Film-Tech Forum ARCHIVE: Optimized for White or Red: BIG problem

This is topic Optimized for White or Red: BIG problem in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.

To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=006866

Posted by Frank Angel (Member # 248) on 01-13-2005, 09:54 PM:

Ever since reading all that has been said about the difference between soundtracks optimized for white or red "excitation" I have really come to the conclusion that this is going to be a looming problem and one royal pain for rep houses playing older prints.

If I understand it correctly, cyan tracks not withstanding, there are two different and distinct types of standard silver tracks -- those that were made before red readers were a twinkle in any engineer's eye and those that were made once reverse red readers were introduced. So in theory, if I am playing older titles, I can very probably get a print that was struck BEFORE red readers and red light optimization and unless I am using a white light playback -- use a reverse scanner with red light, I am in trouble. Correct?

And if I am to believe the guys here who know (and I do), I now understand that playing a track optimized for white light with a red exciter is going to sound like sibilance-city or a cat on a hot tin roof. And what's worse, there is no way at all for any projectionist to know which way the lab has optimized the soundtrack on any give print, because the labs (and it seems the SMPTE as well) have endorsed, what I would consider is a VERY significant change without providing ANY indication on the prints themselves which of the two very different optimization parameters it was made in. Am I still on point here?

So.....this is a really sorry state of affairs IMHO for rep houses, and it certainly is one sloppy way to go about letting those down the line know an important piece of technical information about a print's soundtrack, ESPECIALLY since there is no what to visibly tell which is which. From a archival/historical perspective, this sucks big time. Down the road it will all be guess work. Thing is, I am already down the road.

How is it that there is no SMPTE spec about this, at the very least, marking prints as red-optimized so a projectionist will know, gee, better not play this one with white light. Seems to me, that information is just as important as indicating if the soundtrack is "optimized" for Dolby SR or Dolby A, and like those two very different types, it is impossible to tell which a print is just by looking at the sound track.

If there is no practical way to distinguish between red and white light optimized prints that were made during the transition period, which would seem to be a significant number of titles over a 5-10 year period, no? Then the only practical thing would be to have the ability to switch light sources while previewing the print to determine which sounds best. Ideally this would necessitate some manufacturer to design a reverse scan reader with easily and repeatably switching capabiliity, back and forth between two light sources. So where is this piece of equipment?

Barring the availability of such a unit, seems to me that the only other solution to this untenable position rep houses are put in, would be to forego the advantages of better separation and frequency response that reverse scan provides and reluctantly go back to a front scan design and use the homebrew exciter lamp replacement. I would imagine that a high intensity white LED exciter could just as easily be fashioned the same as has been done with the red, so that both LED units would be built into standard exciter lamp bases for easy swapping out between the two.

What still baffles me is that this issue seems to have just slipped under the radar. BTW, everyone to agreed that playing a silver track excited by red light sounds harsh and splashy -- what happens the other way around? I have one booth that so far has not converted due to all the discussions here that really scared me -- I don't want to degrade my sound by installing new hardware (the Kelmars are sitting up on a shelf). Playing new prints with white exciters doesn't seem to degrade the sound in that theatre as far as I can tell.

Posted by Steve Guttag (Member # 268) on 01-13-2005, 10:01 PM:

Frank,

Cyan prints are obviously optimized for red-light readers only.

High-Magenta prints will play, more or less, equally on red, IR or white light readers.

All others should be considered to be optimized for white or IR readers. However, it has been noted that a pristine old track sounds best with an IR reader, a scuffed or poorly made old track will sound best with a good ol' exciter lamp.

Steve

Posted by Brad Miller (Member # 2) on 01-13-2005, 11:11 PM:

Frank, call up BACP and have them sell you a pair of analog only reverse scanning penthouses, then keep your existing exciter based "basement" readers. (I've strongly considered this for my own screening room.)

Posted by John Hawkinson (Member # 1135) on 01-13-2005, 11:45 PM:

Frank, there is no reason to believe that any labs have produced any conventional silver-applicated tracks that are optimized for red LED readers.

A few months ago, John Pytlak suggested there might be some such prints, but after a bit, he realized that there had been a miscommunication. See pp. 37-38 of the Cyan Tracks thread, specifically the initial post, and then the correction.

--jhawk

Posted by Matthew Jaro (Member # 1786) on 01-14-2005, 10:25 AM:

I've screened older prints with a BACP reverse scan red reader, and it certainly is not sibilant city. The sound track is certainly quite acceptable (if not perfect). Variable density tracks also sound pretty good. The reverse scan has a much better frequency response than the exciter lamps did. When I show a modern print in Optical SR, with a Panastereo processor, it sounds quite good with the reverse scan red LED. Steve can verify that my sound system is not poor by any means. I'm using a Panastereo processor, QSC amps, and tri-amp connections to JBL application engineered speakers.

I imagine that the most serious problem would be with blue sound tracks. I have never run into one of these, fortunately.

I don't know how much better the sound tracks would sound with IR LEDs, but the qualilty with red is certainly not bad and the dialog is quite clear.

I already have penthouses for DTS and Dolby SRD, so a third penthouse would be a horror.

Frank, if you would like to come and see for yourself, you a are more than welcome. I'm in the Washington, DC area.

Posted by Brian Guckian (Member # 1678) on 01-14-2005, 12:23 PM:

Frank, if you're after absolute fidelity then you should go the route Brad has suggested and have both red and tungsten readers.

I don't have the exact info to hand, but I recall that the backgound noise level is a little higher when a traditional silver track (as opposed to a High-Magenta one) is read by a red LED soundhead. BUT this level is still some way below the noise "floor" of the typical theatre (i.e. the reproduced noise is lower than, and therefore masked by, the background noise in the theatre). And this is using "quiet" theatres as a benchmark. (There may be detailed technical information on Dolby's website).

It very much depends on the acoustical specification of your auditorium. This is why reading silver tracks on red readers is perfectly acceptable in the normal theatre environment. Also, I worked in a preview theatre a few years ago with good noise isolation. We ran quite a lot of classic prints for functions, etc. and never had problems with audible noise. Some prints were a little harsh, but we could usually deal with this by adjusting the mono EQ in our CP65 on a print-by-print basis.

I do know that the NFT in London were using seperate reproducing heads to deal with this issue...some of the UK posters here may know more.

A related problem now rearing its head is the availablility of many types of exciter lamp since the market has contracted so much. I believe that Kinoton for example are now using an LED-based system for special applications such as 16mm, to overcome this.

Posted by Matthew Jaro (Member # 1786) on 01-14-2005, 01:54 PM:

I'm not sure I understand the physics of what happens when you read a silver soundtrack with red light. Consider variable-area sound tracks: The modulated portion is basically transparent and the background is black. The contrast ratio between black and transparent should be the determining factor in the track signal-to-noise ratio. Since the track is transparent, what does the color of the light matter? I'm not talking about cyan or magenta tracks.

Similarly, the frequency response is determined by the minimum of the recording slit and the reading slit(or reverse scan LED array) width. Again, the color of the light should not make a difference.

How can the color of the light introduce distortion? I've yet to see a believable explanation of this.

Posted by Gordon McLeod (Member # 33) on 01-15-2005, 07:45 AM:

The colour of the light affects what is called cross modulation distortion. That is why tracks are optiomized for a particular type of playback
do a search on crossmodulation there is lots on the sight about it

Posted by Frank Angel (Member # 248) on 01-15-2005, 11:50 AM:

Gotch. Thanks guys. I didn't catch that correction in the other thread. You can understand how this would be a major concern if playback of older analog tracks were compromised. Brad, I will check out the BACP. I always thought they had an edge over Kelmar anyway.

Posted by Dominic Case (Member # 1830) on 01-16-2005, 06:23 PM:

quote: Matthew Jaro
Since the track is transparent, what does the color of the light matter?

It's not about signal-to-noise, it'a about cross-mod distortion, which is due to image spread at the edges of the waveform of the track. Both negative and positive print have a cerain amount of image spread, which has the effect of reducing and distorting high frequencies. Normally, prints are made at a particular density so that, as much as possible, negative and print image spreads cancel each other out.

However, this is also a matter of colour balance, as the three dye layers each have their own properties. A white reader only sees silver, but it sees it in whichever layer(s) have been exposed & developed. A red reader sees not only silver (in traditional redeveloped tracks) but also the cyan dye, increasing the contribution made by that layer to the soundtrack. It sees little or no trace of the magenta dye present in traditional (dark blue) or high magenta tracks.

Posted by John Pytlak (Member # 331) on 01-17-2005, 10:23 AM:

http://www.dyetracks.org/ci.os.0012.reddye.html

quote:
Crossmod Performance

A conventional silver print, played back on a red LED reader, will not display the same crossmod cancellation characteristics as it would if it were reproduced on a standard tungsten reader. Figure 3 shows that a silver track read on a tungsten reader will display good crossmod cancellation when the negative employed to print the track has a density of 2.18 optical density units (O.D.). The same track needs to be printed from a negative with a 2.6-O.D. density if it is to display good crossmod cancellation when reproduced on a red LED reader. The filter pack used to create the high-magenta silver print in Figure 3 is W12 + 0.2 ND + 0.90 C.

Fortunately, there is a sound negative density, in conjunction with a print color balance, that will produce a release print capable of playing equally well, in terms of crossmod cancellation on both tungsten and red LED readers. This print is called a "high-magenta" applicated print and requires a negative density of ~2.7 O.D. with a print color balance of 1.09 IR, 1.44 Y, 2.02 C, 3.82 M. Figure 4 shows the crossmod results from such a print.

The same negative used to print a high-magenta silver track release print can be used to print a cyan dye track release print. Figure 5 shows the crossmod results from a cyan dye track print reproduced with a red LED reader. The target color balance for this print was 0.36 IR, 0.36 Y, 2.14 C, 0.38 M. The graph shows that the best crossmod performance was achieved with a negative density in the range of 2.6 O.D.

http://www.kodak.com/US/plugins/acrobat/en/motion/support/h44/h44.pdf

quote:
Similarly, sound engineers can use a series of tests to
make sure of the quality of their product. The calibrated
input to the sound engineer’s system is an electronic signal,
namely, the cross-modulation signal. When sound engineers
get back the exposed and processed negative, they can
measure the cross-modulation signal and determine whether
or not the negative was properly exposed and how the
negative should be printed. They do not make their
measurements on a densitometer as does the photographer,
but rather on special equipment for measuring crossmodulation
distortion. They will have made, however, as did
the photographer, a previous series of tests (the crossmodulation
family) on which they will base their decision on
how to print the negative. Finally, when sound engineers
receive the print of the sound track, they can again measure
the cross-modulation signal to determine whether or not the
negative was correctly printed. Used this way, the crossmodulation
test provides the sound engineer with a method
of determining the correct exposure for a photographic
sound track. This method does not rely on subjective
judgments. It provides the engineer with a consistently high
quality product.

Posted by Matthew Jaro (Member # 1786) on 01-17-2005, 11:02 AM:

Thanks for the info Dominic and John. It seems that when I play SR prints (which must have been optimized for white light), they sound really good. The results are much better than I got with my old tungston system. I don't hear any sibilant distortion. I know that many venues now have red light readers in place of tungston. Is the cross modulation distortion great enough to warrent two readers on a system, or is just a minor amount of distortion? Is this something we should really worry about, or is the problem mostly theoretical?

Posted by John Pytlak (Member # 331) on 01-17-2005, 11:16 AM:

quote:
It seems that when I play SR prints (which must have been optimized for white light), they sound really good. The results are much better than I got with my old tungsten system. I don't hear any sibilant distortion. I know that many venues now have red light readers in place of tungsten. Is the cross modulation distortion great enough to warrent two readers on a system, or is just a minor amount of distortion? Is this something we should really worry about, or is the problem mostly theoretical?

The data presented in the May 1996 SMPTE paper by Paul R. Goldberg certainly shows a significant difference in the sound negative density required to produce an optimized cross-modulation distortion between a "white" light reader and red light reader for a cyan+magenta+silver track on the print (2.18 vs. 2.60 density).

As far as perceived quality, Goldberg writes:

quote:
From the previous discussion it can be seen that the industry switch to high-magenta applicated soundtracks is required for two very important reasons: first, the need to mitigate any diminished crossmod performance associated with the rapidly expanding use of red readers and, second, the need to facilitate the industry change-over to cyan dye tracks.

Although a quality of audio reproduction, which the average movie patron would not detect as being different from the current standard, and, may in fact, because of increased stereo separation, wider frequency response at high power levels and significant reduction of illumination nonuniformity distortion perceive as being an audible improvement, is produced by red readers, a closer match, with regard to crossmod performance, can and should be achieved.

Posted by Steve Guttag (Member # 268) on 01-17-2005, 12:23 PM:

It varies from film to film Matt. The DTS Buzz and Bill test reel that was issued in 1993 with a Dolby-A analog track really shows the issue on crossmod.

If you come across a print where you find the problem, we can mount a set of IR LEDs for your readers and optimize for IR and leave the visible reds ready for plug in with a nominal amount of HF attenuation (different focus points)

Steve

Posted by Bruce Hansen (Member # 281) on 01-18-2005, 12:36 PM:

Oh darn, I thought this would be a thread about what wine to serve with dinner, white or red. [Big Grin]