Testing 1 2 ...
5 views
Skip to first unread message
Roberto Muscia
Sep 26, 2023, 2:03:22 PM9/26/23
to FMML Filtermeister Mailing List
Just checking if this post arrives ...
Roberto
Roberto
--
Gemaakt met Opera's e-mailprogramma: http://www.opera.com/mail/
paul.s...@telenet.be
Sep 27, 2023, 4:55:42 AM9/27/23
to Roberto Muscia, filter...@googlegroups.com
Hi Roberto!
Glad to see a sign of life on FM!
I am still active with FM and have several projects that are already partly functional. Unfortunately, the lease on my home is coming to an end and I will have less time left in the coming months.
I hope FM keeps moving!
Best regards,
Paul
Glad to see a sign of life on FM!
I am still active with FM and have several projects that are already partly functional. Unfortunately, the lease on my home is coming to an end and I will have less time left in the coming months.
I hope FM keeps moving!
Best regards,
Paul
Van: "Roberto Muscia" <gro...@muscia.nl>
Aan: "FMML Filtermeister Mailing List" <filter...@googlegroups.com>
Verzonden: Dinsdag 26 september 2023 20:03:18
Onderwerp: [FMML2] Testing 1 2 ...
Aan: "FMML Filtermeister Mailing List" <filter...@googlegroups.com>
Verzonden: Dinsdag 26 september 2023 20:03:18
Onderwerp: [FMML2] Testing 1 2 ...
--
You received this message because you are subscribed to the Google Groups "FilterMeister Mailing List (FMML)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to filtermeiste...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/filtermeister/op.2bv4fsl2fxw4o6%40roberto-asus-g11.kpn.
You received this message because you are subscribed to the Google Groups "FilterMeister Mailing List (FMML)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to filtermeiste...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/filtermeister/op.2bv4fsl2fxw4o6%40roberto-asus-g11.kpn.
Roberto Muscia
Sep 27, 2023, 2:18:47 PM9/27/23
to filter...@googlegroups.com, paul.s...@telenet.be
Howdy Paul,
Are you saying basically you are kicked out of your house soon?? Ouch, sounds stressful! I hope you find a new place in time.
Yes, perhaps a rename of FMML to 'The Silence List' would be more spot on :-) No idea how many people are still subscribed to this list.
Hey, watch the 'several projects' thingy: before you know it they pile up quickly and in the end you have spent valuable time on things that eventually never get finished. Unless with that time 'exercise' was involved that is of course ;-P
Good luck getting things on track again the coming months!
Roberto
Op Wed, 27 Sep 2023 10:55:39 +0200 schreef <paul.s...@telenet.be>:
paul.s...@telenet.be
Sep 28, 2023, 7:38:44 AM9/28/23
to gro...@muscia.nl, filter...@googlegroups.com
Hi Roberto,
I am indeed being 'kicked out' of my home. After years of renovation work, the owner is settling in here. But that's it for this topic, since 'renting' is not included the FilterMeisterWiki.
The reply to the emails automatically fills in the personal sender. Since I have been aware of this for some time, I always manually add googlegroups.com.
FM may have unfortunately come to a (temporary?) technical standstill, but what is there remains very valuable and offers more opportunities than one would think for developing projects. These are primarily intended for personal use. It is particularly interesting to have a tool that allows you to put anything together quite quickly and efficiently. Optically inspired software has not yet said the last word either. It doesn't always have to be AI to play with.
In previous emails I have already indicated that FM becomes much more interesting when blank filter bases are offered, where you only have to enter the variables and the bare formulas. After all, most of the work now goes to the GUI and peripheral structure. To create a coherent program with the separate codes, you quickly get conflicting effects. These problems have been worked out in advance with a ready-made basis. I have already posted something about this. I will come back to this later with new and better packages.
FM actually has a lot of potential, but you have to extract it. I would like to make a cleaned up version where everything is purely 16-bit oriented and the original Filter Factory is removed. And speaking about kicking out: lets replace the gnome by a typographical logo with standing. The gnome was a historical joke. The basic preview is also better overruled (and therefore simply disabled) and replaced by a seamless zoom. These things already work, except for a few beauty details. That's why I'm waiting a little longer before releasing it.
FM can work much faster than commonly believed. To investigate this, I built a tool that allows you to compare any function or program line with an empty operation. An empty operation is simply repeating it x times, in this case without executing anything. Basically an empty scan, that also takes up CPU time. Because the delay that an operation costs is compared to this, the measured time delay is device independent and very precise. This produces surprising results and immediately shows many functions that you better should not apply and change to a different approach. For example, division is much slower than multiplication. In many cases you can divide by x by alternatively multiplying by 1/x...
I'll come back to more later on.
Anyone who read this email, please respond with an (empty) reply, so that we can get an idea of who is still connected... Thanks if you will do this!
I am indeed being 'kicked out' of my home. After years of renovation work, the owner is settling in here. But that's it for this topic, since 'renting' is not included the FilterMeisterWiki.
The reply to the emails automatically fills in the personal sender. Since I have been aware of this for some time, I always manually add googlegroups.com.
FM may have unfortunately come to a (temporary?) technical standstill, but what is there remains very valuable and offers more opportunities than one would think for developing projects. These are primarily intended for personal use. It is particularly interesting to have a tool that allows you to put anything together quite quickly and efficiently. Optically inspired software has not yet said the last word either. It doesn't always have to be AI to play with.
In previous emails I have already indicated that FM becomes much more interesting when blank filter bases are offered, where you only have to enter the variables and the bare formulas. After all, most of the work now goes to the GUI and peripheral structure. To create a coherent program with the separate codes, you quickly get conflicting effects. These problems have been worked out in advance with a ready-made basis. I have already posted something about this. I will come back to this later with new and better packages.
FM actually has a lot of potential, but you have to extract it. I would like to make a cleaned up version where everything is purely 16-bit oriented and the original Filter Factory is removed. And speaking about kicking out: lets replace the gnome by a typographical logo with standing. The gnome was a historical joke. The basic preview is also better overruled (and therefore simply disabled) and replaced by a seamless zoom. These things already work, except for a few beauty details. That's why I'm waiting a little longer before releasing it.
FM can work much faster than commonly believed. To investigate this, I built a tool that allows you to compare any function or program line with an empty operation. An empty operation is simply repeating it x times, in this case without executing anything. Basically an empty scan, that also takes up CPU time. Because the delay that an operation costs is compared to this, the measured time delay is device independent and very precise. This produces surprising results and immediately shows many functions that you better should not apply and change to a different approach. For example, division is much slower than multiplication. In many cases you can divide by x by alternatively multiplying by 1/x...
I'll come back to more later on.
Anyone who read this email, please respond with an (empty) reply, so that we can get an idea of who is still connected... Thanks if you will do this!
Best regards,
Paul
Van: "Roberto Muscia" <gro...@muscia.nl>
Aan: "filtermeister" <filter...@googlegroups.com>, "paul simoens" <paul.s...@telenet.be>
Verzonden: Woensdag 27 september 2023 20:18:42
Onderwerp: Re: [FMML2] Testing 1 2 ...
Aan: "filtermeister" <filter...@googlegroups.com>, "paul simoens" <paul.s...@telenet.be>
Verzonden: Woensdag 27 september 2023 20:18:42
Onderwerp: Re: [FMML2] Testing 1 2 ...
Jim Clatfelter
Sep 28, 2023, 11:20:44 AM9/28/23
to paul.s...@telenet.be, gro...@muscia.nl, filter...@googlegroups.com
To view this discussion on the web visit https://groups.google.com/d/msgid/filtermeister/1901261484.5799416.1695901122326.JavaMail.zimbra%40telenet.be.
paul.s...@telenet.be
Sep 28, 2023, 12:36:37 PM9/28/23
to ernst....@gmx.de, filter...@googlegroups.com
Hi Ernst,
Interesting story. Reproduction is also a key interest for me, but mainly for reflective pictures. To reproduce pictures you can handle 2 approaches: facsimile or an attempt to extract all optical information of the source and try to correct it to what the ideal should be.
In my experience is a scanner not the best instrument. It has the potential of high resolutions, but due to the stepping motor on such a micro scale there are always small motion shift errors in the direction of the scan. E.g. if you scan twice the same image and put it as a layer on top of each other, you will see the image 'breathing' as you swap. And it is extremely slow in operation.
With a scanner is it also difficult to adjust the exposure manually , so far if possible.
Less a problem for transparences, but for reflective originals the lighting device of a scanner emphasizes the structure of the carrier paper and the bronzing effects in the dark parts.
I turned to use a good digital camera with macro lens and a special lighting set up. The rest is software...
Paul
Interesting story. Reproduction is also a key interest for me, but mainly for reflective pictures. To reproduce pictures you can handle 2 approaches: facsimile or an attempt to extract all optical information of the source and try to correct it to what the ideal should be.
In my experience is a scanner not the best instrument. It has the potential of high resolutions, but due to the stepping motor on such a micro scale there are always small motion shift errors in the direction of the scan. E.g. if you scan twice the same image and put it as a layer on top of each other, you will see the image 'breathing' as you swap. And it is extremely slow in operation.
With a scanner is it also difficult to adjust the exposure manually , so far if possible.
Less a problem for transparences, but for reflective originals the lighting device of a scanner emphasizes the structure of the carrier paper and the bronzing effects in the dark parts.
I turned to use a good digital camera with macro lens and a special lighting set up. The rest is software...
Paul
Van: "Ernst Mattar" <ernst....@gmx.de>
Aan: "paul simoens" <paul.s...@telenet.be>
Verzonden: Donderdag 28 september 2023 16:25:50
Aan: "paul simoens" <paul.s...@telenet.be>
Verzonden: Donderdag 28 september 2023 16:25:50
Onderwerp: Re: [FMML2] Testing 1 2 ...
Hi all,
I'm still connected to this email list.
Here is a short description what I wanted to do with Filtermeister:
Some years ago I started the work for automatic recognition of pictures in old normal-8 or super-8 movies. The reason was to build a converter to digitize the old movies. My first testings were with filtermeister, but never had a good result. Maybe I didn't use the right algorithm for pattern recognition. During this time I found in Internet a description of a similar work, but written in Java. My testing with these programs had a very good result. But the author used a normal scanner whose resulution was not very high. Therefore I had to make some modifications to this program.
By chance, I found in ebay a very cheap new scanner which was sold as defect. This was a Reflecta ProScan 10T with a max resolution of 10000dpi. I could repair this scanner and print with a 3D-printer an adapter to allow scanning of movies. For the movie transport I used an stepper motor connected to an Arduino. Arduino and scanner where controlled by an PC so that it could work standalone day and night.
To combine the tousands of resulting pictures to videos including filtering I used standard software from Internet (AviSynth).
Best regards,
Ernst
I'm still connected to this email list.
Here is a short description what I wanted to do with Filtermeister:
Some years ago I started the work for automatic recognition of pictures in old normal-8 or super-8 movies. The reason was to build a converter to digitize the old movies. My first testings were with filtermeister, but never had a good result. Maybe I didn't use the right algorithm for pattern recognition. During this time I found in Internet a description of a similar work, but written in Java. My testing with these programs had a very good result. But the author used a normal scanner whose resulution was not very high. Therefore I had to make some modifications to this program.
By chance, I found in ebay a very cheap new scanner which was sold as defect. This was a Reflecta ProScan 10T with a max resolution of 10000dpi. I could repair this scanner and print with a 3D-printer an adapter to allow scanning of movies. For the movie transport I used an stepper motor connected to an Arduino. Arduino and scanner where controlled by an PC so that it could work standalone day and night.
To combine the tousands of resulting pictures to videos including filtering I used standard software from Internet (AviSynth).
Best regards,
Ernst
To view this discussion on the web visit https://groups.google.com/d/msgid/filtermeister/1901261484.5799416.1695901122326.JavaMail.zimbra%40telenet.be.
Roberto Muscia
Sep 28, 2023, 3:08:45 PM9/28/23
to filter...@googlegroups.com
Hi,
Same here: after experimenting with scanning photos and slides I started reproducing photos using a scanner (HP Scanjet 6300C) from 8-2008 to 10-2008 which was an annoying slow process. Even more because I applied 3-pass scanning which produced quite better (less noisier) results. VueScan features multi-pass scanning and luckily it worked with my scanner. I have seen it fail completely with other scanners as they did not start scanning at the exact same starting position.
The pain-staking slow process seduced me to try a camera to digitize the photos. At first it looked promising but I had to work on better lighting: 2-2010 I started reproducing photos using the camera. So it took 10-2008 to 2-2010 to come up with a working setup and the results were great (evenly lit and free of glare). Basically reproducing works lightning fast but... getting the photos out of the albums undamaged takes a while (stuck with double side tape). While scanning I had plenty of time for that but using the camera that is the part slowing the process down dramatically.
Photo paper texture is better visible on photos reproduced by camera than by scanner indeed but good lighting can minimize texture glare a great deal.
9 photo albums done and never finished the rest of the photo albums :-/
Roberto
Op Thu, 28 Sep 2023 18:36:31 +0200 schreef <paul.s...@telenet.be>:
paul.s...@telenet.be
Sep 29, 2023, 6:02:15 AM9/29/23
to filter...@googlegroups.com, gro...@muscia.nl
Hi Roberto and FM members,
Normally you should get better contrast details with a properly set up lighting than with a flatbed scanner. Viewed microscopically, paper resembles a carpet made up of shiny fibers. From a distance, the eye sees this as a homogeneous mass with only a slight sheen. This gloss obscures much detail that is nevertheless present in the dark areas of the image.
This glare can be almost completely suppressed by polarizing the lighting and applying a polarizing filter rotated 90° to the macro lens. Because everything that shines remains polarized, it can be almost completely extinguished. Everything that does not shine breaks the polarization and therefore remains visible. In practice, this means the chemicals that form the effective image of a photo. The paper carrier becomes, as it were, invisible.
Also know that a digital sensor responds very linearly to light, while our vision is logarithmic. This means that the dark areas in a digital image always make much larger jumps in value than the light areas. This can be dramatically counteracted via HDR technology. Combining a normally exposed image with an overexposed one. In post-processing, the overexposed image is pressed down again and a mask replaces the dark areas of the normal image. It all sounds cumbersome and it is. But once you have this set up, it can be done surprisingly routinely. It's worth it!
The attached photo shows a test shot with polarized lighting. The intention is to develop this into a permanent repro setup.
Normally you should get better contrast details with a properly set up lighting than with a flatbed scanner. Viewed microscopically, paper resembles a carpet made up of shiny fibers. From a distance, the eye sees this as a homogeneous mass with only a slight sheen. This gloss obscures much detail that is nevertheless present in the dark areas of the image.
This glare can be almost completely suppressed by polarizing the lighting and applying a polarizing filter rotated 90° to the macro lens. Because everything that shines remains polarized, it can be almost completely extinguished. Everything that does not shine breaks the polarization and therefore remains visible. In practice, this means the chemicals that form the effective image of a photo. The paper carrier becomes, as it were, invisible.
Also know that a digital sensor responds very linearly to light, while our vision is logarithmic. This means that the dark areas in a digital image always make much larger jumps in value than the light areas. This can be dramatically counteracted via HDR technology. Combining a normally exposed image with an overexposed one. In post-processing, the overexposed image is pressed down again and a mask replaces the dark areas of the normal image. It all sounds cumbersome and it is. But once you have this set up, it can be done surprisingly routinely. It's worth it!
The attached photo shows a test shot with polarized lighting. The intention is to develop this into a permanent repro setup.
Paul
Verzonden: Donderdag 28 september 2023 21:08:40
Onderwerp: Re: [FMML2] Testing 1 2 ... - scanning
Onderwerp: Re: [FMML2] Testing 1 2 ... - scanning
To view this discussion on the web visit https://groups.google.com/d/msgid/filtermeister/op.2bzwsqybfxw4o6%40roberto-asus-g11.kpn.
Roberto Muscia
Sep 29, 2023, 2:58:13 PM9/29/23
to filter...@googlegroups.com
Polarizing the light is an interesting thing! Haven't thought of that. I know from experience using a polarizer it is not removing the glare from all surfaces equally (e.g. with water reflecting it works much better than with glass reflecting) but it very well may work with subtle textured photo paper. Especially if the illumination is already polarized but I expect problems with the latter if the light sources are set up in an oblique manner (light source in line with camera causes horrible reflection).
Nice subject to experiment with, next time :-P
Roberto
Op Fri, 29 Sep 2023 12:02:09 +0200 schreef <paul.s...@telenet.be>:
Reply all
Reply to author
Forward
0 new messages