The Missing Link? -- Sound Capture with Hardware


Andreas thank you, we will wait with impatience.
Good luck in the defense of the thesis!


@dan74: Thank you very much!
Ok, here you go, this is a test of audio extraction from your clip:

Some artefacts that the AEO-Light version had are gone (uneven illumination, stitching, stammer), but others are much more evident now, because my software does not apply noise reduction to the audio.
The most noise in your video is caused by dirt on the left part of the soundtrack (those little dots), other noise sources are the jpeg compression, the ghosting artefacts also a rotation and distortion of the image and so on.
The next planned step for my software is to automatically remove the dirt, that should improve the audio further, but expect no wonders the source is far away from the optimum.
And another thing I have noticed is, that the AEO-Light version is some kind of pitched deeper and I can’t explain why that is.


Andreas, thanks for the video.
Excellent work, even on the worst example, just great.
I look forward to from you software after August.
Thanks again!


This is fantastic, @Andreas!! A totally acceptable solution to the “access first” philosophy behind Kinograph (vs. “perfection/preservation”).

Thanks so much for keeping us informed of your progress. This is a huge help to the project.



Hello everyone…as per my experience the soundtrack varies within a fixed width, the light source is a very
thin line, so it results in the photocell receiving an amount of light proportional to the width of that section of the soundtrack. It is very simple really.The difficulty is in the calibration, the speed has to be constant,
the slit perfectly focused and aligned to the soundtrack, i.e. the line has to fall horizontally across the soundtrack, and even a slight angle will cause the sound to be inferior.

printed circuit board manufacturing


True on all accounts. I’m looking at continuous servo motors that are used on tape recording machines to ensure constant speed. As for focus and alignment - definitely things that can be solved. It’s the electronics I’m clueless about. I have no experience with that kind of circuitry, as I deal mostly with MCU circuits. Got any recommendations?



Thanks for the great work. Very much like to know your progress.


It is still work in progress.
I am focused on the documentation and
on developing and testing cleanup algorithms at the moment.
Updates will follow, when I have finished the thesis.


thanks for the reply.
Did you have a chance to compare the software captured sound with the same sound track recorded through the sound head of a projector.
May be @dan74 has a sound track of the film.


did you have a chance to compare the sound track with the one from the sound head.


Hey. Here is my example of the real sound from a film projector.


No I have not compared it yet, because I have no access to film material and a projector.
But from theoretic standpoint, the quality should be at least the same, if the signal is read only (@ modern 2K scan, without scanner artifacts). But if there are scanner artifacts, they can be eliminated digitally.
The example videos shown in this thread had really strong scanner artifacts, which have been removed completely.
If the optical soundtrack is treated with algorithmic restoration, the quality would be far better than with a projector.
Most optical soundtracks have signal redundancy which can be used to detect and remove dirt and damages. And even the noise, which is produced by the emulsion grain, can be removed without digital artifacts.


Thanks so much for the explanation. There are some examples from Kinetta, where sound extraction s were done using AEO Light. Wish to see your


Thanks for the clip Dan.


Hey Andreas, your software seems like the perfect solution so far.
What are the effects of dolby A or SR using your program? Can they be captured properly without equipment?


I haven’t tested it yet.
But analogue optical dolby audio is an extension of dual bilateral soundtracks (2.0 mono), so it should be readable the same way without special equipment. What you’ll get would be a 2.0 surround file, which should be decodable by a dolby pro logic decoder once flagged as that. What I do not know at the moment is the difference of dynamic range compression of dolby to dual bilateral, and which equalisation was applied to such soundtracks. So maybe there has to be an expander and an equalizer to be applied to a dolby soundtrack.
It is on my to-do list too. :slight_smile:
But at the moment I am under time pressure to finalize the dirt removal and the thesis.
So first results with GUI will take a while.


Andreas, looking forward to it.


Hello Andreas.
Advice on setting up the windows in the program AEO-Light.
I’m in the program make only one change. The picture shown by the arrow -

You portray their settings when did my test scans.


That’s 100 frames, the sound pulls why? How to increase the speed of sound?


Well, that seems like AEO-Light computes the overlap between frames false, so that are
fewer samples per frame to the soundtrack than recorded (actually with gaps in the audio). The scan of the image has not enough overlap, so the software fails to find it. That is compensated via resampling, so the fewer samples are interpolated to match the
duration of a frame, thus results in a seemingly slower/deeper audio playback.
The best thing you can do is, to scan the frames with a bigger overlap, so that gaps do not
occur. If the scanner has not enough sensor area to do so, you can maybe make two shifted scans per frame (shift by 3 perforations instead of 4, in an extra scan only
for the audio). If you have scanned with 3 perforations, you have to change the frame rate to 4/3 of the original rate, in this case that would be 32 fps. And an overlap registration by image should be better.

My thesis isn’t finished yet too. I had to make a pause, due to a broken right hand. So the
thesis will end a little bit later.
At the moment my software is able to detect and remove most dirt, so that about 80-90% of
clicks and pops are removed, and the noise caused by the emulsion grain, is reduced too. Looking forward to bring it over 98% dirt removal and a detection of the rest 2% (for removal by hand), but that is on the agenda after the thesis has finished.