CRT film-outs

[Paul Bruening]

Good and bad news. I've done the scans of a film trailer frame and critically measured the analog sound track. The DG5 can screen grab well beyond the needed resolution. It can also display that track with sufficient resolution. The problem is still converting a sound into an optical track image. I'm still grinding over that.

[John Sprung]

Erm, that was the case in the old days when they used to add a strip of new monochrome emulsion after the colour picture development, and then run the film through a monochrome chain to process the sound track.

 

I remember there being processes in which the track was developed differently than the picture on a release print using a wheel applicator to get chemicals onto that area only. But striping new emulsion I've never heard of. It doesn't sound practical. We need Dominic Case and/or Robert Houlihan to chime in here.

 

 

 

-- J.S.

[Phil Rhodes]

I remember there being processes in which the track was developed differently than the picture on a release print using a wheel applicator to get chemicals onto that area only.

 

That was my impression also.

 

P

[Dirk DeJonghe]

The track on positive prints used to be a silver track where a sound developer application was needed after the first fixer to retain silver in the emulsion. The silver was needed to excite the infra-red sensitive photocell as used in the older style soundhead as installed in the projector since the 1920ies.

About 5 years ago the switch was made to equip projectors with modern soundheads with red LEDs as light source. Silver tracks were no longer needed, the new 'cyan tracks' did not need any additional processing steps and a lot of wash water could be saved as well.

Currently, on 35mm cyan dye tracks are the only game in town, on 16mm silver tracks with applicator are still used but becoming very rare.

[Paul Bruening]

The track on positive prints used to be a silver track where a sound developer application was needed after the first fixer to retain silver in the emulsion. The silver was needed to excite the infra-red sensitive photocell as used in the older style soundhead as installed in the projector since the 1920ies.

About 5 years ago the switch was made to equip projectors with modern soundheads with red LEDs as light source. Silver tracks were no longer needed, the new 'cyan tracks' did not need any additional processing steps and a lot of wash water could be saved as well.

Currently, on 35mm cyan dye tracks are the only game in town, on 16mm silver tracks with applicator are still used but becoming very rare.

 

Dirk,

 

For the first time ever, I looked over at where you are from. My gosh, man! Your English is better than ours.

[Paul Bruening]

Good and bad news. I've done the scans of a film trailer frame and critically measured the analog sound track. The DG5 can screen grab well beyond the needed resolution. It can also display that track with sufficient resolution. The problem is still converting a sound into an optical track image. I'm still grinding over that.

 

I also had to do the math on various lenses available to shoot the screen. So far, the Nikon 85mm has the least distortion. Nikon provides software that can correct distortion. Even the 85's .4% barrel distortion could add a 24Hz stereo flutter to the audible sound if not compensated digitally.

 

I'm still mentating on the kind of registration test that will give me critical measurement of the Mitchell. The usual offset grid might not do it.

[Keith Walters]

I remember there being processes in which the track was developed differently than the picture on a release print using a wheel applicator to get chemicals onto that area only. But striping new emulsion I've never heard of. It doesn't sound practical. We need Dominic Case and/or Robert Houlihan to chime in here.

-- J.S.

You could be right, this is not my area of expertise. The process is called “appliquéing” and involves some sort of roller applicator. I couldn’t find anything more specific on the Internet, you just get virtually the same story repeated over and over; how the switch to cyan track sound is going to have massive but largely unexplained benefits for the environment and so on!

The problem with normal colour film processing is that all of the silver is removed in the bleaching, so perhaps what is laid down is a strip of some sort of protective gel. If that’s the case though, you’d wonder why the film isn’t made that way in the first place. Why would you need to apply it in the lab?

However the process definitely did involve running the processed colour film through a monochrome processing chain, and in fact one of the many advantages of cyan soundtracks is that processing labs could convert their existing mono chains to colour.

[Dirk DeJonghe]

First, in the 20ies, there was black and white positive print film, containing lots of silver. The first soundheads for projectors had a cell sensitive to infrared because the silver in the film would stop the infrared light real well. No redevelopment of the sound track area was needed. Even today, B&W prints are still made the same way with a silver track.

 

With Technicolor prints, the raw print stock was B&W sensitive and first the sound track was exposed then processed in convertional B&W process, dried and then CMY dye-track printed on pin-registered belts. Technicolor prints had a silver track even before the dyes were applied.

 

With colour positive (ECP) there was a problem: after bleaching and fixing none of the silver remains in the emulsion and the infra-red sensitive cell in the projector would have a very bad S/N ratio. To solve this problem, a sound redeveloper (hydroquinonein a thick paste) was applied after the bleach but before the second fixing bath. This would prevent silver being removed by the second fixer but only in the soundtrack area.

 

With the advent of newer 'red light' soundheads a switch was made to eliminate the soundtrack redevelopment since the projector cell would work well with just dye-track instead of silver-track. The new soundheads are backwards compatible and still read the old silver-tracks just fine. The elimination of the hydroquinone and wash water to remove the soundtrack paste from the film before it enters the second fixer is substantial and a very big ecological step.

 

Between the silver-track and the final cyan dye-track there was an interim measure, the magenta track, with track application, intended to be readable on both old and new soundheads. The new cyan track is not readable on old soundheads. Since all 35mm projectors are supposed to have the new 'red' soundhead now, there is no reason to run the soundtrack application any more, except for 16mm as I noted before.

[Paul Bruening]

More news,

 

I got the slides back. Like Phil said, the contrast of the Fujichrome is really noticeable. Dominant, actually. The Sekonic meter did not deliver an accurate reading of the DG5 screen. I'll need about 2 more stops. That slows shutter time down to more like 1 fps. That puts 2hr print production to around 2 days. That's fine since the most I'd ever produce is a dozen prints or so for festival and promotional release. I'm actually at the point where I have to do the rest of the tests on print stock. The screen came off with a bluish cast even with daylight slide stock. LUTs won't correct overall color cast. That will have to be arrived at with filtration.

 

Out of 35 slides shot at varying shutter speeds, no instance of detectable screen refresh occurred even down to 1/60th of a second against the slowest possible refresh of 12Hz. These LCDs may not be an issue like CRTs.

 

On the plus side, the DG5's screen matrix is undetectable even under microscope. I'm very happy about that. When compared to samples of conventionally made, fourth generation images the quality of the DG5/slide images is better. Better is better. But I was hoping for amazingly better and didn't get that. I still have to test that with the 75mm Cooke lens on the Mitchell before I decree the image good enough or not. The Cooke has critical focusing witness marks. The Nikon 85mm that I used for this test is not easy to arrive at critical focus. So, I can't really be sure I had the screen at the best possible focus. The FM2's viewfinder, focus screen is rated down to only f2.8. That means I can't dependably focus at low apertures in these slide tests

 

I still have to test whether the resolution on the Mitchell system is adequate for the analog sound idea.

[Phil Rhodes]

Off the top of my head I'd expect a really good, really slow film stock (such as print media) should be able to see a hint of the grid on a 1920 TFT - I'd expect to need to defocus slightly to lose it. Consider your lens, optimal aperture, and focus setup. You will, of course, need outrageously sharp lenses to see this, and everything else set up absolutely optimally. Are you working in a temperature-controlled environment?

 

Optical filtration is preferred since it doesn't cost you colour resolution, but I'd expect to have to tweak gamma and so forth.

 

I guess what we really need to see is scans of shots of DLAD Marcie (remembering John Pytlak).

 

P

[Paul Bruening]

I guess what we really need to see is scans of shots of DLAD Marcie (remembering John Pytlak).

 

I've got 4th gens of her. That would help me see how much loss occurs progressively (4th gen source, scan on Epson, shot back through DG5).

[Paul Bruening]

How different is the galvanometer in a 16mm sound camera from a 35mm sound camera? Could you re-sprocket and re-gear a 16mm sound camera and get 35mm results from it's 16mm light gate?

[Paul Bruening]

Phil,

 

You and Keith had thrown out some numbers that were the digital equivalents of analog sound. I have done a little math on the DG5 system. 2400 steps X 24 fps = 57,600 steps or samples. That part is better than even CD sound. How would you then rate the equivalent "bits" for mono, stereo and Dolby's surround thing (2 tracks with each side carrying variations- a form of 4 tracking). Assuming I can work out the track imaging and associated challenges, I'd like to be able to state what kind of digital equivalent sound I can produce. Something like, "It can do the equivalent of 16 bit 57,600."

[Ben Syverson]

Does the DG5 print the optical track as part of the picture? That is, does it print out the full silent gate with picture and sound all at once? If it does, it's fairly trivial to find the bitdepth of the audio it can produce. Just take the width (in pixels) of the audio portion of the image, then:

bitdepth = Log2(steps)

 

As for the question of how much bitdepth is possible on an optical soundtrack, that's more tricky. We would need to know the resolving power and SNR of the red/cyan layer of that particular print stock. However, we can ballpark it. The very best consumer color film (Astia) resolves well below 100 lp/mm. The very best fine grained B&W can resolve upwards of 200-300. Let's be extraordinarily generous and assume color release print stock could achieve 200 lp/mm, which is highly doubtful. A mono soundtrack is what, 2.94mm wide? That means 588 line pairs, which given Nyquist, means we can achieve 294 discrete levels.

 

However, there's noise in the form of film grain. Let's again be generous and say that the grain will give us an average error of only 2 levels (less than 1%), which is a tiny deviation of only 10 microns! 20 * Log10(294 / 2) = 43.3 dB. In other words, just over 7 bits. (This is actually wrong, because the layout of the levels is not linear, but whatever). For stereo, it's half that, so just over 6 bits.

 

Still the picture is not so simple as that. Even 1 bit can give you amazing results (see: DSD). Sampling rate and error diffusion matter greatly. The sampling rate (18.66mm * 200 lp/mm) gives us 1866 samples vertically for every frame, which translates to 44,784. Basically, CD quality sampling.

 

Luckily, that 1% noise actually has the effect of increasing the apparent bitdepth, because you can achieve "in-between" levels. (See: dithering in images). However, the result will sound very slightly noisy.

 

This seems to model the reality of optical sound pretty well. We know 35mm optical tracks sound different than a CD (slightly noisier, less dynamic range), and this helps explain why. Hopefully it can also serve as the starting point for some back-of-the-envelope calculations for people building filmout machines as well. :)

Edited February 14, 2010 by Ben Syverson

[Paul Bruening]

Hey Ben,

 

Sorry about the late reply. I haven't checked this thread for a while. If you scrutinize some of our previous posts, laborious as they may be, you'll find answers to your questions that we have tried to tackle to date. The current hang-up on this project remains, "How do I turn sound files into optical track images for addition onto the DG5's picture image?"

[Paul Bruening]

Hey Ben,

 

Sorry about the late reply. I haven't checked this thread for a while. If you scrutinize some of our previous posts, laborious as they may be, you'll find answers to your questions that we have tried to tackle to date. The current hang-up on this project remains, "How do I turn sound files into optical track images for addition onto the DG5's picture image?"

 

Hello again Ben,

 

Have you had a chance to go through the previous posts, yet? I think we could put our brains together on this, putting the goals closer within reach. I don't have any money to compensate you. Post services at cost is all I have to trade.

[Ben Syverson]

Ha, no compensation needed. :) I've gone over the thread and think I mistook the DG5 for a film recorder, when it's actually a monitor! The resolution of the DG5 is 3840 x 2400, right? So the full (silent) aperture area would be 3200 x 2400.

 

If the sound area is 2.94mm wide out of 24.89mm, that translates to 378 horizontal samples for audio. I'm assuming we want stereo, so that's 189 samples per channel. That's 7.56 bit, though it's a bit misleading because the values aren't linear -- so the result will already sound different from a "normal" digital audio file at 7 or 8 bits.

 

Based on the numbers I scribbled in my last post, it looks like the monitor will NOT be the limiting factor in the system, so that's good news. It's more likely to be the film's grain or MTF that limits the effective bitdepth. Another way to say it is, you'll probably never see pixelation in the final film print of your optical track, even with a microscope.

[Paul Bruening]

Ben,

 

I'm hungry for a hacker. I need to convert WAV files into a digital image equivalent of two optical tracks of Dolby Surround. I can't go to film without it. I can't go to big screen without it. I need some coaching on Adobe macros to automate the screen grabs, digital alterations and patching back together for the DG5. I'm stuck in limbo without these two categories solved.

[Ben Syverson]

Paul, there is a waveform visualizer in After Effects which might be really useful in this context. The key will be applying the appropriate audio filtration -- didn't I read that optical sound has some kind of EQ or nonlinear response curve in order to make it physically larger on film? Real waveforms for dialog tend to be very small relative to the full height.