VueScan Suggestions
Ed Hamrick
suggestions for improving VueScan. I also asked that
these suggestions not be posted in this newsgroup, but
instead that I would summarize the responses. I've
included these suggestions below, and have categorized
them under "User's Guide", "Performance", "User Interface",
"Features" and "Scanner Specific Features".
I appreciate all the suggestions, and will take them
into account when considering changes to VueScan.
Regards,
Ed Hamrick
User's Guide
------------
2 PDF file of User's Guide with more easy explanations
Improve documentation
Index for User's Guide
More comprehensive documentation, integrated with VueScan
Improve User's Guide
Performance
-----------
2 Delay re-doing preview when cropping adjusted
Speed up sluggish GUI
Further work on preview (better speed and resolution)
Faster processing
User Interface
--------------
Fix UI glitches on Mac OS X
Wizards to help make Color adjustments easier
Simplify user interface
Allow displaying preview at near full-screen
Split-screen mode for before/after correction
Make left window a dockable panel
Display all previews at the same time and allow changes to each
Hotkeys for next/previous frame
Add option to reset all options in current tab
Features
--------
2 Use .ini files from old versions
2 Curve tool
2 Copy feature
Color processing like older versions of VueScan
Fractional size reduction option
Better grain reduction
Fill Flash correction checkbox
Save to PNG format
Automatic frame region
Dark frame subtraction, other noise reduction
More explicit control over exposure time
Support for 3rd party plugins
IT-8 calibration
User defined negative types
Spot meter
Better IR cleaning, especially for shallow scratches
Color saturation control
Batch scan with different settings for each scan
User selectable ICC profiles
Fast thumbnail scan on Nikon
Option for different amounts of sharpening
API for other programs
Photoshop plugin
Output key settings with the files
Scanner Specific Features
-------------------------
2 Calibrate command on Nikon
2 Support for non-standard frame sizes on Nikon scanners
Better frame offset/auto cropping on LS-4000
Auto film strip frame offset on LS-4000
Per-color exposure control on Nikon
Support for Microtek 6800
Fix problem with FS4000 and Infrared Clean
Bram
looked on your website http://www.hamrick.com/ but couldn't find a
"suggestion" link there.
Best regards,
Bram
Ed Hamrick <use...@hamrick.com> schreef in berichtnieuws
ai9eqd$t1a$1...@nntp-m01.news.aol.com...
Pierre Gaston
"Ed Hamrick" <use...@hamrick.com> wrote in message
news:ai9eqd$t1a$1...@nntp-m01.news.aol.com...
> I recently asked that people send me their three top
> suggestions for improving VueScan. I also asked that
> these suggestions not be posted in this newsgroup, but
> instead that I would summarize the responses. I've
> included these suggestions below, and have categorized
> them under "User's Guide", "Performance", "User Interface",
> "Features" and "Scanner Specific Features".
>
> I appreciate all the suggestions, and will take them
> into account when considering changes to VueScan.
>
> Regards,
> Ed Hamrick
Ed:
I noticed that you don't post the address to where checks for the software
can be sent. Why not?
Ed Hamrick
> I noticed that you don't post the address to where checks for the software
> can be sent. Why not?
Go to www.hamrick.com, click "Register", then click
"Register by mail".
Regards,
Ed Hamrick
Bram
Here is a communications problem: we are not talking about registering the
software, I already have doen this a long time ago. I am looking for a place
to submit my suggestions. And today also for the suggested list I saw
yesterday, because now I don't have this list any more.
Best regards,
Bram
Erik Krause
you wrote...
> Here is a communications problem: we are not talking about registering the
> software, I already have doen this a long time ago. I am looking for a place
> to submit my suggestions. And today also for the suggested list I saw
> yesterday, because now I don't have this list any more.
Ed posted this when he asked for suggestions a few days ago:
--
Erik Krause
Erik Krause
you wrote...
> I recently asked that people send me their three top
> suggestions for improving VueScan. I also asked that
> these suggestions not be posted in this newsgroup, but
> instead that I would summarize the responses. I've
> included these suggestions below, and have categorized
> them under "User's Guide", "Performance", "User Interface",
> "Features" and "Scanner Specific Features".
I suppose you posted this here for discussion. So let me comment ist...
> User's Guide
> ------------
> 2 PDF file of User's Guide with more easy explanations
> Improve documentation
> Index for User's Guide
> More comprehensive documentation, integrated with VueScan
> Improve User's Guide
I think most beginners ask for easy to follow step by step
instructions. True?
> Performance
> -----------
> 2 Delay re-doing preview when cropping adjusted
> Speed up sluggish GUI
> Further work on preview (better speed and resolution)
> Faster processing
I personally am happy with the actual state...
> User Interface
> --------------
[...]
> Allow displaying preview at near full-screen
Well, I use two monitors and have extended the window to the left
monitor, so I have near full screen preview ;-)
> Split-screen mode for before/after correction
Good idea. Hotkey toggle would do almost the same...
> Make left window a dockable panel
Very good idea!
> Display all previews at the same time and allow changes to each
> Hotkeys for next/previous frame
> Add option to reset all options in current tab
Good points...
> Features
> --------
> 2 Use .ini files from old versions
> 2 Curve tool
both ok.
> 2 Copy feature
Don't understand...
> Color processing like older versions of VueScan
> Fractional size reduction option
No need. Resampling is better done by a specialized app.
> Better grain reduction
perhaps the mentioned SUSAN filter?
> Fill Flash correction checkbox
what's that?
> Save to PNG format
> Automatic frame region
> Dark frame subtraction, other noise reduction
interesting
[...]
> IT-8 calibration
perhaps interaction with Marti's ICC profiler...
[...]
> Option for different amounts of sharpening
This is far better done in Photoshop...
[...]
> Scanner Specific Features
> -------------------------
[...]
> Per-color exposure control on Nikon
good point
--
Erik Krause
Don
Ed Hamrick wrote:
> I recently asked that people send me their three top
> suggestions for improving VueScan. I also asked that
> these suggestions not be posted in this newsgroup, but
> instead that I would summarize the responses. I've
> included these suggestions below, and have categorized
> them under "User's Guide", "Performance", "User Interface",
> "Features" and "Scanner Specific Features".
I'm surprised that one suggestion wasn't improving the install routine
(ie allowing install dir. to be specified), or did I miss it in the list?
Don
UrbanVoyeur
posts go there to retire.
(2) This is a multi threaded conversation and you are just one of the
voices. Ed was replying to another poster with different question. (HINT:
you are BRAM, not Pierre Gaston)
(3) Your question has been asked and answered before: sup...@hamrick.com
"Bram" <bvd...@planet.nl> wrote in message
news:aibvc6$pt6$1...@reader08.wxs.nl...
Ed Hamrick
> I'm surprised that one suggestion wasn't improving the install routine
> (ie allowing install dir. to be specified), or did I miss it in the list?
There were a lot of things that I was surprised nobody asked for.
That's the main reason I asked people to limit their suggestions
to their top three - it tends to eliminate things that would be
nice to have, but that isn't really very important.
Regards,
Ed Hamrick
Ed Hamrick
> > Improve User's Guide
> I think most beginners ask for easy to follow step by step
> instructions. True?
My main design goal is to make it possible for beginners to
scan without ever looking at the User's Guide. One of the
best user interface design guides has lots of interesting
comments on this:
http://www.joelonsoftware.com/uibook/chapters/fog0000000062.html
> > 2 Copy feature
> Don't understand...
People want to be able to print directly from VueScan -
i.e. use VueScan and a scanner as a copier.
> > Fill Flash correction checkbox
> what's that?
I have no idea - someone suggested it.
Regards,
Ed Hamrick
Erik Krause
you wrote...
> > > 2 Copy feature
> > Don't understand...
>
> People want to be able to print directly from VueScan -
> i.e. use VueScan and a scanner as a copier.
On Windows this could easily be done with vueprint as external viewer
(command line: vuepro32 /print ;-)
For Linux an even better solution could be achieved with external
Viewer: lpr -Plp -#1 -r (more in message Id
<46ac41bb.02050...@posting.google.com>)
For Mac OS something similar should be possible.
--
Erik Krause
Bart van der Wolf
> "Erik Krause" <erik....@gmx.net> wrote:
SNIP
> > > Fill Flash correction checkbox
> > what's that?
>
> I have no idea - someone suggested it.
I don't know for sure either, but perhaps the request was more profound than
the person suggesting it realised. For the theory have a look at:
http://www.cs.huji.ac.il/%7Ewerman/Papers/hdrc.pdf and
for a nice collection of examples: http://wind.winona.msus.edu/~stafford/p/
The idea of luminance gradient leveling can be applied to a single (RAW /
HDR)image, or a composite of differently exposed versions of a scene (like a
really good long exposure function).
It is also somewhat like the Retinex technology from NASA, but I'm more
impressed with the HDR approach.
That would not only fix overexposed flash foregrounds, if that's what the
suggestion was about.... But it is probably a bit complicated to encode
efficiently :-(
Bart
Erik Krause
you wrote...
> > > > Fill Flash correction checkbox
> > > what's that?
> >
> > I have no idea - someone suggested it.
>
> I don't know for sure either, but perhaps the request was more profound than
> the person suggesting it realised. For the theory have a look at:
> http://www.cs.huji.ac.il/%7Ewerman/Papers/hdrc.pdf and
> for a nice collection of examples: http://wind.winona.msus.edu/~stafford/p/
You need multiple different exposed images for that. In a sinb´gle
image where the film is overexposed the best software solution can't
find any details.
> The idea of luminance gradient leveling can be applied to a single (RAW /
> HDR)image, or a composite of differently exposed versions of a scene (like a
> really good long exposure function).
> It is also somewhat like the Retinex technology from NASA, but I'm more
> impressed with the HDR approach.
Yes, it's very interesting! But I think this is more a thing for Helmut
Dersch and his PanoTools than for scanning software. For VueScan I'm
happy with the logarithmic curve option...
> That would not only fix overexposed flash foregrounds, if that's what the
> suggestion was about.... But it is probably a bit complicated to encode
> efficiently :-(
Not only that. Different scanned images are never aligned properly, you
have to align them in Photoshop (etc.) before combining. And how to
deal with moving leafs, clouds or persons and so on? Nothing for a
scanner software...
--
Erik Krause
Bart van der Wolf
"Erik Krause" <erik....@gmx.net> wrote in message
news:MPG.17b5b86fa...@ID-18456.user.dfncis.de...
> Hi, Bart van der Wolf
> You need multiple different exposed images for that. In a sinb´gle
> image where the film is overexposed the best software solution can't
> find any details.
It also works on a single image, it adjusts/levels the rapid changes of
luminance, so more contrast can be given to the more gradual changing areas.
That's the interesting thing for scanning. And it gets better when the
multiple exposures are multiple scans/passes, because it allows correct
exposure of transparent film areas and extra exposure for the dense areas.
SNIP
> > That would not only fix overexposed flash foregrounds, if that's what
the
> > suggestion was about.... But it is probably a bit complicated to encode
> > efficiently :-(
>
> Not only that. Different scanned images are never aligned properly, you
> have to align them in Photoshop (etc.) before combining. And how to
> deal with moving leafs, clouds or persons and so on? Nothing for a
> scanner software...
Correct, if you talk about separate handheld or moving object exposures, but
the better scanners are capable of multi(pass) scanning with good/perfect
registration of the individual scans.
Bart
Emo
Tom Harrison
news:_3039.16098$t%5.79...@newssvr28.news.prodigy.com...
> Do students get a discount?
The software is only $40 -- this is an equal opportunity discount :-)
Tom
Emo
"Tom Harrison" <tNoOmShP...@attbi.com> wrote in message
news:67239.776920$cQ3.118977@sccrnsc01...
> > Do students get a discount?
>
> The software is only $40 -- this is an equal opportunity discount :-)
>
> Tom
Wish I had the equal opportunity funds :(
craig gonter
So far all I have to say is that you get what you paid for.... I consider it
in beta.
Erik Krause
> > You need multiple different exposed images for that. In a sinb´gle
> > image where the film is overexposed the best software solution can't
> > find any details.
>
> It also works on a single image, it adjusts/levels the rapid changes of
> luminance, so more contrast can be given to the more gradual changing areas.
> That's the interesting thing for scanning. And it gets better when the
> multiple exposures are multiple scans/passes, because it allows correct
> exposure of transparent film areas and extra exposure for the dense areas.
But do we need that? Vuescan and my LS40 do a very good job scanning
even the densest areas of Fuji Velvia. Long exposure pass takes the
noise out of very dark parts. Logarithmical curve lowers the contrast
range. Anything else should be done by specialized software.
--
Erik Krause
UrbanVoyeur
customer support is excellent, and it produces great results. It is an
excellent value and is not "beta" in any way.
"craig gonter" <cgo...@capital.net> wrote in message
news:be0baed5.02080...@posting.google.com...
Cliff Norton
> So far all I have to say is that you get what you paid for.... I consider it
> in beta.
I consider this comment to be in beta. When you actually have a point
to convey warranting a 1.0 version, please feel free to repost.
Cliff
Bart van der Wolf
Do we need that? Well certain images do benefit. If you consider that e.g.
an existing light room interior with a window, or a landscape with shadow
and sun patches and sky (reflecting of water surfaces or wet or waxy leafs),
or a flash picture with overexposed near and under exposed far regions, has
a huge contrast (scene dynamic) range. Then, if you evenly distribute the
luminances over the total available range of RGB values, then the shadow
areas (or the flash/sunlit areas, or both) will have a relatively low
contrast, they (each) are only covered by a small percentage of the total
range. If the high contrast transition between them is locally reduced, then
there will be more levels left for shadows or highlights. That will increase
feature visibility in both highlights AND shadows.
If given the choice, I'd rather have the scanner software do it's magic on
the raw files than me trying to fix it in postprocessing.
Bart
Erik Krause
you wrote...
> I don't know for sure either, but perhaps the request was more profound than
> the person suggesting it realised. For the theory have a look at:
> http://www.cs.huji.ac.il/%7Ewerman/Papers/hdrc.pdf and
> for a nice collection of examples: http://wind.winona.msus.edu/~stafford/p/
>
> The idea of luminance gradient leveling can be applied to a single (RAW /
> HDR)image, or a composite of differently exposed versions of a scene (like a
> really good long exposure function).
> It is also somewhat like the Retinex technology from NASA, but I'm more
> impressed with the HDR approach.
Theres a software out right now that does high dynamic range
compression: HDR-Shop
http://www.debevec.org/HDRShop
I asked whether we need that. Ok, I know now: I need it :-) (see other
posting)
--
Erik Krause
Pasi Savolainen
> I appreciate all the suggestions, and will take them
> into account when considering changes to VueScan.
I hope you'll count my voice ;)
> User's Guide
> ------------
> 2 PDF file of User's Guide with more easy explanations
I hope this link helps:
<http://freshmeat.net/projects/htmldoc/?topic_id=87%2C154%2C857%2C856%2C853>
It's under GPL, but you can't get compiled binaries without paying, however
I believe that Ed can manage to compile a program :D
There's even a web-based PDF generator (made a page of 'contents', it didn't
follow links..)
> More comprehensive documentation, integrated with VueScan
Any chances to get tooltips? I don't see them in Linux version, hadn't ran
windows version yet. I recall wxWindows had them pretty easy way, simple
parameter to constructor? (Must say that winapi had it in very horrible way
when i tried to do that..)
> Performance
> -----------
> Speed up sluggish GUI
I think I see that vuescan is multithreaded now, but doens't take UI
advantages of it as of yet (.42), so it seems like this is coming along?
> Further work on preview (better speed and resolution)
> Faster processing
I must say that on a P233, it takes pretty long time to generate a preview,
and recalculate almost any color adjustments. Certainly much longer than
GIMP takes to recalculate Levels adjustments for same (16bit) picture. A
part of this may be because GIMP shows it progress in a visible way.
On a Tangent, what is that delay after showing 'Updating Buffer 98%', and
showing the actual preview picture, is it supposed to be there?
(unscientific test, 1410 dpi, 1280x2000 pix, 48bit, mirror, takes 20sec
(showing 98%), and 100% cpu usage).
> User Interface
> --------------
> Wizards to help make Color adjustments easier
I don't know about Wizards, but if colors adjustments could be done in same
way as Levels adjustments (etc. histogram in same 'subwindow', and controls
below). I must apologise if I understand the concept of color correction
wrong, and this is impossible to implement this way.
> Simplify user interface
Can't really agree, but in the way that most used controls (when going
advanced route) should be on same panel. I constantly swap between color and
device tab (in device I mostly switch dpi & number of frame).
> Split-screen mode for before/after correction
I think in some Polaroid 35SE software there was a preview mode which split
screen in _adjustable_ halves, and the anchor points for that line could be
pretty much anywhere, so I could split almost any point of a picture in half
for comparision.
> Display all previews at the same time and allow changes to each
This would be really great. Also I think that vuescan should take advantage
of disk-space, not only memory, when keeping previews of same strip of
film.
> Hotkeys for next/previous frame
Absolutely.
One thing I can't believe anyone hasn't mention, is a 'Save As' dialog. I
don't think I'm the only one who just scans some random picture, without
need of further adjustments.
At the moment I have this routine: Scan from strip some images, fire up
gqview (irfanview equivalent), with that choose right picture and fire up
GIMP (PS equiv) from that. I have now a directory with scan00XXX.tiff files,
and i sure can't distinguish much from them. Setting output directory for
every scanning session (I do maybe 20-30 scans at a time) seems a bit tedious.
Am I missing something?
> Features
> --------
> Save to PNG format
I made a comparison between .TIFF and .PNG, and png wins at about 72% (eg.
30% smaller). there is a script and sample output in article
<http://groups.google.com/groups?selm=afqiq6%243lfu%241%40midnight.cs.hut.fi>,
unfortunately it's in finnish, but I believe anyone can understand the
results. testpicture set was rather 'normal'.
> Dark frame subtraction, other noise reduction
AFAIK this works only for digicams. In a scanner there could be a dark lines
subtraction, but I'm pretty darn sure any scanning software already has it.
(rather difficult situation in case of multiline scanners).
> Support for 3rd party plugins
wxwindows 'supports' python, and python supports anything?
> User defined negative types
or rather remembered neg/slide rgb exposure & basic color adjustments
(selectable from a drop down box?)
> API for other programs
API could be a plugin.
Also, one more (IMHO Great) feature would be to do 'mouse zoom'. In this,
you'd have a window with 'blown up', zoomed (each pixel at least in 2x2
size) picture of the point where mouse cursor is. I know thot this can be
done with a most video card's own utilities, but they _don't_ have the extra
info vuescan has. In my 'vision' this updates as mouse cursor moves over the
image, and helps to set more precise autofocus spot, and check for slide's
focus.
And, to not forget, thanks again Ed!
--
Psi -- I need no force <http://www.iki.fi/pasi.savolainen>
Vivake -- Virtuaalinen valokuvauskerho <http://members.lycos.co.uk/vivake/>
Jan Exner
>> Simplify user interface
> Can't really agree, but in the way that most used controls (when
> going advanced route) should be on same panel. I constantly swap
> between color and device tab (in device I mostly switch dpi & number
> of frame).
Oh yes! A pair of buttons on the color tab ('next frame'/'previous
frame') would be grand!
>> Hotkeys for next/previous frame
> Absolutely.
Or hotkeys, of course.
And I sometimes think the interface should disable itself while the
background is working. I often find myself clicking somewhere while
VueScan is scanning or saving, and then wondering why nothing
happens... I guess I am used to UIs having an input event queue.
Grüße,
Jan
--
Jan Exner · ex...@gmx.net · 0x9E0D3E98 · http://www.jan-exner.de/
Neues aus Frankreich http://www.jan-exner.de/france.html
Bart van der Wolf
Yes, amazing isn't it. I agree it works fine for assembling a high dynamic
range image from several low dynamic range ones (as you mentioned earlier).
Unfortunately the multiple exposure examples (obviously) involve stationary
subjects like architecture, but they do rock! However, the principle also
applies to a single (LDR) image, or a modified "long exposure" (HDR) scan.
Also unfortunate is that the HDRShop software doesn't do (or I have not
found) the final step, reconstruction with the attenuated gradients, leaving
us now with an enormous dynamic range which looks dull overall, like a
printfilm scanned with a Black and Whitepoint set to 0%. Adding an S-shaped
curves correction helps, but doesn't even come close to what can be done.
I'm currently (amongst others) experimenting with the postprocessing step,
to see if something simple can be done with HDRShop+Photoshop. On success,
I'll share the outcome.
Bart
Erik Krause
you wrote...
> I'm currently (amongst others) experimenting with the postprocessing step,
> to see if something simple can be done with HDRShop+Photoshop. On success,
> I'll share the outcome.
I'm looking forward to it. Meanwhile I've done some experimenting
myself. I started with two masked adjustment layers. As a mask I took
the desaturated image itself, blurred by gaussian blur or (better)
median filter. One layer got that mask inverted. So I can adjust
shadows and highlights independently with the two layer levels or
curves.
Unfortunately this gives halos around edges with high contrast because
of the blurring of the masks. You can paint back the edge with the
history brush, but I looked for a faster solution.
It comes with the "Glowing Edges" filter. Desaturate a copy of your
image and copy it again. Select "Glowing Edges" in one of the copies. I
have used Edge width 10, Edge Brightness 5 and Smoothness 15.
Invert this image and use it as a layer mask in the other desaturated
one. Set selection to that layer mask and blur the image with the
median filter. The radius should not be much bigger than the edge width
in "Glowing Edges" (12 with me).
The result is a mask with preserved high contrast edges and blurred low
contrast ones. I don't understand the maths in the hdrc.pdf, but I
suppose, that's more or less what they mean by "gradient attenuation".
The mask should finally be applied to the two adjustment layers of the
original image, inverted to one of them of course.
All this should be tested on different images and put in a photoshop
action finally.
--
Erik Krause
Bart van der Wolf
SNIP
> The result is a mask with preserved high contrast edges and blurred low
> contrast ones. I don't understand the maths in the hdrc.pdf, but I
> suppose, that's more or less what they mean by "gradient attenuation".
Perhaps an easy way to mentally visualize what is done is this:
Imagine a relatively flat landscape with a high plateau mountain in its
center, viewed from above. The mountain has steep edges, but the top is
relatively flat just like the lower surroundings. If the surroundings
correspond to values/heights close to zero, and the plateau values/heights
close to 255, you can imagine that relatively many values are wasted in
accomodating values to represent the steep edges, and few values are left to
describe the flat areas. If we could proportionally shrink the height of the
steep edges, that would allow us to increase the contrast in the relatively
flat areas, while staying within the 0-255 range. The gradient of the edge
transition is attenuated (while avoiding halos) and more dynamic range
resolution can be spent on large surfaces (without exaggeration which would
increase graininess).
Mathmatically that is a much more complex issue because each pixel is
considered a plateau or a canyon when compared to its neigbors, but there's
the challenge. It still has to look natural, no halos, but visually
improved.
Bart
Erik Krause
you wrote...
> > The result is a mask with preserved high contrast edges and blurred low
> > contrast ones. I don't understand the maths in the hdrc.pdf, but I
> > suppose, that's more or less what they mean by "gradient attenuation".
>
> Perhaps an easy way to mentally visualize what is done is this:
I think I understood it. You describe almost what I think that I do wis
my methode. Let's proof it.
As in your example I do the following: I construct a mask, that
preserves steep edges but blurs the more or less flat areas. This mask
can be used to manually adjust the height of the flat areas. I can rise
the low areas and lower the high areas. Since the mask preserves the
steep edges, there are no halos.
I put this all in a PS6 action tomorrow. If you (or anyone else) want
to try, I can send it per email.
--
Erik Krause
Jorma U. Heikkinen
<I put this all in a PS6 action tomorrow. If you (or anyone else) want
to try, I can send it per email.<
Please do.
Ukko Heikkinen
--