Inconsistency in Atmos playback on 5.1 systems

[BillTubes]

I have read often here in this forum of Atmos playback not having much surround content on a 5.1 system. It is has been commented that the Dolby system automatically changes to the speaker configuration that it plays back on. But why is the immersive quality so inconsistent?

In my experience most of the Atmos content is very immersive when played on my 5.1 system, but other content has issues. It doesn’t seem to be random.

For example Steven Wilson mixes always sound great when played on my system like the latest Moondance. I remember reading an interview where he mentioned that he plays his Atmos mixes on a 5.1 bed to check the surround content and makes adjustments for a satisfactory experience. I am not sure what knobs he turns to make this happen. Other Atmos mixes sound wrong on the my 5.1 playback. My example here is Giles Martin’s Beatles Revolver and Red Album. I had to turn up the volume on the rears 10db to get proper balance. Since these are streaming from the Apple 4k, I was not sure if this is part of the problem.

My question is this;

Does the mixing engineer assign where channels will fold down for different speaker configurations ?

If any of you knowledgeable folks can shed light on this it would be greatly appreciated.

Thanks, Bill

 

[LuvMyQuad]

Myself and a friend are looking into how well Atmos downmixes. Specifically for 5.1 and Quad playback, is it better to use a 5.1 substream if available, or downmix the 7.1 using an AVP. We prepared a 1st test for listening and evaluation.

Below are screenshots of the 7.1 (top) and 5.1 substream (bottom) tracks from the Peter Gabriel IO album. Both have been ripped to MKV and then to FLAC using Audiomuxer. Using make MKV, the disc the 7.1 identifies as Dolby True HD. The 5.1 stream identifies as AC3. Both are 24/48kHz.

Can someone explain why the 7.1 tracks are smaller in size than the 5.1 tracks? I would have expected the opposite.

7.1

5.1

 

[ssully]

Was the AC3 stream decoded to PCM?

 

[LuvMyQuad]

Was the AC3 stream decoded to PCM?

I'll check, but I think so.

 

[ssully]

Given the file sizes, I would predict yes.

 

[BillTubes]

My file sizes for IO 7.1 are almost the same as yours. I ripped to MKV then used DVD Audioextractor to make flac files. On my system the 7.1 files are played on foobar and data sent to my OPPO 105 where the 8 channels are output to 6 Analog channels.
Maybe the conversion from 7.1 to 5.1 is another variable.

BTW the Atmos mix for this Peter Gabriel BluRay sounds very good on my 5.1 system.

 

[LuvMyQuad]

My file sizes for IO 7.1 are almost the same as yours. I ripped to MKV then used DVD Audioextractor to make flac files. On my system the 7.1 files are played on foobar and data sent to my OPPO 105 where the 8 channels are output to 6 Analog channels.
Maybe the conversion from 7.1 to 5.1 is another variable.

BTW the Atmos mix for this Peter Gabriel BluRay sounds very good on my 5.1 system.

That's all good news then going forward.

Given the file sizes, I would predict yes.

And the answer is yes they were converted to PCM. They would have needed to be processed to FLAC. You seem to understand the file size question. Is it because the lossy AC3 stream gets padded out in a conversion to lossless FLAC while the 7.1 stream remained lossless? Is it something else?

 

[ssully]

That's all good news then going forward.

And the answer is yes they were converted to PCM. They would have needed to be processed to FLAC.

Not really. AC3 and DTS in FLAC is possible without decoding. Audiomuxer has a tool for doing this. It puts AC3 and DTS files in a SPDIF/WAV wrapper , which can then be converted to FLAC (which does not reduce file size, but does allow excellent tagging for use in media players).

You really only want to do this if you intend to be 'purist' about your files -- i.e., keeping them in their original bitstream format, letting the decoding happen every time you play them (using, e.g., the decoders in your AVR).

I eventually I got over that tic, once I convinced myself there was no sonic difference. Now I decode to PCM/FLAC files as I rip (but I note the lossy origin in tags).

 

[ssully]

As for the size question, I can't say why your decoded 5.1 files are larger than your decoded 7.1 files, without analyzing the files themselves.

Your 7.1 file sizes match the files I ripped from my I/O BluRay disc btw.

 

[LuvMyQuad]

.
Maybe the conversion from 7.1 to 5.1 is another variable.

Yes. I'd like to find out. Which "mixdown" works best for 5.1 playback... The converted AC3 stream or the lossless stream that is hardware dependant. And how much hardware dependant is it?

@B&W Driver is working on getting it down to quad. I think Ralphie @4-earredwonder still runs quad on one of his rigs

My son needs the 5.1 for his system.

I can only play back 5.1.4.

 

[B&W Driver]

As this thread seems to be all about the same questions I've been puzzling over - namely, how do Atmos mixes sound on a quad or a 5.1 system - I'm posting here to facilitate following the discussion. A very informative thread, indeed!

 

[HomerJAU]

Atmos playing on a 5,1 or 7.1 can sound great BUT like every other surround format it depends on the mix.

Many Atmos mixes released for streaming are ordinary and done by inexperienced audio engineers (surround mixing wise). There have been reports that some mixes are authored with the engineer on headphones alone, the mix is never heard on a real Atmos system during the mixing process.

 

[AYanguas]

My question is this;

Does the mixing engineer assign where channels will fold down for different speaker configurations ?

If any of you knowledgeable folks can shed light on this it would be greatly appreciated.

Thanks, Bill

My guess is that the substreams 2.0 / 5.1 / 7.1 that are generated by the render tool from the Atmos master to the final consumer file are "automatically" processed by the Dolby Algorithms. Apple providing different algorithms for its Spatial Audio.

The mixing engineer has very little or no control of that "fold down" process. So if the downmixed substreams does not sound very good to the "ears" of the mixing engineer, he has two options:

1) Tweak the Master Atmos mix to give a fold down that would sound better. If he tweeks for binaural headphones 2.0, then the resulting 5.1 would be "unpredictable" (to say something).
2) If the Master Atmos sounds good for Speakers, and he does not want to change it, then make an additional separate traditional 5.1 mix.

Obviously, the second option is "more expensive", and some engineers/producers limit that work either due to cost or inexperience.

With respect to the @LuvMyQuad investigations, I think they would eventually result in the discovery of patterns that certain sound placements or pannings that work well with speakers don't work well in the automatic 5.1 fold down. While others do.

Experienced mixing engineers, I suppose, should have learned about these patterns and make their mixes with those conditions if what they want is to facilitate a "fold down" instead of making a "good" Atmos mix for speakers. Or maybe they just do not have a full speaker system in his mixing studio, and check the Atmos mix with headphones.