Film-Tech Cinema Systems
Film-Tech Forum


Post New Topic  Post A Reply
my profile | my password | register | search | faq & rules | forum home
  next oldest topic   next newest topic
» Film-Tech Forum   » Operations   » Digital Cinema Forum   » Digital Cinema Audio 101 (Page 1)

 
This topic comprises 2 pages: 1  2 
 
Author Topic: Digital Cinema Audio 101
Armand Daiguillon
Film Handler

Posts: 12
From: Plantation FL USA
Registered: Jan 2018


 - posted 01-21-2018 01:34 AM      Profile for Armand Daiguillon   Email Armand Daiguillon   Send New Private Message       Edit/Delete Post 
Hello all,

I was hoping someone could help shed some light for me on some basics of how today's digital cinema works.
Specifically with audio.
Back in the day if you had a victoria 5 projector with super cool SDDS and were having trouble ... I could help.
But i've been out of the game for a quite a while and some of the new fundamentals seem to escape me.
So any light anybody could shed would be greatly appreciated.
Here is what I don't get.
I've been looking at Barco projectors with integrated media servers and I just don't get the audio.
First they use rj45 plugs for the aes/ebu outputs rather than db25 which seems to be what everybody else uses - but what I dont get is why use either ?
Why is it setup to output 8 channels (or more) to begin with ?
The projector/media server does not decode the digital to analog so why so many outputs ?
My home stereo system has one SPDIF (which from what I read uses basically the same spec as aes/ebu) that carries the digital signal and it decodes from that one cable all the channels.
So if its not decoding the digital, what exactly comes out the 8 channels aes/ebu output of this projector/media server ?
I assume most dcp's have several audio formats encoded - 2 channel stereo, and Dolby Digital (5.1) maybe another 7.1 track ?
(Is DTS still even a thing ?...)
So if the dcp has several audio tracks what does it output through those aes/ebu 8 channels if it has no decoder built in ?
Is Stereo on 1 set of two pairs and then different digital tracks on the others ?
Does it go to a Dolby processor and if it doesn't "read" a dolby digital signal on any of the 2 pairs does it fall back
to some "pass through function" of just 2 channel ?
Is the aes/ebu output on the projector sending a "generic" pcm signal on some channels but and encoded digital formats on others ?
I'm just kinda lost ..I've looked all over the internet but can't find anything that explains exactly how this process works nowadays.
The way I see it why have 25 pin outputs at all ? If AES/EBU is the same spec as SPDIF why not just have one
audio out cable SPDIF from the projector/media server to the audio processor ? The newest spec can hanle 64 channels on one 2 pair ?
And why even need a separate audio processor - why not just have the digital to analog converter built in with direct outputs to your amps ?

I realize its alot of questions but I just dont see the logic in the setup as is ...and i'm assuming its because I dont get the true picture of how it works.
So any light anybody can shed on this would be greatly appreciated.

Thanks !!

 |  IP: Logged

Stephan Shelley
Jedi Master Film Handler

Posts: 675
From: castro valley, CA, usa
Registered: Nov 2014


 - posted 01-21-2018 02:12 AM      Profile for Stephan Shelley   Email Stephan Shelley   Send New Private Message       Edit/Delete Post 
DCI specifies 16 channels of uncompressed digital audio. Not all the channels have been officially assigned. Each AES pair can carry 2 channels of digital audio. Channels 1/2 are L/R, 3/4 C/Sub, 5/6 Ls/Rs, 7/8 HI/VI, 11/12 BLS/BRS for 7.1 surround. Dbox uses 13 and Dolby Atmos uses 14 for a sync signal. Some places were using 15/16 for HI/VI. GDC has a 16 channel system. Yes DTS has an object orientated system.

 |  IP: Logged

Armand Daiguillon
Film Handler

Posts: 12
From: Plantation FL USA
Registered: Jan 2018


 - posted 01-21-2018 03:13 AM      Profile for Armand Daiguillon   Email Armand Daiguillon   Send New Private Message       Edit/Delete Post 
>DCI specifies 16 channels of uncompressed digital audio. Not >all the channels have been officially assigned. Each AES pair >can carry 2 channels of digital audio. Channels 1/2 are L/R, >3/4 C/Sub, 5/6 Ls/Rs, 7/8 HI/VI, 11/12 BLS/BRS for 7.1 >surround. Dbox uses 13 and Dolby Atmos uses 14 for a sync >signal. Some places were using 15/16 for HI/VI. GDC has a 16 >channel system. Yes DTS has an object orientated system.

First, thanks for the response, I get some of it ...
Here's what I dont get, If each of these pairs is uncompressed digital, is it pcm ?
If its not decoding the dolby or dts, where is it getting the other channel data from ? Or are these dcp's encoded with multichannels in a "standard" digital format ? And if thats the case why do you need dolby/dts at all ?
I guess what i'm most unclear is where the dolby or dts specific signal lies ?

 |  IP: Logged

Marco Giustini
Film God

Posts: 2539
From: Reading, UK
Registered: Nov 2007


 - posted 01-21-2018 05:56 AM      Profile for Marco Giustini   Email Marco Giustini   Send New Private Message       Edit/Delete Post 
Indeed there is no Dolby/DTS anymore in Digital Cinema - this is if you exclude Dolby Atmos.

5.1 and 7.1 are on the DCP in uncompressed PCM form, 16 channels of them. Those 16 channels are then wired to the sound processor (usually only 6 or 8 of them) which does the EQ and routing - NO decoding, unless you consider converting digital PCM in analogue sound as "decoding".

In theory you could make your own sound processor by using any professional device capable of accepting PCM over an AES connection - you would need at least 3xAES to make a 5.1 sound as each AES pair will carry 2 channels.

RJ45 is being used more and more by manufacturers as it seems to be an easy way to connect devices and, apparently, a cat5 cable is very good at carrying AES sound. No need to solder or make your own cable. But it's just the form factor, it's the same AES being used by others via DB25.

 |  IP: Logged

Armand Daiguillon
Film Handler

Posts: 12
From: Plantation FL USA
Registered: Jan 2018


 - posted 01-21-2018 01:31 PM      Profile for Armand Daiguillon   Email Armand Daiguillon   Send New Private Message       Edit/Delete Post 
Interesting, so does that mean that dcp's are more limited in scope than other formats ?
An .mkv video file can carry multiple audio tracks (for different languages) using AC3 (Dolby) or AAC.
It would seem you are saying all dcp files are "standardized" to pcm and have just one audio track (7.1) available ?
And basically dolby digital and dts as I knew them back in the day are basically obsolete concepts in digital cinema ?
That would explain alot .....

 |  IP: Logged

Lyle Romer
Phenomenal Film Handler

Posts: 1380
From: Davie, FL, USA
Registered: May 2002


 - posted 01-21-2018 01:55 PM      Profile for Lyle Romer   Email Lyle Romer   Send New Private Message       Edit/Delete Post 
"Back in the day" the digital sound systems were a means to get a digital sound track played in sync with 35mm film. The storage method for digital audio didn't "match" the analog method of storing the picture on film.

Storing digital sound on film (like Dolby and SDDS) had limits of bandwidth due to the practical printing of small dots that could be reliably read.

DTS was limited in bandwidth because when it was designed, hard drives were expensive and DVDs didn't exist yet. For economy they needed to put the soundtrack on CD or CD-ROM. Since uncompressed audio on a CD would take 2 discs for every 2 channels (for most movies). 6 track would require 6 discs (and therefore 6 CD-ROM drives). That would have made the player more expensive and created complexity in keeping all of the discs from getting misplaced.

I'm not even sure if the 386 (I think, maybe it was 486) based PC that was the original player could have played multiple audio CDs simultaneously. Using the compressed format, the PC just had to handle one data stream and output it to their decoder hardware.

With digital cinema, storage of both audio and video are on the same media so you don't need a "system" to get digital audio with digital video. Also, storage is cheap so it is practical to ship both together on a hard drive or download the file via satellite or other means.

 |  IP: Logged

Harold Hallikainen
Jedi Master Film Handler

Posts: 738
From: Denver, CO, USA
Registered: Aug 2009


 - posted 01-21-2018 02:16 PM      Profile for Harold Hallikainen   Author's Homepage   Email Harold Hallikainen   Send New Private Message       Edit/Delete Post 
A single digital cinema package MAY have multiple soundtracks, though this is rare. Each movie (called a "composition") has a Composition Playlist that tells what files will be played at what time during the playback of the composition. Movies are still broken into reels where each reel has an image file and an audio file that runs about 20 minutes. The CPL ties these all together. There may be multiple CPLs in a digital cinema package. There may be a CPL for 5.1 sound, another for 7.1 sound, etc. DCPs also can carry subtitles and closed captions as separate "timed text" files.

With digital cinema, the storage space and bandwidth are available to use uncompressed audio, so no encoding or compression is required. The standards also try to stay away from proprietary technology as much as possible. Thus, the audio is done as a single file with up to 16 channels in it, with a sample depth of 24 bits and a sample rate of 48 kbps. The standards allow for 96 kbps, but, as far as I know, only one movie has been distributed using 96 k audio.

In a previous post, there was mention of object-based audio such as Dolby Atmos or DTS-X. In a typical "channel-based" audio system, such as 5.1 or 7.1DS, the mixer pans audio fragments to drive combinations of the available speakers or arrays of speakers (surrounds are arrays). The individual feeds to each speaker or array are recorded as 6 or 8 channels of audio. These audio channels are played back in the auditorium to speakers in corresponding locations to reproduce an approximation of what was heard on the dub stage.

In object-based audio, the individual audio fragments (mono audio tracks) along with panning instructions (and other information, such as decorrelation, object-size, etc.) are recorded. In addition, a channel-based mix is also recorded. Most sound is handled in the channel-based mix and then objects are added on top of this "bed." The audio is broken into frames, just as the image is. So, there are typically 24 chunks of audio per second. Each of these chunks are individually playable. They include all the audio and all the metadata necessary to play that chunk. A renderer looks at each bed or object "element" in the frame. This element includes panning instructions, etc. for objects, and includes speaker destinations for the bed. The elements also include "pointers" to the referenced audio elements (mono audio for 1/24 second). There can be a total of up to 128 object elements and bed channels in a frame (assuming 48k audio). In the playback system, a "renderer" evaluates the metadata in the bed and object elements and combines the referenced audio according to those instructions.

Summarizing, in channel-based audio, the speaker feeds are recorded. The recorded feeds are sent directly (with equalization, etc.) to the corresponding speakers. In object-based audio, audio fragments and panning instructions are recorded. The renderer performs the panning instructions on the audio and sends it to the corresponding speakers (typically up to 64 individually driven speakers, but there is technically no limit).

Harold

 |  IP: Logged

Dave Macaulay
Film God

Posts: 2105
From: Toronto, Canada
Registered: Apr 2001


 - posted 01-21-2018 03:56 PM      Profile for Dave Macaulay   Email Dave Macaulay   Send New Private Message       Edit/Delete Post 
Yes multiple channels are possible on a single digital audio data signal.
The digital cinema audio standard is to use uncompressed PCM audio, so you get two discrete audio channels on a single AES signal.
The original standard use AES signal 1=L&R, 2=C&SW, 3=LS&RS. The 7.1 rear channels BLS&BRS are on 6 (ch 11 & 12) because #5 was already assigned to HI/VI before 7.1 was developed.
Many processors can decode ProLogic multichannel from the L/R single AES signal but this is not used for DCP playback - mostly for a Blu-Ray sending HDMI to an IMB type system - usually because most media blocks and all Dolby processors don't speak DTS... CP750 processors will (I think) decode Dolby Digital compressed multichannel audio as well, the CP850 will even decode Dolby Atmos and TrueHD from alternate content sources that have these formats recorded.
But most cinemas only play DCPs, and all DCPs have 5.1 or 7.1 sound on 6 or 8 channels (3 or 4 AES signals). Those few movies that are mono or stereo only use the relevant channels with the rest silent, although all are typically still played to the auditorium as processors tend to only have 5.1 or 7.1 inputs and settings. Digital silence is very silent so you only hear the system's quiescent noise.
DCPs actually have 8 AES pairs - 16 audio channels - available. Some of the "extra" ones carry HI/VI assistive audio and some are reserved for ATMOS or other immersive sound schemes and for DBOX seat movement data.
All audio channels will have watermarking added by the server media block. The other data channels are not watermarked.

 |  IP: Logged

Steve Guttag
We forgot the crackers Gromit!!!

Posts: 12134
From: Annapolis, MD
Registered: Dec 1999


 - posted 01-22-2018 07:10 AM      Profile for Steve Guttag   Email Steve Guttag   Send New Private Message       Edit/Delete Post 
Pair 4, originally, was for LC/RC and pair 8 was for HI/VI. However, early players/systems were only set up for 8-channels (4 pairs) so HI/VI quickly moved to pair 4, forcing LC/RC to move to pair 5 leaving BLs/BRs to take pair 6. Pair 7 seems to have been usurped by proprietary audio control information and pair 8, officially, remains an HI/VI track. So, in a sense, all 16-channel are now spoken for though pair 8 could/should be reassigned and pair 7 should probably be done a better way (not use audio channels for control work). In fact, I think HI/VI shouldn't consume full bandwidth audio channels. They need the quality of a telephone, not a Hi-Fi system.

 |  IP: Logged

Harold Hallikainen
Jedi Master Film Handler

Posts: 738
From: Denver, CO, USA
Registered: Aug 2009


 - posted 01-22-2018 09:59 AM      Profile for Harold Hallikainen   Author's Homepage   Email Harold Hallikainen   Send New Private Message       Edit/Delete Post 
The current ISDCF recommended assignment list is at http://isdcf.com/papers/ISDCF-Doc4-Audio-channel-recommendations.pdf . Note the addition of sign language on channel 15. This is VP9 encoded video as described in the draft document at http://isdcf.com/papers/ISDCF-Doc13-Sign-Language-Video-Encoding-for-Digital-Cinema.pdf . Ideally this would be an aux data track, but equipment in the field cannot yet deal with aux data tracks.

On full range audio for HI/VI, we often use an LCR mix generated in the sound processor for HI since not all compositions have an HI track (especially trailers). This generally works well, but there have been some features where one character's dialog is only in the surrounds.

We've also experimented with putting VI audio on the LFE track. Since LFE has a very sharp cutoff at 125 Hz (SMPTE specified filter), 200 or 300 Hz on up is available for VI audio.

Harold

 |  IP: Logged

Sean McKinnon
Phenomenal Film Handler

Posts: 1564
From: Peabody Massachusetts
Registered: Sep 2000


 - posted 01-29-2018 08:46 PM      Profile for Sean McKinnon   Email Sean McKinnon   Send New Private Message       Edit/Delete Post 
One cool thing you can do with Q-Sys (or symmetrix or other DSP) is create a ducking scheme where if there is nothing coming in on the HI track you can send a LCR mix down for your alt content (and send a mic mixer output or whatever) but if there is signal on HI it will duck the other sources and just pass the HI track. You can also mix the output of that HI ducker with your VI information in a less extreme ducking set up to gently duck the HI information a few db when there is narration so the listener gets a nice mix but the VI info is still clear and easy to pick out.

 |  IP: Logged

Harold Hallikainen
Jedi Master Film Handler

Posts: 738
From: Denver, CO, USA
Registered: Aug 2009


 - posted 01-29-2018 09:35 PM      Profile for Harold Hallikainen   Author's Homepage   Email Harold Hallikainen   Send New Private Message       Edit/Delete Post 
Very clever! The JSD-100 compares the HI track to the LCR mix. If the LCR mix is something like 20 dB louder than the HI track, it uses the LCR mix. Using compressors in Q-SYS is clever!

Harold

 |  IP: Logged

Bill Brandenstein
Master Film Handler

Posts: 348
From: Santa Clarita, CA
Registered: Jul 2013


 - posted 01-30-2018 03:33 PM      Profile for Bill Brandenstein   Email Bill Brandenstein   Send New Private Message       Edit/Delete Post 
quote: Harold Hallikainen

Thus, the audio is done as a single file with up to 16 channels in it, with a sample depth of 24 bits and a sample rate of 48 kbps. The standards allow for 96 kbps, but, as far as I know, only one movie has been distributed using 96 k audio.

Just for the record, Harold, this looks like a typo; I think you meant to say "sample depth of 24 bits and a sample rate of 48KHz. The standards allow for 96KHz..."

And Harold, having never heard of a movie with so-called high resolution / 96KHz audio, would appreciate the title, if possible!

 |  IP: Logged

Harold Hallikainen
Jedi Master Film Handler

Posts: 738
From: Denver, CO, USA
Registered: Aug 2009


 - posted 01-30-2018 05:42 PM      Profile for Harold Hallikainen   Author's Homepage   Email Harold Hallikainen   Send New Private Message       Edit/Delete Post 
Thanks for fixing my typo! As I recall, the movie with 96k audio was released by Sony several years ago. Searching back through email, I can't find the title.

Harold

 |  IP: Logged

Martin Brooks
Jedi Master Film Handler

Posts: 804
From: Forest Hills, NY, USA
Registered: May 2002


 - posted 01-30-2018 06:26 PM      Profile for Martin Brooks   Author's Homepage   Email Martin Brooks   Send New Private Message       Edit/Delete Post 
He might be thinking of Blu-ray release. "The Right Stuff" was released with a 96KHz audio sampling rate, but initially, the discs were created at 48KHz. They re-pressed them and if you wrote to them, they sent you a new disc, but frankly, I can't hear any difference between the two versions and I find the sound on this Blu-ray to be quite terrible in any case - it lacks fidelity and impact. It almost sounds as if the tracks came from the Dolby Optical and not from the 70mm 6-track mix.

But aside from the fact that the soundtrack doesn't sound great, there's no reason for it to sound better at 96KHz than 48KHz, because it's very unlikely that there's any audio above 24KHz anyway. And if there isn't, the increased sampling rate will make absolutely no difference.

 |  IP: Logged



All times are Central (GMT -6:00)
This topic comprises 2 pages: 1  2 
 
Post New Topic  Post A Reply Close Topic    Move Topic    Delete Topic    next oldest topic   next newest topic
 - Printer-friendly view of this topic
Hop To:



Powered by Infopop Corporation
UBB.classicTM 6.3.1.2

The Film-Tech Forums are designed for various members related to the cinema industry to express their opinions, viewpoints and testimonials on various products, services and events based upon speculation, personal knowledge and factual information through use, therefore all views represented here allow no liability upon the publishers of this web site and the owners of said views assume no liability for any ill will resulting from these postings. The posts made here are for educational as well as entertainment purposes and as such anyone viewing this portion of the website must accept these views as statements of the author of that opinion and agrees to release the authors from any and all liability.

© 1999-2018 Film-Tech Cinema Systems, LLC. All rights reserved.