Cinema DTS - main thread

[spoRv]

Cinema DTS - main thread

Documents:

APT-X100: a low-delay, low bit-rate, sub-band adpcm audio coder for broadcasting (PDF)
https://mega.nz/file/bgtVmT4C#Zf5OTIEVi2...sqCi5DtvIE

Updated: 2020-04-29

When useful links will be posted in this thread, they will be added to the first post.

[CSchmidlapp]

[CSchmidlapp]

Hi All

So lets say.....
We have a Cinema DTS disc (Pre and post 1999)
We have Foobar2000 installed with the latest apt-x100 plugin (Version 0.1.10).
Default settings.
We convert all the reels to a 5.1.wav, with the LFE crossover handled by the apt-x100 plugin.

Now what?

The documentation refers to channel volume adjustments (Based on Pre and Post 1999 discs)
There is more information here https://fanrestore.com/thread-71.html
In this thread https://fanrestore.com/thread-2539.html
And in this one https://fanrestore.com/thread-501.html

Presumably there is speed change needed if syncing with a Bluray (24 to 23.976)?

And a resample to 48000kHz to bring it into BD spec's if needed.

[little-endian]

Oh yes, there are too many different threads already for effectively too little discussion, indeed.

As for the speed change, the correct answer just like lawyers love to say, is "it depends", as while most Blu-rays run at 24/1.001 fps, some European video masters come along with 24 fps straight, so be aware.

For me, the technical "cleanest" way is to leave the Cinema DTS as it is (44.1 kHz / 16 Bit LPCM once decoded) and reflag the video to also run at 24 fps and then muxing all that together after having synced "the shit out of it", so to say. Stretching the audio to match 24/1.001 and also resample it should still be transparent and inaudible, of course but with nowadays players, I seems unnecessary to desperately stick to any Blu-ray specs.

[CSchmidlapp]

Oh yes, there are too many different threads already for effectively too little discussion, indeed.

As for the speed change, the correct answer just like lawyers love to say, is "it depends", as while most Blu-rays run at 24/1.001 fps, some European video masters come along with 24 fps straight, so be aware.

For me, the technical "cleanest" way is to leave the Cinema DTS as it is (44.1 kHz / 16 Bit LPCM once decoded) and reflag the video to also run at 24 fps and then muxing all that together after having synced "the shit out of it", so to say. Stretching the audio to match 24/1.001 and also resample it should still be transparent and inaudible, of course but with nowadays players, I seems unnecessary to desperately stick to any Blu-ray specs.

Absolutely, the fact we able to get access to the original theatrical discs is pretty amazing.
Being able to decode and play with the files on our home pc's is a bit of a geeks dream.

The most important aspect of all this is the channel volume question.
From what I'm gathering from the limited material spread across this forum and the wider interwebs, is that channel volume has to be altered depending on a few variants.
The measurement in db is not the same in an NLE to that of a theater amplification system?
So the +10db applyed to the the LFE channel in the documentation does not translate to me just punching up the LFE volume in premiere does it?

There are quite a few films in my collection (mainly from Hong Kong!) that are 24fps, but the majority (most of my collection are from Hollywood or there European distributers ) run at 23.976.
I'm personally putting together versions that hold legacy mixes, like many here, The Cinema DTS will share the same file as the other mixes from a multitude of sources like Laserdisc.
Due to the syncing process, The CDTS track is going to be encoded in a lossless codec due to it coming from a lossy one in the first place.
Id say 23.976 is most likely everybody's finishing FPS.

The up-sample to 48kHz is a good question. As you pointed out sticking to Blu-ray spec considering the reality of file based media seems pretty mute.

[CSchmidlapp]

For anybody starting out like me Ive done the following to get files for syncing in premiere...

Using Foobar2000 with the latest apt-x100 plugin (Version 0.1.10).
I convert all the reels to a wav on Default settings (the LFE crossover handled by the apt-x100 plugin).
Giving me a 16bit, 44100kHz, 6 channel (5.1) wav (repeated for each reel).

I then used Eac3to to slowdown to 23.976 and resample to 48kHz.

I use the UsEac3to GUI, so I loaded the file and adjusted the settings.
The command line parameters were...

Code:

%_.wav -24.000 -changeTo23.976 -resampleTo48000 -down16

again repeated for each reel.


For the record, in this particular case (which is perfect) the BD looks as though it uses the theatrical mix.
Comparing the transcoded Cinema DTS files to the Bluray 5.1 mix while syncing, it looks like all the channel levels are overall correct.
The LFE channel needed no adjustment and matched.
The rears seem a touch louder on the CDTS but nothing to drastic.

The overall tone of the CDTS is 'walmer' though, with the BD version sounding more clear in the high end.
I think maybe this could be EQing on the BD's master, or most probably that it comes from a higher quality / lossless encode.

[deleted user]

I personally prefer doing the sync at 24 fps (retiming the Blu Ray first) so that the sync is done losslessly and the resampling for the final output. That way the raw sync is archived and could also be resampled with a better resampler later. Personally I use iZotope SRC 64 Bit for that. Here's a comparison of resampler quality:
https://src.infinitewave.ca/

[little-endian]

For the record, in this particular case (which is perfect) the BD looks as though it uses the theatrical mix.

What movie did you compare in particular?
As stated in the other thread, when comparing the audio of the Cinema DTS with the Blu-ray counterpart, for these at least, the bass consistently sounds "thicker" on the Blu-ray:

Armageddon
Men in Black
Collateral
Hackers

With Gravity, which I recently synchronized, it's difficult to say as the mix is surprisingly quite different.

The LFE channel needed no adjustment and matched.

Well, in the case of Men in Black, the effective LFE (which in the case of Cinema DTS is really a confusing term as there is none) levels differ:

Men in Black, first ~ 14 minutes (CDTS above, BD below)



Also a member here (who unfortunately isn't too keen on participating in any bigger discussion) was in contact with someone dealing with Cinema DTS professionally and that certain individual claimed the foobar plugin's decoding result to be wrong level-wise whereas the lack of any detail about what should be wrong in particular, doesn't exactly make things easier. 

I compared the result from that person with the supposedly correct levels with the foobar result and indeed, the LFE is a bit higher in his, but somehow, the BD mix still has a stronger bass with my setup (which certainly is not a reference).

The overall tone of the CDTS is 'walmer' though, with the BD version sounding more clear in the high end. 
I think maybe this could be EQing on the BD's master, or most probably that it comes from a higher quality / lossless encode.

I'd point the "most probably" to the EQing and not any codec differences, as I consider ADPCM pretty much being transparent here, including AC3 and DTS at Blu-ray's core bitrates (and LaserDiscs with AC3 at only 384 kbps for 5.1 proves that the world doesn't quite end here either). But of course, the power of marketing can be quite strong and so anything but lossless and at least 192 kHz / 24 Bit is coming straight from hell.

Maybe borisanddoris can elaborate on that because having read several articles and discussions about bass management and channel levels in general, now I'm confused where the only 6dB difference comes into play depending on the movie release date if one is supposedly calibrated to 85dBSPL and the other to 91dBSPL as mentioned here. Maybe I misunderstand something here as if all AVRs at home effectively apply a 10dB gain for the LFE, wouldn't be the bass of Cinema DTS tracks then be playing 4dB too high if 91dBSPL is the reference here and not 95dBSPL? Doesn't match my experience so we're back to what the foobar plugin does when setting the LFE reconstruction to "0dB".

[CSchmidlapp]

@little-endion The movie Ive been working on is Timecop.
Ive not done an extensive study, just by ear and a quick look at the waveform while syncing.
I do prefer the BD mix in this case.
The CDTS does sound alittle more 'compressed' if that's the right word.
Unsure where that is coming from hence my questioning of EQ and Codec Compression.
Would the encoding engineers compensate for the deficiency's / characteristics of a particular codec compression method?
Ive long discussed with members here about the level colorists grade to the characteristics of a given 35mm film release print type.

@TomArrow. Thanks for your input bud, and yes I think Ill follow that method in the future.

[little-endian]

Ah, I guess I've to get some Blu-ray source of Time Cop then and compare as well. There might be different Blu-ray releases of that movie, so just for the protocol - which one have you checked?

The problem in principle here is that we simply don't know the original. Could be, that they did a remix for the Blu-ray or theoretically that they had a better source for the later Blu-ray release (odd, but possible).

For instance, the original CDTS of Jurassic Park from 1993, while considered to be a very good mix, the voices have some distortions which interestingly don't occur on the BD release which also raises the interesting question whether they "cleaned" it up using the same masters which were used to encode the CDTS and thus that one being just the way it was originally produced, or they had access to better sources when they encoded the DTS-HD MA track for the Blu-ray.