46 PM4/7/12
to
In article <6idon7div4ipocdf6...@4ax.com>, Mxsmanic
can resolve so there's no issue if it does reduce it.
> The whole issue of aliasing is past history; I don't know why people still
> worry about it.
aliasing is definitely not past history. it's still very real.
> When you have a large number of pixels, you don't need to
> worry about aliasing.
not as much.
> So the ultimate solution for aliasing is more pixels.
> There will always be aliasing of details that are sufficiently small, if the
> lens can resolve them, but the aliasing will not be noticeable or troublesome.
it might be noticeable.
> Even in the old days, when sensors had far fewer pixels, I never really
> noticed problems with aliasing.
then you are either blind or don't know what to look for.
> I'd rather take the risk of aliasing than put
> up with the blur of anti-aliasing.
it doesn't blur.
<mxsm...@gmail.com> wrote:
> > IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> > theoretically gets rid of *colour moire, and possibly reduces luminance
> > moire (by more random placement of green sensels than a repeating bayer
> > pattern), so there's less need for a low pass filter.
>
> It would also reduce color resolution. If it's actually being done, it sounds
> like a gimmick.
not really, but bayer already has higher colour resolution than the eye
> > IIRC this camera uses a "pseudo random" pattern for RGBG filters, which
> > theoretically gets rid of *colour moire, and possibly reduces luminance
> > moire (by more random placement of green sensels than a repeating bayer
> > pattern), so there's less need for a low pass filter.
>
> It would also reduce color resolution. If it's actually being done, it sounds
> like a gimmick.
can resolve so there's no issue if it does reduce it.
> The whole issue of aliasing is past history; I don't know why people still
> worry about it.
> When you have a large number of pixels, you don't need to
> worry about aliasing.
> So the ultimate solution for aliasing is more pixels.
> There will always be aliasing of details that are sufficiently small, if the
> lens can resolve them, but the aliasing will not be noticeable or troublesome.
> Even in the old days, when sensors had far fewer pixels, I never really
> noticed problems with aliasing.
> I'd rather take the risk of aliasing than put
> up with the blur of anti-aliasing.
nospam
Apr 7, 2012, 5:07:47 PM4/7/12
to
In article
<c2be0db2-5406-417a...@j14g2000vbc.googlegroups.com>,
since it's no longer made.
<c2be0db2-5406-417a...@j14g2000vbc.googlegroups.com>,
RichA <rande...@gmail.com> wrote:
> Didn't Sony produce a sensor with yellow pixels at one point in a
> P&S? Anyone know how that turned out?
they used emerald as the 4th colour and it didn't work out too well
> Didn't Sony produce a sensor with yellow pixels at one point in a
> P&S? Anyone know how that turned out?
since it's no longer made.
nospam
Apr 7, 2012, 5:07:48 PM4/7/12
to
In article <s4qon79bihq6ihc9j...@4ax.com>, Bruce
<docne...@gmail.com> wrote:
> More green is welcome, because that is where the Bayer pattern is
> deficient - and that's in spite of having 50% of the pixels against
> 25% for each of red and blue.
how is it deficient if it has twice as many?
> More green is welcome, because that is where the Bayer pattern is
> deficient - and that's in spite of having 50% of the pixels against
> 25% for each of red and blue.
nospam
Apr 7, 2012, 5:07:49 PM4/7/12
to
In article <eu9qn75o8o9ag8ddl...@4ax.com>, Mxsmanic
> And remember that lossy compression often reduces blue
> and red resolution even further.
nothing you can see.
<mxsm...@gmail.com> wrote:
> Photos already have problems with shimmering reds and blues because there are
> too few pixels of both.
no they don't.
> Photos already have problems with shimmering reds and blues because there are
> too few pixels of both.
> And remember that lossy compression often reduces blue
> and red resolution even further.
nospam
Apr 7, 2012, 5:07:50 PM4/7/12
to
In article <as9qn71an3ag7256c...@4ax.com>, Mxsmanic
conversion to rgb. there's a reason nobody does this anymore.
<mxsm...@gmail.com> wrote:
> Using cyan, yellow, and magenta instead of red, blue, and green increases
> light sensitivity by a factor of 2, but at the expensive of color resolution
> and accuracy.
no, it's a wash. you get better sensitivity but you lose it in
> Using cyan, yellow, and magenta instead of red, blue, and green increases
> light sensitivity by a factor of 2, but at the expensive of color resolution
> and accuracy.
conversion to rgb. there's a reason nobody does this anymore.
nospam
Apr 7, 2012, 5:07:51 PM4/7/12
to
In article <rl9qn79i8akdtcebt...@4ax.com>, Mxsmanic
> That's way
> preferable to blurring every single photo just on the off chance that there
> might be a problem with aliasing.
no it isn't because the fix will affect real details.
> > You mostly haven't had that choice. And if you're working fast and in
> > field conditions, you're MUCH better off with the AA filter.
>
> Nope, I want the output of the camera to be as unmanipulated as possible, so
> that I can make any necessary adjustments in post. Changes made in the camera
> cannot be undone, and if they ruin a photo, you're out of luck.
shoot raw.
> That's why so many DSLRs actually produce poor video, incidentally.
actually, very good video.
<mxsm...@gmail.com> wrote:
> > I've seen recent examples of pictures containing horrible aliasing.
>
> If there is horrible aliasing, it can be adjusted in post.
no it can't.
> > I've seen recent examples of pictures containing horrible aliasing.
>
> If there is horrible aliasing, it can be adjusted in post.
> That's way
> preferable to blurring every single photo just on the off chance that there
> might be a problem with aliasing.
> > You mostly haven't had that choice. And if you're working fast and in
> > field conditions, you're MUCH better off with the AA filter.
>
> Nope, I want the output of the camera to be as unmanipulated as possible, so
> that I can make any necessary adjustments in post. Changes made in the camera
> cannot be undone, and if they ruin a photo, you're out of luck.
> That's why so many DSLRs actually produce poor video, incidentally.
nospam
Apr 7, 2012, 5:07:51 PM4/7/12
to
In article <ylfkpqbm...@dd-b.net>, David Dyer-Bennet
particularly if the moire is low frequency banding.
<dd...@dd-b.net> wrote:
> > If there is horrible aliasing, it can be adjusted in post. That's way
> > preferable to blurring every single photo just on the off chance that there
> > might be a problem with aliasing.
>
> Nope, horrible aliasing (generally moire) is really terribly hard to
> adjust in post.
it's impossible. if you remove the moire you will remove real detail,
> > If there is horrible aliasing, it can be adjusted in post. That's way
> > preferable to blurring every single photo just on the off chance that there
> > might be a problem with aliasing.
>
> Nope, horrible aliasing (generally moire) is really terribly hard to
> adjust in post.
particularly if the moire is low frequency banding.
nospam
Apr 7, 2012, 5:07:52 PM4/7/12
to
In article <nlhsn756qnhc7vulb...@4ax.com>, TheRealSteve
there are wide gamut lcd displays that can display argb, and that's
without that nonsense.
without that nonsense.
nospam
Apr 7, 2012, 5:07:53 PM4/7/12
to
In article <3a8sn79psk00dindv...@4ax.com>, Mxsmanic
> A Bayer filter is compensation for an inadequate number of pixels to begin
> with. If you have enough pixels--or three sensors--you don't need a Bayer
> pattern.
nonsense, and 3 sensors is *very* impractical. even pro video is moving
to bayer (red).
<mxsm...@gmail.com> wrote:
> > You're probably not going to get it; because it doesn't match the
> > behavior of the human eye very well, so it'll look less good than a less
> > "fair" design does (to human eyes).
>
> It will look just as well. After all, the real world has the three colors
> balanced, too.
the eye's response is not equal on all colours.
> > You're probably not going to get it; because it doesn't match the
> > behavior of the human eye very well, so it'll look less good than a less
> > "fair" design does (to human eyes).
>
> It will look just as well. After all, the real world has the three colors
> balanced, too.
> A Bayer filter is compensation for an inadequate number of pixels to begin
> with. If you have enough pixels--or three sensors--you don't need a Bayer
> pattern.
to bayer (red).
Peter Irwin
Apr 7, 2012, 6:16:28 PM4/7/12
to
more like 2/3 green.
Peter.
Mxsmanic
Apr 7, 2012, 10:09:21 PM4/7/12
to
nospam writes:
> it's impossible. if you remove the moire you will remove real detail,
> particularly if the moire is low frequency banding.
So does an AA filter.
> it's impossible. if you remove the moire you will remove real detail,
> particularly if the moire is low frequency banding.
Mxsmanic
Apr 7, 2012, 10:10:59 PM4/7/12
to
nospam writes:
> no it isn't because the fix will affect real details.
An AA filter simply deprives you of those details to begin with, whether you
> no it isn't because the fix will affect real details.
want it to or not.
> shoot raw.
If raw is available. But that won't help if there's a AA filter on the sensor.
> actually, very good video.
Actually very poor. The performance of a DSLR for video typically is nothing
like its performance for still photos. I've seen and done the side-by-side
comparisons.
Mxsmanic
Apr 7, 2012, 10:12:11 PM4/7/12
to
nospam writes:
> not really, but bayer already has higher colour resolution than the eye
> can resolve so there's no issue if it does reduce it.
Color resolution doesn't depend just on the filter pattern. And comparisons to
> not really, but bayer already has higher colour resolution than the eye
> can resolve so there's no issue if it does reduce it.
what the eye can resolve involve many factors. I still see shimmering reds and
blues in photos, so there's still a problem.
Mxsmanic
Apr 7, 2012, 10:14:02 PM4/7/12
to
nospam writes:
> the eye's response is not equal on all colours.
You may want to do something with the photos besides look at them. And the
> the eye's response is not equal on all colours.
solution to inadequate color is more and better pixels.
> nonsense, and 3 sensors is *very* impractical. even pro video is moving
> to bayer (red).
nospam
Apr 8, 2012, 12:02:09 AM4/8/12
to
nospam
Apr 8, 2012, 12:02:10 AM4/8/12
to
In article <c1t1o790epo9vmiu4...@4ax.com>, Mxsmanic
if you mean image analysis, then there are special cameras for that
purpose.
> > nonsense, and 3 sensors is *very* impractical. even pro video is moving
> > to bayer (red).
>
> Three sensors have worked very well for video for decades.
they have, and now they've been replaced by something better, a bayer
sensor.
<mxsm...@gmail.com> wrote:
> > the eye's response is not equal on all colours.
>
> You may want to do something with the photos besides look at them. And the
> solution to inadequate color is more and better pixels.
what else would i want do with photos than look at them?
> > the eye's response is not equal on all colours.
>
> You may want to do something with the photos besides look at them. And the
> solution to inadequate color is more and better pixels.
if you mean image analysis, then there are special cameras for that
purpose.
> > nonsense, and 3 sensors is *very* impractical. even pro video is moving
> > to bayer (red).
>
> Three sensors have worked very well for video for decades.
sensor.
nospam
Apr 8, 2012, 12:02:06 AM4/8/12
to
In article <7qs1o7dk6gsqq1glb...@4ax.com>, Mxsmanic
wrong. an aa filter prevents the moire from happening in the first
place, so the real detail won't need to be altered.
place, so the real detail won't need to be altered.
nospam
Apr 8, 2012, 12:02:08 AM4/8/12
to
In article <vqs1o79iipr0f7qdd...@4ax.com>, Mxsmanic
cannot resolve, so you aren't losing anything. take away the aa filter
and you won't get those details, you'll get fake details, aka alias
artifacts.
> > shoot raw.
>
> If raw is available. But that won't help if there's a AA filter on the sensor.
two totally different things.
> > actually, very good video.
>
> Actually very poor. The performance of a DSLR for video typically is nothing
> like its performance for still photos. I've seen and done the side-by-side
> comparisons.
millions of others have seen both and they're buying dslrs *for* video
because they are very good and a fraction of the price of what they
were buying, particularly pros.
<mxsm...@gmail.com> wrote:
> > no it isn't because the fix will affect real details.
>
> An AA filter simply deprives you of those details to begin with, whether you
> want it to or not.
no it doesn't. what an aa filter does is reduce the details the sensor
> > no it isn't because the fix will affect real details.
>
> An AA filter simply deprives you of those details to begin with, whether you
> want it to or not.
cannot resolve, so you aren't losing anything. take away the aa filter
and you won't get those details, you'll get fake details, aka alias
artifacts.
> > shoot raw.
>
> If raw is available. But that won't help if there's a AA filter on the sensor.
> > actually, very good video.
>
> Actually very poor. The performance of a DSLR for video typically is nothing
> like its performance for still photos. I've seen and done the side-by-side
> comparisons.
because they are very good and a fraction of the price of what they
were buying, particularly pros.
Alfred Molon
Apr 8, 2012, 5:07:08 AM4/8/12
to
In article <070420121707464413%nos...@nospam.invalid>, nospam says...
well appreciate the colour details.
Besides, the Bayer interpolation impacts the overall image resolution,
not just the colour one.
But again, it's pointless to conduct this discussion with you, because
you refuse to accept simple facts, and routinely go into a "no-no-no"
denial mode. Maybe I should filter out your posts and save my time on
other issues.
--
Alfred Molon
------------------------------
Olympus E-series DSLRs and micro 4/3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site
> not really, but bayer already has higher colour resolution than the eye
> can resolve so there's no issue if it does reduce it.
Irrelevant, because if you enlarge the image enough the eye can very
> can resolve so there's no issue if it does reduce it.
well appreciate the colour details.
Besides, the Bayer interpolation impacts the overall image resolution,
not just the colour one.
But again, it's pointless to conduct this discussion with you, because
you refuse to accept simple facts, and routinely go into a "no-no-no"
denial mode. Maybe I should filter out your posts and save my time on
other issues.
--
Alfred Molon
------------------------------
Olympus E-series DSLRs and micro 4/3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site
David J Taylor
Apr 8, 2012, 5:46:10 AM4/8/12
to
"Alfred Molon" <alfred...@yahoo.com> wrote in message
news:MPG.29eb7384...@news.supernews.com...
[]
> --
>
> Alfred Molon
Alfred,
If you enlarge a print enough you can probably see the individual fibres
in the paper, and if you enlarge a display enough you can see the
individual RGB pixels. The image need only contain enough resolution
(whether it be colour or monochrome) for the intended application.
What point is there in presenting a final image of greater resolution than
the eye can perceive? In engineering terms, it is a waste.
David
news:MPG.29eb7384...@news.supernews.com...
[]
> Irrelevant, because if you enlarge the image enough the eye can very
> well appreciate the colour details.
>
> Besides, the Bayer interpolation impacts the overall image resolution,
> not just the colour one.
[]
> well appreciate the colour details.
>
> Besides, the Bayer interpolation impacts the overall image resolution,
> not just the colour one.
> --
>
> Alfred Molon
Alfred,
If you enlarge a print enough you can probably see the individual fibres
in the paper, and if you enlarge a display enough you can see the
individual RGB pixels. The image need only contain enough resolution
(whether it be colour or monochrome) for the intended application.
What point is there in presenting a final image of greater resolution than
the eye can perceive? In engineering terms, it is a waste.
David
Mxsmanic
Apr 8, 2012, 7:04:16 AM4/8/12
to
nospam writes:
> wrong. an aa filter prevents the moire from happening in the first
> place, so the real detail won't need to be altered.
The AA filter works by removing the real detail.
> wrong. an aa filter prevents the moire from happening in the first
> place, so the real detail won't need to be altered.
Mxsmanic
Apr 8, 2012, 7:06:11 AM4/8/12
to
nospam writes:
> no it doesn't. what an aa filter does is reduce the details the sensor
> cannot resolve, so you aren't losing anything.
The same is true of correcting moiré in post. They are exactly equivalent.
> no it doesn't. what an aa filter does is reduce the details the sensor
> cannot resolve, so you aren't losing anything.
The problem is that the cutoff of an AA filter is never sharp enough, so
either it always removes details that the sensor _can_ resolve, or it lets
through details that it cannot.
> millions of others have seen both and they're buying dslrs *for* video
> because they are very good and a fraction of the price of what they
> were buying, particularly pros.
Mxsmanic
Apr 8, 2012, 7:07:22 AM4/8/12
to
David J Taylor writes:
> What point is there in presenting a final image of greater resolution than
> the eye can perceive? In engineering terms, it is a waste.
Sometimes a "waste" is justified in order to provide a safety margin.
> What point is there in presenting a final image of greater resolution than
> the eye can perceive? In engineering terms, it is a waste.
Mxsmanic
Apr 8, 2012, 7:08:32 AM4/8/12
to
nospam writes:
> what else would i want do with photos than look at them?
You could subject them to computer analysis.
> what else would i want do with photos than look at them?
> if you mean image analysis, then there are special cameras for that
> purpose.
> they have, and now they've been replaced by something better, a bayer
> sensor.
much better images with much higher color and luminance resolution.
David J Taylor
Apr 8, 2012, 7:57:05 AM4/8/12
to
"Mxsmanic" <mxsm...@gmail.com> wrote in message
news:q4s2o71njvo3ue4v6...@4ax.com...
The AA filter works by removing the detail which cannot be accurately
recorded by a sensor made up of discrete pixels (at regular intervals).
The lack of such filtering will lead to artefacts such as steps on near
horizontal or near vertical edges, and false-colours on finely textured
material, and these artefacts cannot be removed in post-processing.
David
news:q4s2o71njvo3ue4v6...@4ax.com...
recorded by a sensor made up of discrete pixels (at regular intervals).
The lack of such filtering will lead to artefacts such as steps on near
horizontal or near vertical edges, and false-colours on finely textured
material, and these artefacts cannot be removed in post-processing.
David
David J Taylor
Apr 8, 2012, 7:59:02 AM4/8/12
to
"Mxsmanic" <mxsm...@gmail.com> wrote in message
news:las2o7pru0kr1tj4b...@4ax.com...
so large that the individual pixels are seen).
David
David J Taylor
Apr 8, 2012, 8:00:59 AM4/8/12
to
"Mxsmanic" <mxsm...@gmail.com> wrote in message
news:mbs2o75k5ml5lgnta...@4ax.com...
[]
> The Bayer sensor is cheaper, not better. A three-sensor camera would
> provide
> much better images with much higher color and luminance resolution.
Alignment issues and lens issues make such an arrangement either very
> provide
> much better images with much higher color and luminance resolution.
costly or physically impossible at high resolutions, and the higher colour
resolution isn't needed for human vision.
David
Martin Brown
Apr 8, 2012, 9:00:31 AM4/8/12
to
On 08/04/2012 13:00, David J Taylor wrote:
> "Mxsmanic" <mxsm...@gmail.com> wrote in message
> news:mbs2o75k5ml5lgnta...@4ax.com...
> []
>> The Bayer sensor is cheaper, not better. A three-sensor camera would
>> provide
>> much better images with much higher color and luminance resolution.
Only in your fevered imagination! There is a small hit in resolution
> "Mxsmanic" <mxsm...@gmail.com> wrote in message
> news:mbs2o75k5ml5lgnta...@4ax.com...
> []
>> The Bayer sensor is cheaper, not better. A three-sensor camera would
>> provide
>> much better images with much higher color and luminance resolution.
with Bayer but it is not significant except on pathological test cards.
The only time when Bayer is at a noticeable resolution disadvantage over
tricolour sensor cameras is when the image is pure monochromatic red or
blue. This happens for instance when using a solar prominence filter for
instance. I can't think of any other real world situations where the
luminance detail is compromised by using Bayer filter mask.
> Alignment issues and lens issues make such an arrangement either very
> costly or physically impossible at high resolutions, and the higher
> colour resolution isn't needed for human vision.
they do is to downsample the chroma info before encoding as Y Cr Cb.
There is no point transmitting detail that the human eye cannot see.
--
Regards,
Martin Brown
nospam
Apr 8, 2012, 11:37:01 AM4/8/12
to
In article <q4s2o71njvo3ue4v6...@4ax.com>, Mxsmanic
only slightly, since a perfect filter does not exist.
nospam
Apr 8, 2012, 11:37:02 AM4/8/12
to
In article <06s2o7h9veg8n90vc...@4ax.com>, Mxsmanic
> > millions of others have seen both and they're buying dslrs *for* video
> > because they are very good and a fraction of the price of what they
> > were buying, particularly pros.
>
> They aren't cinematographers.
some are.
<mxsm...@gmail.com> wrote:
> The problem is that the cutoff of an AA filter is never sharp enough, so
> either it always removes details that the sensor _can_ resolve, or it lets
> through details that it cannot.
there's always a tradeoff. we don't live in a perfect world.
> The problem is that the cutoff of an AA filter is never sharp enough, so
> either it always removes details that the sensor _can_ resolve, or it lets
> through details that it cannot.
> > millions of others have seen both and they're buying dslrs *for* video
> > because they are very good and a fraction of the price of what they
> > were buying, particularly pros.
>
> They aren't cinematographers.
nospam
Apr 8, 2012, 11:37:03 AM4/8/12
to
In article <mbs2o75k5ml5lgnta...@4ax.com>, Mxsmanic
and software.
meanwhile, the vast majority of photos are taken for human consumption.
> > if you mean image analysis, then there are special cameras for that
> > purpose.
>
> A special camera isn't always available.
image analysis isn't always necessary.
> > they have, and now they've been replaced by something better, a bayer
> > sensor.
>
> The Bayer sensor is cheaper, not better. A three-sensor camera would provide
> much better images with much higher color and luminance resolution.
except that maintaining alignment with high resolution sensors is all
but impossible even in controlled lab conditions, and the colour
resolution is already higher than what you can see so more is a waste.
<mxsm...@gmail.com> wrote:
> > what else would i want do with photos than look at them?
>
> You could subject them to computer analysis.
i could and if i wanted to do that, i'd get the appropriate equipment
> > what else would i want do with photos than look at them?
>
> You could subject them to computer analysis.
and software.
meanwhile, the vast majority of photos are taken for human consumption.
> > if you mean image analysis, then there are special cameras for that
> > purpose.
>
> A special camera isn't always available.
> > they have, and now they've been replaced by something better, a bayer
> > sensor.
>
> The Bayer sensor is cheaper, not better. A three-sensor camera would provide
> much better images with much higher color and luminance resolution.
but impossible even in controlled lab conditions, and the colour
resolution is already higher than what you can see so more is a waste.
nospam
Apr 8, 2012, 11:37:04 AM4/8/12
to
In article <MPG.29eb7384...@news.supernews.com>, Alfred Molon
defects. did you do that with film and analyze each individual grain?
> Besides, the Bayer interpolation impacts the overall image resolution,
> not just the colour one.
only slightly.
> But again, it's pointless to conduct this discussion with you, because
> you refuse to accept simple facts, and routinely go into a "no-no-no"
> denial mode. Maybe I should filter out your posts and save my time on
> other issues.
it's you who won't accept basic facts about image processing and human
vision.
<alfred...@yahoo.com> wrote:
> > not really, but bayer already has higher colour resolution than the eye
> > can resolve so there's no issue if it does reduce it.
>
> Irrelevant, because if you enlarge the image enough the eye can very
> well appreciate the colour details.
who does that? if you are going to look for defects, you will find
> > not really, but bayer already has higher colour resolution than the eye
> > can resolve so there's no issue if it does reduce it.
>
> Irrelevant, because if you enlarge the image enough the eye can very
> well appreciate the colour details.
defects. did you do that with film and analyze each individual grain?
> Besides, the Bayer interpolation impacts the overall image resolution,
> not just the colour one.
> But again, it's pointless to conduct this discussion with you, because
> you refuse to accept simple facts, and routinely go into a "no-no-no"
> denial mode. Maybe I should filter out your posts and save my time on
> other issues.
vision.
Alfred Molon
Apr 8, 2012, 3:09:18 PM4/8/12
to
In article <jlrmp3$stj$1...@dont-email.me>, David J Taylor says...
instance) only 6MP?
If a camera has 24MP, it must have 24 million dots of detail, if not the
camera may have 24 MP nominal resolution, but an effective resolution
which is lower.
> What point is there in presenting a final image of greater resolution than
> the eye can perceive? In engineering terms, it is a waste.
Then why would you need a 24MP camera if the eye can appreciate (for
> the eye can perceive? In engineering terms, it is a waste.
instance) only 6MP?
If a camera has 24MP, it must have 24 million dots of detail, if not the
camera may have 24 MP nominal resolution, but an effective resolution
which is lower.
Alfred Molon
Apr 8, 2012, 3:13:39 PM4/8/12
to
In article <jlruls$uh5$1...@dont-email.me>, David J Taylor says...
rotating disk with three colour filters, eleminating all potential
alignment issues.
> Alignment issues and lens issues make such an arrangement either very
> costly or physically impossible at high resolutions,
Irrelevant point. You could for instance have a B&W sensor and a
> costly or physically impossible at high resolutions,
rotating disk with three colour filters, eleminating all potential
alignment issues.
Alfred Molon
Apr 8, 2012, 3:18:44 PM4/8/12
to
In article <NZfgr.29381$pc1....@newsfe11.iad>, Martin Brown says...
pixels in real world scenes. An example would be a meadow with flowers,
or colourful clothes.
Another big issue is the difference between colour and luminance
resolution, which makes it impossible to properly dimension the AA
filter. Dimension it for the colour and you cut off a lot of resolution,
dimension it for the luminance and you end up with moiré.
But frankly, I'm tired of repeating these points all the time to people
who don't want to hear them or understand them. One day all sensors will
be full-colour and this meaningless discussion about Bayer will cease.
> I can't think of any other real world situations where the
> luminance detail is compromised by using Bayer filter mask.
Believe it or not, there can be changes in colour between adjacent
> luminance detail is compromised by using Bayer filter mask.
pixels in real world scenes. An example would be a meadow with flowers,
or colourful clothes.
Another big issue is the difference between colour and luminance
resolution, which makes it impossible to properly dimension the AA
filter. Dimension it for the colour and you cut off a lot of resolution,
dimension it for the luminance and you end up with moiré.
But frankly, I'm tired of repeating these points all the time to people
who don't want to hear them or understand them. One day all sensors will
be full-colour and this meaningless discussion about Bayer will cease.
Floyd L. Davidson
Apr 8, 2012, 3:42:23 PM4/8/12
to
Alfred Molon <alfred...@yahoo.com> wrote:
>In article <NZfgr.29381$pc1....@newsfe11.iad>, Martin Brown says...
>> I can't think of any other real world situations where the
>> luminance detail is compromised by using Bayer filter mask.
>
>Believe it or not, there can be changes in colour between adjacent
>pixels in real world scenes. An example would be a meadow with flowers,
>or colourful clothes.
Which is why higher resolution sensors are beneficial.
>In article <NZfgr.29381$pc1....@newsfe11.iad>, Martin Brown says...
>> I can't think of any other real world situations where the
>> luminance detail is compromised by using Bayer filter mask.
>
>Believe it or not, there can be changes in colour between adjacent
>pixels in real world scenes. An example would be a meadow with flowers,
>or colourful clothes.
>Another big issue is the difference between colour and luminance
>resolution, which makes it impossible to properly dimension the AA
>filter. Dimension it for the colour and you cut off a lot of resolution,
>dimension it for the luminance and you end up with moiré.
>But frankly, I'm tired of repeating these points all the time to people
>who don't want to hear them or understand them. One day all sensors will
>be full-colour and this meaningless discussion about Bayer will cease.
the pixel density will increase past the point where an AA
filter is actually needed. That is going to happen in the
relatively near future, and is probably somewhere between 175
and 250 MP.
The distinction of significance is that to make a tri-color
sensor more efficient than a Bayer filtered sensor will require
approximate three times the pixel density, or perhaps 600
million or so sensor locations on a 36x24mm sensor. It
certainly is possible that in the distant future that might well
be a viable technology, but it is also very likely that a 200 MP
Bayer sensor is always going to be less expensive, and just as
effective.
--
Floyd L. Davidson <http://www.apaflo.com/>
Ukpeagvik (Barrow, Alaska) fl...@apaflo.com
nospam
Apr 8, 2012, 4:09:35 PM4/8/12
to
In article <MPG.29ec01caa...@news.supernews.com>, Alfred
people want to pay, it's not going to happen, especially if the
difference can't be seen.
> You could for instance have a B&W sensor and a
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
unless the subject moves.
Molon <alfred...@yahoo.com> wrote:
> > Alignment issues and lens issues make such an arrangement either very
> > costly or physically impossible at high resolutions,
>
> Irrelevant point.
it's not irrelevant at all. if it's impossible to build at a price
> > Alignment issues and lens issues make such an arrangement either very
> > costly or physically impossible at high resolutions,
>
> Irrelevant point.
people want to pay, it's not going to happen, especially if the
difference can't be seen.
> You could for instance have a B&W sensor and a
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
nospam
Apr 8, 2012, 4:09:38 PM4/8/12
to
In article <MPG.29ec00c7b...@news.supernews.com>, Alfred
an 8x10 print at 300 ppi (considered tack sharp), only needs 2400x3000
pixels, or 7.2 megapixels. 24 mp is overkill. even an 11x14 print
doesn't need 24 mp.
plus, most people don't print anymore, they upload to facebook or
flickr or some other online gallery, or they carry their photo album on
their smartphone or tablet. you certainly don't need 24 mp for that.
> If a camera has 24MP, it must have 24 million dots of detail, if not the
> camera may have 24 MP nominal resolution, but an effective resolution
> which is lower.
if a camera has 24 mp, it has 24 million pixels. that says nothing
about how much detail will be in any given photo.
Molon <alfred...@yahoo.com> wrote:
> > What point is there in presenting a final image of greater resolution than
> > the eye can perceive? In engineering terms, it is a waste.
>
> Then why would you need a 24MP camera if the eye can appreciate (for
> instance) only 6MP?
you don't.
> > What point is there in presenting a final image of greater resolution than
> > the eye can perceive? In engineering terms, it is a waste.
>
> Then why would you need a 24MP camera if the eye can appreciate (for
> instance) only 6MP?
an 8x10 print at 300 ppi (considered tack sharp), only needs 2400x3000
pixels, or 7.2 megapixels. 24 mp is overkill. even an 11x14 print
doesn't need 24 mp.
plus, most people don't print anymore, they upload to facebook or
flickr or some other online gallery, or they carry their photo album on
their smartphone or tablet. you certainly don't need 24 mp for that.
> If a camera has 24MP, it must have 24 million dots of detail, if not the
> camera may have 24 MP nominal resolution, but an effective resolution
> which is lower.
about how much detail will be in any given photo.
nospam
Apr 8, 2012, 4:09:39 PM4/8/12
to
In article <MPG.29ec02fb8...@news.supernews.com>, Alfred
resolution than you can see.
> Another big issue is the difference between colour and luminance
> resolution, which makes it impossible to properly dimension the AA
> filter. Dimension it for the colour and you cut off a lot of resolution,
> dimension it for the luminance and you end up with moiré.
everything has tradeoffs. meanwhile, millions of incredibly sharp
photos are being taken with bayer sensors and aa filters.
> But frankly, I'm tired of repeating these points all the time to people
> who don't want to hear them or understand them.
did it ever occur to you that the reason so many people keep trying to
explain this to you is because *you* don't understand it?
> One day all sensors will
> be full-colour and this meaningless discussion about Bayer will cease.
you know this how?
Molon <alfred...@yahoo.com> wrote:
> > I can't think of any other real world situations where the
> > luminance detail is compromised by using Bayer filter mask.
>
> Believe it or not, there can be changes in colour between adjacent
> pixels in real world scenes. An example would be a meadow with flowers,
> or colourful clothes.
believe it or not, bayer can capture that, and at a much higher chroma
> > I can't think of any other real world situations where the
> > luminance detail is compromised by using Bayer filter mask.
>
> Believe it or not, there can be changes in colour between adjacent
> pixels in real world scenes. An example would be a meadow with flowers,
> or colourful clothes.
resolution than you can see.
> Another big issue is the difference between colour and luminance
> resolution, which makes it impossible to properly dimension the AA
> filter. Dimension it for the colour and you cut off a lot of resolution,
> dimension it for the luminance and you end up with moiré.
photos are being taken with bayer sensors and aa filters.
> But frankly, I'm tired of repeating these points all the time to people
> who don't want to hear them or understand them.
explain this to you is because *you* don't understand it?
> One day all sensors will
> be full-colour and this meaningless discussion about Bayer will cease.
Mxsmanic
Apr 8, 2012, 6:44:09 PM4/8/12
to
nospam writes:
> In article <MPG.29ec00c7b...@news.supernews.com>, Alfred
> Molon <alfred...@yahoo.com> wrote:
>
> > Then why would you need a 24MP camera if the eye can appreciate (for
> > instance) only 6MP?
>
> you don't.
>
> an 8x10 print at 300 ppi (considered tack sharp), only needs 2400x3000
> pixels, or 7.2 megapixels. 24 mp is overkill. even an 11x14 print
> doesn't need 24 mp.
Not so.
> In article <MPG.29ec00c7b...@news.supernews.com>, Alfred
> Molon <alfred...@yahoo.com> wrote:
>
> > Then why would you need a 24MP camera if the eye can appreciate (for
> > instance) only 6MP?
>
> you don't.
>
> an 8x10 print at 300 ppi (considered tack sharp), only needs 2400x3000
> pixels, or 7.2 megapixels. 24 mp is overkill. even an 11x14 print
> doesn't need 24 mp.
Six million pixels is roughly what you can scan within a photo that is held at
the "standard" distance from your eyes, which is roughly equvalent to the
diagonal of the image. However, the number of pixels you need in a digital
photo depends on the viewing distance. The shorter the viewing distance, the
greater the number of pixels the image must contain in order to satisfy the
visual acuity of the eyes.
A more accurate estimate can be made by dividing your viewing distance by
6875. This gives the maximum pixel size that the image should contain. So if
you are viewing an image from 3 feet away, the pixels in the image must be no
larger than 3 feet / 6875 = 0.009 inch, or about 190 pixels per inch.
By extension, if you have a print that will be viewed from three feet away,
and it measures 2.5 feet by 1.5 feet, it will need at least 5700 x 3420
pîxels, or 19.5 megapixels, to match the visual actuity of a person with
normal vision under good viewing conditions. If the print is smaller, or if
it's further away, it doesn't need as many pixels. If it's closer or larger,
it needs more.
All of this reduces to the fact that a normal human retina can resolve details
as small as 30 seconds of arc in static images. Three hundred and sixty
degrees, times 60 minutes, times 60 seconds, divided by 30, divided by two,
divided by pi, equals 6875.
Mxsmanic
Apr 8, 2012, 6:46:27 PM4/8/12
to
nospam writes:
> believe it or not, bayer can capture that, and at a much higher chroma
> resolution than you can see.
False. The resolution depends mainly on the pixel density, not on the color
> believe it or not, bayer can capture that, and at a much higher chroma
> resolution than you can see.
pattern used. The pattern of color filters affects the the ratio of
chrominance to luminance, not the absolute resolution.
> everything has tradeoffs. meanwhile, millions of incredibly sharp
> photos are being taken with bayer sensors and aa filters.
a greater number of pixels on sensors, not from the color filter patterns that
they use.
> you know this how?
It happened to video.
Mxsmanic
Apr 8, 2012, 6:48:06 PM4/8/12
to
David J Taylor writes:
> Alignment issues and lens issues make such an arrangement either very
> costly or physically impossible at high resolutions, and the higher colour
> resolution isn't needed for human vision.
Incorrect on both points. Precise alignment isn't needed, and the higher color
> Alignment issues and lens issues make such an arrangement either very
> costly or physically impossible at high resolutions, and the higher colour
> resolution isn't needed for human vision.
resolution is just as useful as higher luminance resolution. And all photos
are not necessarily intended just for viewing by human eyes.
Mxsmanic
Apr 8, 2012, 6:48:44 PM4/8/12
to
Alfred Molon writes:
> Irrelevant point. You could for instance have a B&W sensor and a
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
Alignment isn't a problem to begin with. And using a rotating filter would
> Irrelevant point. You could for instance have a B&W sensor and a
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
introduce color aberrations in moving subjects.
Mxsmanic
Apr 8, 2012, 6:50:23 PM4/8/12
to
nospam writes:
> i could and if i wanted to do that, i'd get the appropriate equipment
> and software.
What equipment would that be, if not a camera?
> i could and if i wanted to do that, i'd get the appropriate equipment
> and software.
> except that maintaining alignment with high resolution sensors is all
You don't need to maintain alignment. You might not even want them aligned.
> ... and the colour resolution is already higher than what you can
> see so more is a waste.
The color resolution depends on the size of the image and the number of
pixels. It's not possible to say that any specific method of color capture
exceeds the capacity of human vision.
nospam
Apr 8, 2012, 6:56:29 PM4/8/12
to
In article <ol44o79ujk0d6h5lb...@4ax.com>, Mxsmanic
> Six million pixels is roughly what you can scan within a photo that is held at
> the "standard" distance from your eyes, which is roughly equvalent to the
> diagonal of the image. However, the number of pixels you need in a digital
> photo depends on the viewing distance. The shorter the viewing distance, the
> greater the number of pixels the image must contain in order to satisfy the
> visual acuity of the eyes.
obviously if you are going to plant your nose an inch from the print
you are going to see pixels and/or printer dots.
at normal viewing distances, 300 ppi is all you need, which means 8
megapixels is sufficient for an 8x10.
<mxsm...@gmail.com> wrote:
> > > Then why would you need a 24MP camera if the eye can appreciate (for
> > > instance) only 6MP?
> >
> > you don't.
> >
> > an 8x10 print at 300 ppi (considered tack sharp), only needs 2400x3000
> > pixels, or 7.2 megapixels. 24 mp is overkill. even an 11x14 print
> > doesn't need 24 mp.
>
> Not so.
it is so.
> > > Then why would you need a 24MP camera if the eye can appreciate (for
> > > instance) only 6MP?
> >
> > you don't.
> >
> > an 8x10 print at 300 ppi (considered tack sharp), only needs 2400x3000
> > pixels, or 7.2 megapixels. 24 mp is overkill. even an 11x14 print
> > doesn't need 24 mp.
>
> Not so.
> Six million pixels is roughly what you can scan within a photo that is held at
> the "standard" distance from your eyes, which is roughly equvalent to the
> diagonal of the image. However, the number of pixels you need in a digital
> photo depends on the viewing distance. The shorter the viewing distance, the
> greater the number of pixels the image must contain in order to satisfy the
> visual acuity of the eyes.
you are going to see pixels and/or printer dots.
at normal viewing distances, 300 ppi is all you need, which means 8
megapixels is sufficient for an 8x10.
nospam
Apr 8, 2012, 6:56:30 PM4/8/12
to
In article <6654o7t8bngq95rul...@4ax.com>, Mxsmanic
> > you know this how?
>
> It happened to video.
video is moving to 1 chip bayer sensors. so much for that idea.
<mxsm...@gmail.com> wrote:
> > believe it or not, bayer can capture that, and at a much higher chroma
> > resolution than you can see.
>
> False. The resolution depends mainly on the pixel density, not on the color
> pattern used. The pattern of color filters affects the the ratio of
> chrominance to luminance, not the absolute resolution.
and it's still ~5x higher than what your eye can resolve.
> > believe it or not, bayer can capture that, and at a much higher chroma
> > resolution than you can see.
>
> False. The resolution depends mainly on the pixel density, not on the color
> pattern used. The pattern of color filters affects the the ratio of
> chrominance to luminance, not the absolute resolution.
> > you know this how?
>
> It happened to video.
nospam
Apr 8, 2012, 6:56:31 PM4/8/12
to
In article <ib54o7p46mdcfsame...@4ax.com>, Mxsmanic
> and the higher color
> resolution is just as useful as higher luminance resolution.
not to humans it isn't.
> And all photos
> are not necessarily intended just for viewing by human eyes.
almost all are, and the few photos that aren't can use specialized
sensors designed explicitly for that purpose.
<mxsm...@gmail.com> wrote:
> David J Taylor writes:
>
> > Alignment issues and lens issues make such an arrangement either very
> > costly or physically impossible at high resolutions, and the higher colour
> > resolution isn't needed for human vision.
>
> Incorrect on both points. Precise alignment isn't needed,
yes it is.
> David J Taylor writes:
>
> > Alignment issues and lens issues make such an arrangement either very
> > costly or physically impossible at high resolutions, and the higher colour
> > resolution isn't needed for human vision.
>
> Incorrect on both points. Precise alignment isn't needed,
> and the higher color
> resolution is just as useful as higher luminance resolution.
> And all photos
> are not necessarily intended just for viewing by human eyes.
sensors designed explicitly for that purpose.
nospam
Apr 8, 2012, 7:01:55 PM4/8/12
to
In article <qe54o7tsvov681t4m...@4ax.com>, Mxsmanic
the vast majority of photos.
> > except that maintaining alignment with high resolution sensors is all
> > but impossible even in controlled lab conditions ...
>
> You don't need to maintain alignment. You might not even want them aligned.
you definitely need them aligned or you will have artifacts.
> > ... and the colour resolution is already higher than what you can
> > see so more is a waste.
>
> The color resolution depends on the size of the image and the number of
> pixels. It's not possible to say that any specific method of color capture
> exceeds the capacity of human vision.
it's very possible if you understand how bayer works.
if you don't understand how bayer works, then it won't be possible.
<mxsm...@gmail.com> wrote:
> > i could and if i wanted to do that, i'd get the appropriate equipment
> > and software.
>
> What equipment would that be, if not a camera?
a camera designed for 1:1 luma/chroma resolution. that's not needed for
> > i could and if i wanted to do that, i'd get the appropriate equipment
> > and software.
>
> What equipment would that be, if not a camera?
the vast majority of photos.
> > except that maintaining alignment with high resolution sensors is all
> > but impossible even in controlled lab conditions ...
>
> You don't need to maintain alignment. You might not even want them aligned.
> > ... and the colour resolution is already higher than what you can
> > see so more is a waste.
>
> The color resolution depends on the size of the image and the number of
> pixels. It's not possible to say that any specific method of color capture
> exceeds the capacity of human vision.
if you don't understand how bayer works, then it won't be possible.
Peter Irwin
Apr 8, 2012, 8:31:28 PM4/8/12
to
nospam <nos...@nospam.invalid> wrote:
>
> obviously if you are going to plant your nose an inch from the print
> you are going to see pixels and/or printer dots.
>
> at normal viewing distances, 300 ppi is all you need, which means 8
> megapixels is sufficient for an 8x10.
If you take the resolution of the eye as 50 cycles per degree,
>
> obviously if you are going to plant your nose an inch from the print
> you are going to see pixels and/or printer dots.
>
> at normal viewing distances, 300 ppi is all you need, which means 8
> megapixels is sufficient for an 8x10.
and you have the optimistic view that 300 pixels represents
150 cycles then your minimum viewing distance for perfect
resolution is 19.1 inches.
50 x 57.3
---------
150
If you assume the conventional 30 cycles per degree,
then your minimum viewing distance is 11 1/2 inches.
Neither of these are in the nose to the paper category.
Peter.
Mxsmanic
Apr 8, 2012, 10:30:45 PM4/8/12
to
nospam writes:
> at normal viewing distances, 300 ppi is all you need, which means 8
> megapixels is sufficient for an 8x10.
This is incorrect. The number of pixels per inch that are required varies with
> at normal viewing distances, 300 ppi is all you need, which means 8
> megapixels is sufficient for an 8x10.
the viewing distance, and thus is not fixed at 300 ppi.
For an 8x10 held at the diagonal distance of 13 inches, the number of pixels
required to match the eye's visual limits is 529 ppi, not 300 ppi.
Mxsmanic
Apr 8, 2012, 10:32:04 PM4/8/12
to
Peter Irwin writes:
> If you assume the conventional 30 cycles per degree,
> then your minimum viewing distance is 11 1/2 inches.
The actual resolving power is 30 arc-seconds, or one 1/120 of a degree.
> If you assume the conventional 30 cycles per degree,
> then your minimum viewing distance is 11 1/2 inches.
Mxsmanic
Apr 8, 2012, 10:32:47 PM4/8/12
to
nospam writes:
> and it's still ~5x higher than what your eye can resolve.
Since it is a ratio, it is dimensionless.
> and it's still ~5x higher than what your eye can resolve.
> video is moving to 1 chip bayer sensors. so much for that idea.
Mxsmanic
Apr 8, 2012, 10:33:38 PM4/8/12
to
nospam writes:
> yes it is.
No, it is not. In fact, it may not even be desirable.
> almost all are, and the few photos that aren't can use specialized
> sensors designed explicitly for that purpose.
Forensic analysis doesn't have that option.
> yes it is.
No, it is not. In fact, it may not even be desirable.
> almost all are, and the few photos that aren't can use specialized
> sensors designed explicitly for that purpose.
Mxsmanic
Apr 8, 2012, 10:35:14 PM4/8/12
to
nospam writes:
> a camera designed for 1:1 luma/chroma resolution
Which cameras would those be?
> a camera designed for 1:1 luma/chroma resolution
> you definitely need them aligned or you will have artifacts.
original image, and artifacts can be produced whether they are aligned or not.
> it's very possible if you understand how bayer works.
nospam
Apr 8, 2012, 11:09:54 PM4/8/12
to
In article <oai4o79n52o2chom9...@4ax.com>, Mxsmanic
<mxsm...@gmail.com> wrote:
> > at normal viewing distances, 300 ppi is all you need, which means 8
> > megapixels is sufficient for an 8x10.
>
> This is incorrect. The number of pixels per inch that are required varies with
> the viewing distance, and thus is not fixed at 300 ppi.
as i said, normal viewing distances.
> > at normal viewing distances, 300 ppi is all you need, which means 8
> > megapixels is sufficient for an 8x10.
>
> This is incorrect. The number of pixels per inch that are required varies with
> the viewing distance, and thus is not fixed at 300 ppi.
nospam
Apr 8, 2012, 11:09:58 PM4/8/12
to
In article <khi4o7picvfo31cne...@4ax.com>, Mxsmanic
chip, as are all the dslrs that do video.
<mxsm...@gmail.com> wrote:
> > video is moving to 1 chip bayer sensors. so much for that idea.
>
> We shall see.
not shall. it's happening *now*. video cameras such as red are single
> > video is moving to 1 chip bayer sensors. so much for that idea.
>
> We shall see.
chip, as are all the dslrs that do video.
nospam
Apr 8, 2012, 11:10:00 PM4/8/12
to
In article <lii4o7lsp6qs760kl...@4ax.com>, Mxsmanic
designed for that purpose.
<mxsm...@gmail.com> wrote:
> > almost all are, and the few photos that aren't can use specialized
> > sensors designed explicitly for that purpose.
>
> Forensic analysis doesn't have that option.
forensic analysis is a small niche. they can get specialized equipment
> > almost all are, and the few photos that aren't can use specialized
> > sensors designed explicitly for that purpose.
>
> Forensic analysis doesn't have that option.
designed for that purpose.
nospam
Apr 8, 2012, 11:10:03 PM4/8/12
to
In article <dki4o71tddksl72v1...@4ax.com>, Mxsmanic
<mxsm...@gmail.com> wrote:
> > it's very possible if you understand how bayer works.
>
> I do understand how it works, but it looks that I may be in the minority.
based on what you've posted, you definitely do not.
> > it's very possible if you understand how bayer works.
>
> I do understand how it works, but it looks that I may be in the minority.
TheRealSteve
Apr 8, 2012, 11:12:30 PM4/8/12
to
On Sun, 8 Apr 2012 12:57:05 +0100, "David J Taylor"
<david-...@blueyonder.co.uk.invalid> wrote:
>"Mxsmanic" <mxsm...@gmail.com> wrote in message
>news:q4s2o71njvo3ue4v6...@4ax.com...
>> nospam writes:
>>
>>> wrong. an aa filter prevents the moire from happening in the first
>>> place, so the real detail won't need to be altered.
>>
>> The AA filter works by removing the real detail.
>
>The AA filter works by removing the detail which cannot be accurately
>recorded by a sensor made up of discrete pixels (at regular intervals).
>The lack of such filtering will lead to artefacts such as steps on near
>horizontal or near vertical edges, and false-colours on finely textured
>material, and these artefacts cannot be removed in post-processing.
Wrong. Since a perfect AA filter exists only mathematically, real ones
do remove real detail and do not remove all alias artifacts.
nospam
Apr 8, 2012, 11:54:21 PM4/8/12
to
In article <2pk4o71gghusloihv...@4ax.com>, TheRealSteve
> Since a perfect AA filter exists only mathematically, real ones
> do remove real detail and do not remove all alias artifacts.
yes, we live in an imperfect world.
<st...@example.com> wrote:
> >The AA filter works by removing the detail which cannot be accurately
> >recorded by a sensor made up of discrete pixels (at regular intervals).
> >The lack of such filtering will lead to artefacts such as steps on near
> >horizontal or near vertical edges, and false-colours on finely textured
> >material, and these artefacts cannot be removed in post-processing.
>
> Wrong.
it's not wrong.
> >The AA filter works by removing the detail which cannot be accurately
> >recorded by a sensor made up of discrete pixels (at regular intervals).
> >The lack of such filtering will lead to artefacts such as steps on near
> >horizontal or near vertical edges, and false-colours on finely textured
> >material, and these artefacts cannot be removed in post-processing.
>
> Wrong.
> Since a perfect AA filter exists only mathematically, real ones
> do remove real detail and do not remove all alias artifacts.
David J Taylor
Apr 9, 2012, 1:57:04 AM4/9/12
to
"Alfred Molon" <alfred...@yahoo.com> wrote in message
news:MPG.29ec01caa...@news.supernews.com...
> In article <jlruls$uh5$1...@dont-email.me>, David J Taylor says...
>
> Alfred Molon
.. and how many digital cameras have a colour disk these days? Not quite
the handy tool for street photography, is it? We're not talking space
missions here!
Your point is irrelevant in the context of our discussion.
David
news:MPG.29ec01caa...@news.supernews.com...
> In article <jlruls$uh5$1...@dont-email.me>, David J Taylor says...
>> Alignment issues and lens issues make such an arrangement either very
>> costly or physically impossible at high resolutions,
>
>> costly or physically impossible at high resolutions,
>
> Irrelevant point. You could for instance have a B&W sensor and a
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
> --
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
>
> Alfred Molon
.. and how many digital cameras have a colour disk these days? Not quite
the handy tool for street photography, is it? We're not talking space
missions here!
Your point is irrelevant in the context of our discussion.
David
David J Taylor
Apr 9, 2012, 2:09:53 AM4/9/12
to
"Mxsmanic" <mxsm...@gmail.com> wrote in message
news:ib54o7p46mdcfsame...@4ax.com...
requirements, but we are talking here about "photography for human
viewing". The higher colour resolution cannot be perceived by the eye,
and if you want your red, green and blue images to be shifted and rotated
relative to each other then I think I will be giving your images a pass!
At the moment, the Bayer sensor is the best for most people. No-one is
forcing you to buy one, though, as alternatives are available.
David
Martin Brown
Apr 9, 2012, 3:37:11 AM4/9/12
to
On 08/04/2012 20:13, Alfred Molon wrote:
> In article<jlruls$uh5$1...@dont-email.me>, David J Taylor says...
> In article<jlruls$uh5$1...@dont-email.me>, David J Taylor says...
>> Alignment issues and lens issues make such an arrangement either very
>> costly or physically impossible at high resolutions,
>
>> costly or physically impossible at high resolutions,
>
> Irrelevant point. You could for instance have a B&W sensor and a
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
In space, but not in time so you get massive and offensive colour
> rotating disk with three colour filters, eleminating all potential
> alignment issues.
fringing on every moving object. That will look really good on TV!
Tricolour imaging is used in astronomy but it is a serious problem if
the target is a comet and moves between successive exposures.
The whole point of single shot digital colour is to collect all the
information needed to provide an image that matches the capabilities of
the human eye. Some of the "problems" that you complain about are not
the fault of Bayer masks, but a defect in the most common JPEG decoders.
And you have to be pixel peeping to see any of them.
--
Regards,
Martin Brown
Martin Brown
Apr 9, 2012, 4:15:54 AM4/9/12
to
On 08/04/2012 20:18, Alfred Molon wrote:
> In article<NZfgr.29381$pc1....@newsfe11.iad>, Martin Brown says...
blue tapered stripes to actually cause significant problems for a Bayer
mosaic - and even then with the antialias filter the results are not
bad. The only time I have seen significant chroma sampling problems was
photographing in pure red H-alpha light at 656nm. And even then most of
the perceived problems were due to a fault in the common JPEG decoder.
(and also the minor red leaks in the nominally green and blue filters)
> Another big issue is the difference between colour and luminance
> resolution, which makes it impossible to properly dimension the AA
> filter. Dimension it for the colour and you cut off a lot of resolution,
> dimension it for the luminance and you end up with moiré.
Not true. The AA filter is scaled to just slightly more than is strictly
needed for luminance which is also enough to introduce some linear
dependancy between adjacent pixels. The chroma reconstruction from a
Bayer matrix is surprisingly accurate for everything except a handful of
pathological test charts designed to break it.
image reconstruction from sparse data. It is clear that you will never
learn and have been suckered by the Foveon PR scam.
There is no point at all in recording chroma information that the human
eye cannot see in ordinary photography.
--
Regards,
Martin Brown
> In article<NZfgr.29381$pc1....@newsfe11.iad>, Martin Brown says...
>> I can't think of any other real world situations where the
>> luminance detail is compromised by using Bayer filter mask.
>
> Believe it or not, there can be changes in colour between adjacent
> pixels in real world scenes. An example would be a meadow with flowers,
> or colourful clothes.
Indeed. But it takes a very special form of pattern involving red and
>> luminance detail is compromised by using Bayer filter mask.
>
> Believe it or not, there can be changes in colour between adjacent
> pixels in real world scenes. An example would be a meadow with flowers,
> or colourful clothes.
blue tapered stripes to actually cause significant problems for a Bayer
mosaic - and even then with the antialias filter the results are not
bad. The only time I have seen significant chroma sampling problems was
photographing in pure red H-alpha light at 656nm. And even then most of
the perceived problems were due to a fault in the common JPEG decoder.
(and also the minor red leaks in the nominally green and blue filters)
> Another big issue is the difference between colour and luminance
> resolution, which makes it impossible to properly dimension the AA
> filter. Dimension it for the colour and you cut off a lot of resolution,
> dimension it for the luminance and you end up with moiré.
needed for luminance which is also enough to introduce some linear
dependancy between adjacent pixels. The chroma reconstruction from a
Bayer matrix is surprisingly accurate for everything except a handful of
pathological test charts designed to break it.
>
> But frankly, I'm tired of repeating these points all the time to people
> who don't want to hear them or understand them. One day all sensors will
> But frankly, I'm tired of repeating these points all the time to people
> be full-colour and this meaningless discussion about Bayer will cease.
The problem here is your lack of understanding of the mathematics of
image reconstruction from sparse data. It is clear that you will never
learn and have been suckered by the Foveon PR scam.
There is no point at all in recording chroma information that the human
eye cannot see in ordinary photography.
--
Regards,
Martin Brown
C. Neil Ellwood
Apr 9, 2012, 5:56:41 AM4/9/12
to
On Sun, 08 Apr 2012 13:08:32 +0200, Mxsmanic wrote:
> nospam writes:
>
>> what else would i want do with photos than look at them?
>
> You could subject them to computer analysis.
>
>> if you mean image analysis, then there are special cameras for that
>> purpose.
>
> A special camera isn't always available.
>
>> they have, and now they've been replaced by something better, a bayer
>> sensor.
>
> The Bayer sensor is cheaper, not better. A three-sensor camera would
> provide much better images with much higher color and luminance
> resolution.
You could also use one of the old cameras designed for tri-chrome carbro
photography which gave very good results. Not really as convenient as a
dslr and very cumbersome (especially in the whole plate size.
--
Neil
Reverse ‘a’ and ‘r’
Remove ‘l’ to get address.
> nospam writes:
>
>> what else would i want do with photos than look at them?
>
> You could subject them to computer analysis.
>
>> if you mean image analysis, then there are special cameras for that
>> purpose.
>
> A special camera isn't always available.
>
>> they have, and now they've been replaced by something better, a bayer
>> sensor.
>
> The Bayer sensor is cheaper, not better. A three-sensor camera would
> provide much better images with much higher color and luminance
> resolution.
You could also use one of the old cameras designed for tri-chrome carbro
photography which gave very good results. Not really as convenient as a
dslr and very cumbersome (especially in the whole plate size.
--
Neil
Reverse ‘a’ and ‘r’
Remove ‘l’ to get address.
Alfred Molon
Apr 9, 2012, 7:44:22 AM4/9/12
to
In article <jlttng$k74$1...@dont-email.me>, David J Taylor says...
there were. Although I think the Mars rover used such a monochrome
camera with rotating colour filters.
--
Alfred Molon
------------------------------
Olympus E-series DSLRs and micro 4/3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site
> .. and how many digital cameras have a colour disk these days? Not quite
> the handy tool for street photography, is it? We're not talking space
> missions here!
Of course there are no such cameras now on the market. I never claimed
> the handy tool for street photography, is it? We're not talking space
> missions here!
there were. Although I think the Mars rover used such a monochrome
camera with rotating colour filters.
--
Alfred Molon
------------------------------
Olympus E-series DSLRs and micro 4/3 forum at
http://tech.groups.yahoo.com/group/MyOlympus/
http://myolympus.org/ photo sharing site
Alfred Molon
Apr 9, 2012, 7:52:46 AM4/9/12
to
In article <TUwgr.7$%b4...@newsfe21.iad>, Martin Brown says...
> Not true. The AA filter is scaled to just slightly more than is strictly
> needed for luminance which is also enough to introduce some linear
> dependancy between adjacent pixels. The chroma reconstruction from a
> Bayer matrix is surprisingly accurate for everything except a handful of
> pathological test charts designed to break it.
Wrong, because in a bayer sensor the luminance resolution is lower than
the colour resolution. It's impossible to correctly dimension the AA
filter - you have to dimension either for the luminance or for the
colour resolution.
Your claim that "chroma reconstruction from a Bayer matrix is
surprisingly accurate" is wrong as well.
> The problem here is your lack of understanding of the mathematics of
> image reconstruction from sparse data. It is clear that you will never
> learn and have been suckered by the Foveon PR scam.
Your problem is your lack of basic understanding of data sampling
principles. I suggest that you have a look at the Nyquist theorem. You
would probably understand that data needs to be sampled at twice the
maximum frequency input.
> There is no point at all in recording chroma information that the human
> eye cannot see in ordinary photography.
Irrelevant, because you have no way to know what size the final print
will have. It may be big enough for a human to be able to appreciate the
fine detail.
> Not true. The AA filter is scaled to just slightly more than is strictly
> needed for luminance which is also enough to introduce some linear
> dependancy between adjacent pixels. The chroma reconstruction from a
> Bayer matrix is surprisingly accurate for everything except a handful of
> pathological test charts designed to break it.
the colour resolution. It's impossible to correctly dimension the AA
filter - you have to dimension either for the luminance or for the
colour resolution.
Your claim that "chroma reconstruction from a Bayer matrix is
surprisingly accurate" is wrong as well.
> The problem here is your lack of understanding of the mathematics of
> image reconstruction from sparse data. It is clear that you will never
> learn and have been suckered by the Foveon PR scam.
principles. I suggest that you have a look at the Nyquist theorem. You
would probably understand that data needs to be sampled at twice the
maximum frequency input.
> There is no point at all in recording chroma information that the human
> eye cannot see in ordinary photography.
will have. It may be big enough for a human to be able to appreciate the
fine detail.
Peter Irwin
Apr 9, 2012, 8:08:48 AM4/9/12
to
on an eye chart. The optotypes subtend 5 minutes of arc
at the designated distance and are each five line widths high.
(The square characters are 5x5 pixels on a eye chart, and the 20/20
line is 5 arcminutes high at 20 feet.) For 300 pixels per inch
20/20 corresponds to a viewing distance of 11.46 inches.
The actual limit of the eye is more like 50 cycles per degree
under ideal conditions. This translates to a distance of 19.1
inches for 300 pixels per inch.
Peter.
--
pir...@ktb.net
David J Taylor
Apr 9, 2012, 8:50:19 AM4/9/12
to
"Alfred Molon" <alfred...@yahoo.com> wrote in message
news:MPG.29ecebfbb...@news.supernews.com...
[]
> Wrong, because in a bayer sensor the luminance resolution is lower than
> the colour resolution.
[]
> the colour resolution.
> --
>
> Alfred Molon
Wrong - there are four luminance samples for each colour sample.
Luminance resolution is higher, just like the human eye.
David
Mxsmanic
Apr 9, 2012, 9:11:16 AM4/9/12
to
then they necessarily remove detail that the sensor could otherwise capture,
or they leave detail that it cannot.
Mxsmanic
Apr 9, 2012, 9:17:23 AM4/9/12
to
Peter Irwin writes:
> The conventional 30 cycles per degree is equivalent to 20/20
> on an eye chart.
Then the chart is very generous.
> The conventional 30 cycles per degree is equivalent to 20/20
> on an eye chart.
The 30-arc-second limit is imposed by the dimensions of the cells in the
retina themselves. Admittedly, excellent viewing conditions are required to
reach this limit, but it's best to target this limit if you want images that
are always guaranteed to look sharp.
Vernier acuity or integration of a moving target can raise this, although
again it requires very good or ideal conditions.
It is loading more messages.
0 new messages