Film-Tech Forum ARCHIVE: Digital sound.

This is topic Digital sound. 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=008896

Posted by Stephen Furley (Member # 1277) on 08-26-2007, 04:52 PM:

I've just been to see 'The Simpsons Movie' at the local multiplex. I've not been there before; some things were good, some were bad, but that's another story. They could certainly do with more staff; not only could I not find anywhere to buy a ticket, I couldn't even find anybody to ask where I could buy a ticket! Eventually I found a few self-service credit card machines, rather hidden away, but surely there must be a box office somewhere. What happens if you haven't got a credit card?

Both adverts/trailers and the feature were in some digital sound format, as there were split surrounds, it was most likely Dolby.

They ran Pearl & Dean adverts, though rather more than we do, and no less than five trailers, about 25 minutes in total, which seems a bit excessive; the feature wasn't much more than an hour and a half.

For the last few weeks P & D have been sending in the adverts reel with a slight gap at some of the splices, so if you don't bloop them you get a click on the analogue track as they go through. There are also annoying mid-frame negative splices in the black before some of the adverts (why?) which need painting out, and which they hadn't done. But wait, I said they were running digital sound, but we still got the click between many of the adverts; I wouldn't have expected this. What was going on? Was the system reverting to analogue just as the splice went through, so we were hearing the analogue track at that point? This didn't happen at any of the splices in the feature.

I don't have much experience with digital sound (on film) most of the cinemas I use don't have it, and many of the films I watch don't either. I think I've heard DTS twice, SDDS once and Dolby about six times, and run it myself once.

[ 08-28-2007, 05:19 AM: Message edited by: Stephen Furley ]

Posted by Dominic Espinosa (Member # 2122) on 08-26-2007, 04:57 PM:

Back in the day I had the problem with Dolby Digital.
All my trailers would revert right before the splice, *pop*, then back into digital.

My guess is either the digital track ends before the splice or it's just not high enough quality to maintain a low enough fail rate when it reaches the splices.

I too rarely had the issue with features and since then all digital formats I've worked with have given no such issue (but I didn't show rolling stock at any of those locations).

Posted by Thomas Pitt (Member # 4322) on 08-26-2007, 05:37 PM:

Stephen, was it a VUE cinema? They have a system called "Ticket At Any Till" - meaning you don't have to go to the box office to buy a ticket. You can buy one from the refreshment stand, or sometimes from one of the floor staff who will just go to a terminal.

Back on topic, I myself have never experienced DTS sound. The two VUE cinemas I go to both have Dolby Digital, and I've been to a Cineworld a couple of times that has SDDS.

I've heard the digital sound drop out to analogue several times during the Pearl & Dean snipe, as if a splice wasn't made correctly. What I do know is that if the error rate gets too high, both SDDS and Dolby will instantly switch back to the analogue soundtrack until the error rate goes back down again. This stops you hearing digital errors such as white noise, repeating segments of sound etc. when the digital 'noise' is high.

I'm also guessing the splicing tape can obscure the digital data enough for the error rate to go high and cause the analogue switchover. I've certainly heard analogue switchovers in the middle of a movie, often when the film is dirty/damaged or has a splice.

This is also the reason why cinemas are supposed to leave on the MPAA/BBFC card at the beginning of each trailer. If you splice it out (most VUEs do), you'll often get an analogue switchover at the start of the trailer - switching back to digital a couple of seconds later. This makes the start of the trailer sound odd to the audience.
Also, taking one of the Surf's Up trailers as a good example. If you splice out the BBFC certificate at the beginning and just go straight onto the trailer, you'll mute a bit of dialogue at the very beginning!

Posted by Darryl Spicer (Member # 711) on 08-26-2007, 05:43 PM:

Dolby Digital vs. DTS

From above thread about 5 posts down by Steve Guttag

quote:
Apparently no one informed the general technical community about splice-caching?

For quite some time now, the DA-20 has been able to do what Dolby calls splice-cache. How it works is that using one of the low-bit-rate data tracks, they replicate the audio information around the reel-change splice and this is stored in the buffer. In the event there is a loss of data around the reel-change splice the buffer data is inserted and the splice plays through without a drop out.

This will only work for the reel change splices on those films so encoded and ones where the data track has not been so damaged to prevent the storage. This will not help poorly made reel repair splices or cutting out poor lab splices.

Steve

Posted by Stephen Furley (Member # 1277) on 08-26-2007, 06:04 PM:

quote: Thomas Pitt
Stephen, was it a VUE cinema? They have a system called "Ticket At Any Till" - meaning you don't have to go to the box office to buy a ticket. You can buy one from the refreshment stand, or sometimes from one of the floor staff who will just go to a terminal.

Yes, it was. I can see some sense in that idea, but in that case there really ought to be a large sign saying that cinema tickets are available there; I didn't see one. I'll have to have another look.

quote:
From above thread about 5 posts down by Steve Guttag

quote: Apparently no one informed the general technical community about splice-caching?

For quite some time now, the DA-20 has been able to do what Dolby calls splice-cache. How it works is that using one of the low-bit-rate data tracks, they replicate the audio information around the reel-change splice and this is stored in the buffer. In the event there is a loss of data around the reel-change splice the buffer data is inserted and the splice plays through without a drop out.

This will only work for the reel change splices on those films so encoded and ones where the data track has not been so damaged to prevent the storage. This will not help poorly made reel repair splices or cutting out poor lab splices.

That's interesting; I hadn't heard about that. That might explain why the problem didn't affect the feature; on the other hand maybe the cinema staff are just making better splices than P&D at the moment; I don't know.

Posted by Mike Blakesley (Member # 26) on 08-26-2007, 10:47 PM:

I hadn't known about that either. Thanks for digging that up.

I wish Dolby would print the sound a few seconds ahead of the picture and then buffer it to play back properly. That way it would be possible to cut off useless "dates" at the end of trailers, replace them with black film and still have the fadeout music play thru properly. I did that frequently with DTS using the free-wheel feature. It's the thing I miss most about that system.

Posted by Lyle Romer (Member # 1266) on 08-27-2007, 09:37 AM:

quote: Mike Blakesley
I wish Dolby would print the sound a few seconds ahead of the picture and then buffer it to play back properly.

I'm sure if the system was designed with the memory cost of today they would have done that (it would have allowed for more sophisticated error correction also). With cost per byte of RAM in 1990 (or whenever the design was first spec'd out) doing that would have added huge cost to the already $20,000 system.

Posted by Stephen Furley (Member # 1277) on 08-27-2007, 02:10 PM:

It's printed 26 frames in advance; with penthouse readers it must be read about two seconds ahead. Printing it even further ahead would cause bigger problems where different film elements, e.g. adverts and trailers need to be spliced together, and then you have the problem with changeovers. As it is they have to be timed from the motor start; if you advanced it much further you'd have to start reading the sound from the incoming machine before it was up to speed and stable.

I'd rather it was printed in level sync with the picture, which would require all readers to be penthouse type, of course. they couldn't be mounted too far above the gate in changeover installations though.

Posted by Michael Schaffer (Member # 1204) on 08-27-2007, 02:15 PM:

Yes, Stephen, digital sound in cinemas has been around for a good 15 years. [Wink]

Posted by Stephen Furley (Member # 1277) on 08-27-2007, 02:45 PM:

Is it really that long? I saw 'Schlindler's List in DTS at the Empire Leicester Square, and Dolby was out before that, so I suppose it must be getting on for that long now.

Many of the films that I watch are several decades older than that, though I suppose some may have had the sound re-recorded on digital tracks for new prints. The majority of the cinemas where I see films do not have digital sound installed, at least not for film. I've shown several recent independent films that did not have digital tracks, though most do.

We only have digital sound for digital cinema, but at least two of the films that we've run in digital, and we've probably only run about ten films in total on the digital projector so far only had two channel sound. One was 'This is England', which is still sitting on the server as it's coming back next Month for it's third run; I can't remember what the other one was, but in was the first digital film we ran after we had the sound system upgraded to split surrounds. I was disappointed because it was the first digital film we'd run for several weeks, and I was looking forward to hearing the upgraded sound system. New amplifiers had also been installed, and a faulty Cat. 441 card replaced.

Not everything is in digital, and not everything in digital is six channels. Strictly speaking, I think the two channel films are actually six channels, but with four of them silent, as the DMA doesn't see them as a two channel source and allow the matrix decoder to be used, so you only get sound out of two speakers.

I noticed something strange about the sound on 'Becoming Jane', which I think was the second digital film we ran. The lights on the DMA were showing nothing on the Ls and Rs channels, but there was a signal much of the time on the normally unused channels 7 and 8. Despite this there was sound on the surround speakers.

Posted by Mark J. Marshall (Member # 1409) on 08-27-2007, 03:41 PM:

CDS was out in 1990.

Posted by Brian Guckian (Member # 1678) on 08-27-2007, 04:38 PM:

quote: Stephen Furley
They ran Pearl & Dean adverts, though rather more than we do, and no less than five trailers, about 25 minutes in total, which seems a bit excessive; the feature wasn't much more than an hour and a half.

Many adverts in the UK and Ireland are still Dolby SR-only, and thus there will be several reversions during shows, not to mention if policy logos etc. are also in SR.

If reversions are very noticeable then it suggests the analogue chain is not being kept tuned [Frown]

Also, in Ireland at any rate, some of the Distributors have been releasing local short films before features for the last few years, many of which still have SR-only soundtracks.

quote: Stephen Furley
Not everything is in digital, and not everything in digital is six channels. Strictly speaking, I think the two channel films are actually six channels, but with four of them silent, as the DMA doesn't see them as a two channel source and allow the matrix decoder to be used, so you only get sound out of two speakers.

Interesting - is this an example of Dolby E? Or is it a standard Lt Rt track? (But if so, that would be 4.1, not 5.1, when decoded).

Can the DMA-8 matrix decode Lt Rt, and can it decode Dolby E as standard, or are all these features additional extras?

Also wondering why your installation isn't decoding the tracks - the films in question could hardly have been mixed only in L-R stereo?

Posted by Stephen Furley (Member # 1277) on 08-27-2007, 05:32 PM:

quote: Brian Guckian
Interesting - is this an example of Dolby E? Or is it a standard Lt Rt track? (But if so, that would be 4.1, not 5.1, when decoded).

Can the DMA-8 matrix decode Lt Rt, and can it decode Dolby E as standard, or are all these features additional extras?

Also wondering why your installation isn't decoding the tracks - the films in question could hardly have been mixed only in L-R stereo?

No, it's not Dolby E; it's a PCM track with sound on only L and R channels.

The DMA8 Plus has four digital inputs; Digital 1 has four AES pairs, on a 25 pin connector, for the digital cinema server, digital 2 and three are each one AES pair on 75 Ohm BNC connectors, and digital 4 is one pair on Toslink optical. All of these can be matrix decoded when fed a two channel input, but only the first pair on the digital 1 input.

Each of the inputs can be PCM, Dolby Digital or Dolby E. All of this is available as standard. For each of the three input formats there's a matrix decoder setting in the setup software which sets whether the input will be matrix decoded if it has two, no more and no less, channels. For PCM signals this can be set on or off. For Dolby Digital and Dolby E signals it can be set on, off or auto. When set to auto the decoder is turned on of off according to a flag in the meta-data of the encoded stream. On our DMA the decoder setting is on for PCM, and if I feed it a normal two channel PCM source the decoder does indeed turn on. When fed the track from 'This is England', it does not. The four AES pairs on the input 1 are assigned as L and R, C and Sub, Ls and Rs and two extra ones, currently unused. This means that the L and R channels which have sound are both on the first AES pair, and therefore, if seen as a two channel signal should be decoded, which they are not. I can only think that there is data on the other channels, albeit silent, and that therefore the DMA sees it as a full multi-channel source, and does not attempt to matrix decode it, but I don't know for certain if this is the correct explanation. I don't have the means to analyse the data stream and see what is actually there.

Whether this was intended to be a Lt and Rt matrixed track, and in error has been distributed in a form which cannot be decoded, or whether it is really intended to be just simple L and R, possibly for television release, I'm not sure. It's certainly not ideal, in a cinema auditorium, to have sound on only the left and right speakers.

I've just realised that I've got the part about the DMA matrix decoder settings slightly wrong; when manually set to be on, either 'Pro logic' or 'Pro logic II decoding can be chosen, rather than simply 'on'.

Posted by Brian Guckian (Member # 1678) on 08-27-2007, 05:44 PM:

Many thanks indeed for the info, Stephen. Clearly an extremely versatile unit and it's good to learn its features.

So it sounds like a mastering problem, or indeed that the film was intended to be reproduced only as L-R stereo. Stranger things have happened!

But it would be interesting to find out what the explanation is, error or no.

Posted by Stephen Furley (Member # 1277) on 08-27-2007, 06:11 PM:

I would have thought it was an error, it we hadn't had two films like it within a few weeks.

It can do a few other things as well; there's a six channel analogue pass-through input, intended mainly for digital sound on film via a Dolby DA-20 or DTS unit. It can control most models of Dolby CP, select the format on them, prevent them reverting to Analogue when in format 10 and the sound is from a digital source etc. There are a couple of pins which are to do with a subwoofer setting on a CP-200, but I don't understand exactly what that is about. Each of the five inputs can be selected by front panel button, or by a simple switch or relay contact, but other settings must be made by the setup software running on a pc via USB, or by ascii strings sent via Ethernet or RS-232 ports. It's only one U high, unlike the original DMA8, which is useful if you're short of rack space. As well as PCM, Dolby Digital and Dolby E there's also an 'aux' option which is reserved for future use. Exactly what they have in mind, if anything, I don't know.

There's a surround delay setting, as on a Cat. 150 and also a global delay for all channels, to keep the sound in sync with the picture; we need three frames worth of that to allow for the delay in the digital picture getting onto the screen. It's a nice unit, but not cheap.

Edit:

The other digital film we had with two channel sound was 'Night of the Sunflowers', a Spanish film.

[ 08-28-2007, 05:19 AM: Message edited by: Stephen Furley ]

Posted by Steve Guttag (Member # 268) on 08-28-2007, 06:34 AM:

The Dolby-E portion of the DMA8 Plus is not impimented yet but will be once the firmware comes out of beta test...

As to the ProLogic...it is very hard for a decoder should pass the signal as PCM or go to ProLogic...only on AC3 and like sources will there be information in the metadata to tell a decoder what to do. Furthermore, Cinemas calibrate their surrounds 3dB lower (82dBc/side) than everyone else (85dBc/side). The DMA-8 is supposed to look for flags to compensate for that but rarely do any sources have those flags set except conventional consumer sources like DVDs. On several installations now I have line amplifiers to deal with the 3dB difference and have been begging Dolby to provide for forcing the surround level change as a user option. Their concern being that it will ALWAYS be boosted since "more surrounds is better."

Most "professional" video tracks are indeed mixed 2-channel stereo as that is what they have in their video suites. However, that does not translate well to the big screen when 10s of feet exist between left and right speakers. I always force a prologic decode on such programs unless the person who made the video is present and throws a hissy-fit about it.

The DMA8 makes this a bit too complicated to do (need a computer or means of RS232 commands) which is why I prefer the Dolby DP564.

Stephen, are you using all of your inputs? If not, perhaps you could use a DA to send the 1/2 pair to both the 4xAES and to one of the remaining stereo inputs...set one to prologic all the time and leave the 4xAES discrete.

Alternately, you could use something that would spit out the necessary RS232 commands to control the DMA8 Plus. Look at Mediatech. They have button panels that will spit out whatever you program into them.

Steve

Posted by Stephen Furley (Member # 1277) on 08-28-2007, 04:01 PM:

quote: Steve Guttag
The Dolby-E portion of the DMA8 Plus is not impimented yet but will be once the firmware comes out of beta test...

I was not aware of that; the manual says nothing to indicate that it's not fully implemented yet. I don't have any Dolby E encoded material to test it with, and probably won't have in the future either.

quote: Steve Guttag
As to the ProLogic...it is very hard for a decoder should pass the signal as PCM or go to ProLogic...only on AC3 and like sources will there be information in the metadata to tell a decoder what to do. Furthermore, Cinemas calibrate their surrounds 3dB lower (82dBc/side) than everyone else (85dBc/side). The DMA-8 is supposed to look for flags to compensate for that but rarely do any sources have those flags set except conventional consumer sources like DVDs. On several installations now I have line amplifiers to deal with the 3dB difference and have been begging Dolby to provide for forcing the surround level change as a user option. Their concern being that it will ALWAYS be boosted since "more surrounds is better."

Yes, the decoder switching is going to be manual for PCM material, and ours is turned on, since if I feed it 'real' 2-channel PCM material, this is decoded. This is what leads me to think that in the case of the two 'problem' films, the DMA thinks it is getting a multi-channel track, even thoough only two of the tracks actually have any sound on them.

quote: Steve Guttag
Most "professional" video tracks are indeed mixed 2-channel stereo as that is what they have in their video suites. However, that does not translate well to the big screen when 10s of feet exist between left and right speakers. I always force a prologic decode on such programs unless the person who made the video is present and throws a hissy-fit about it.

If you drop the 's' off the end, and just make that 10 feet, thin you're not far off the size of our screen, but even at that size the 'hole in the middle' is noticeable from some seats.

quote: Steve Guttag
The DMA8 makes this a bit too complicated to do (need a computer or means of RS232 commands) which is why I prefer the Dolby DP564.

What is the DP564?

quote: Steve Guttag
Stephen, are you using all of your inputs? If not, perhaps you could use a DA to send the 1/2 pair to both the 4xAES and to one of the remaining stereo inputs...set one to prologic all the time and leave the 4xAES discrete.

Alternately, you could use something that would spit out the necessary RS232 commands to control the DMA8 Plus. Look at Mediatech. They have button panels that will spit out whatever you program into them.

Digital 2 is free at the moment, but likely to be used in future. Was it you, in the past who said that the digital 1 input is 110 Ohm balanced? If that is the case then we would need a matching transformer. Then there is the problem that somebody who only projects there occasionally my be confused by digital cinema content coming in on two different inputs. This might be worth considering if we get more films like this in future, but maybe it was an error, and we won't get any more.

A device to send RS-232 commands won't help, since the decoder 'on' setting has already been made; it's just that i doesn't work with this particular content.

While you're begging Dolby to do other things how about begging them for an option to select any one of the four pairs on digital 1, ignore the other three pairs, and put the chosen pair through the matrix decoder if you choose to do so. It should be a fairly minor software modification I would have thought. Also, they could allow whichever pair was selected to be decoded as Dolby Digital, or Dolby E, not just the first pair.

Posted by Steve Guttag (Member # 268) on 08-28-2007, 06:58 PM:

Forget it...it is apparently intrinsic to the design that the 4xAES input only works off of pair 1/2 for its magic. To the point that if pair 1/2 don't exist or are disconnect, the other lines will NOT pass through.

Yes, the 4xAES is 110-Ohm balanced...aka AES3...aka AES-EBU. You would need a DA in addition to the matching transformer...trust me, the DA was the pricy thing in the deal.

As for confusion, that is as easy as a label. On one you have "discrete channels" and the other you have "Pro-logic".

If you have the DMA8 set for prologic and it gives you discrete...would imply that it is seeing an AC-3 signal that is encoded 5.1 with only Left and Right active...a prologic decode is a prologic decode. Now it could be that if it senses audio on the other channels it presumes discrete (automatic).

The DP564 is Dolbys reference PCM decoder and one of their better products though it is "broadcast" and not generally available for cinema. It is a shame too since it is a marvelous piece that has alternate content application.

DP564 page at Dolby

The DMA8 Plus shipping firmware is 1.3.2.3. Neither Dolby-E nor 96KHz PCM are implemented yet. I have a unit waiting for the Dolby-E part before being installed...been waiting about a year now! The Dolby-E is in beta test now...I want them to get through with that before installing it...I don't need to find their problems on this particular installation. I hear the release is coming soon though.

Steve

Posted by Stephen Furley (Member # 1277) on 08-29-2007, 03:27 AM:

quote: Steve Guttag
Forget it...it is apparently intrinsic to the design that the 4xAES input only works off of pair 1/2 for its magic. To the point that if pair 1/2 don't exist or are disconnect, the other lines will NOT pass through.

That makes sense; I remember now that somewhere in the manual it says that the clock for all pairs is always taken from pair 1; so no pair 1, no clock.

quote: Steve Guttag
Yes, the 4xAES is 110-Ohm balanced...aka AES3...aka AES-EBU. You would need a DA in addition to the matching transformer...trust me, the DA was the pricy thing in the deal.

As for confusion, that is as easy as a label. On one you have "discrete channels" and the other you have "Pro-logic".

It might be an option to consider if we do get more films like this.

Posted by John Hawkinson (Member # 1135) on 09-23-2007, 07:56 PM:

Steve said, "The Dolby-E portion of the DMA8 Plus is not impimented yet but will be once the firmware comes out of beta test...."

On the off chance anyone who cared didn't know...:

quote: Dolby/Ken Jacquert

Date: Sun, 23 Sep 2007 17:29:14 -0700

For your information, there is a new production release of DMA8Plus firmware and setup software. We recommend that you download these latest files [...], install this latest setup software on your laptop, and upgrade your DMA8Plus units at your convenience.

The new DMA8Plus setup software is version 1.2.9.0. This version includes the latest fixes, sample rate display, support for setup software connections through a firewall, and includes a built-in "Dolby Software Update" to be used for upgrading your unit to the latest firmware version (via the "Action" menu).

The new DMA8Plus firmware is version 1.4.1.9. This version includes the latest fixes, 96kHz PCM (Tru96 processing), support for setup software connections through a firewall, and support for Dolby E decoding.

Posted by Steve Guttag (Member # 268) on 09-23-2007, 11:26 PM:

Not that Dolby wouldn't want the information passed on...but your quoted text John looks like it came from a private email which all Dolby emails have a proviso at the bottom that nothing in the email is to be disclosed publically.

I believe what is currently available is a "public beta" of what should become the release version for the DMA8 Plus. This is the last stage of beta to broaden the range of testers and see if they got all of the little things right.

Steve

Posted by John Hawkinson (Member # 1135) on 09-24-2007, 12:21 AM:

Wow, Steve, I almost never read those things, I'm afraid I've gotten inured to them. Thanks.

But in point of fact, it actually says, "If you are not the intended recipient, delete this message. If you are not the intended recipient, ..." So as long as you are the intended recipient, you are not enjoined from further distribution. I think that makes it fine in this case. (That is, the disclaimer is intended to deal with people intercepting e-mail, not to prevent the legitimate recipient from sharing it.)

--jhawk

Posted by Steve Guttag (Member # 268) on 09-24-2007, 08:41 AM:

Most any "technology" company puts restrictions on discussing their propritary information with others. Again, I doubt that Dolby would not want it known they have the DMA8 Plus fully functional.

After one has signed enough "NDAs" (Non-Disclosure Agreements) and are constantly dealing in non-released information, I tend to be very reserved in posting such information and instead have the manufacturer make such posts so everything is above board.

I think you are fine on this one though...as you pointed out...nothing said anything about non-disclosure of the info contained in the email.

Then again, there is this site's rule #6 about posting information in an email unless specific permission was given to post it (as opposed to specific instructions for non-disclosure). [Wink]

Steve

PS Cue markers?

Posted by David Graham Rose (Member # 1450) on 09-24-2007, 12:44 PM:

Quote Stephen Furley
"I noticed something strange about the sound on 'Becoming Jane', which I think was the second digital film we ran. The lights on the DMA were showing nothing on the Ls and Rs channels, but there was a signal much of the time on the normally unused channels 7 and 8. Despite this there was sound on the surround speakers."

Greetings

In the UK track 7 and/or 8 often has audio description narration for the visually impaire.

David