Dolby Atmos decoding hardware ???

[mkt]

I saw this mentioned on ASR "This software is able to convert 3D surround sound to formats that can be played on your PC. What was only possible with expensive AVRs before, like true Dolby Atmos playback, can be done for free with Cavern"
I don't understand it but it looks interesting?

 

[mkt]

I saw this mentioned on ASR "This software is able to convert 3D surround sound to formats that can be played on your PC. What was only possible with expensive AVRs before, like true Dolby Atmos playback, can be done for free with Cavern"
I don't understand it but it looks interesting?

And videos?

 

[humprof]

I'm a little slow...technically speaking, Atmos can already be played back on any cheap notebook PC--and if it has an HDMI port, you can route it to your AVR, to boot. I guess they mean this technology will give you Atmos-y effects on your two laptop speakers alone? I'll confess that parts of that YouTube video sound 3D-ish on my ancient Lenovo, but not all of it. Certainly not as good as, say, Apple Spatial Audio on my MacBook Pro. I need more info.

 

[JediJoker]

...Atmos can already be played back on any cheap notebook PC--and if it has an HDMI port, you can route it to your AVR, to boot.

Yes, but it can't be decoded by most operating systems natively, so it's no mean feat to have Atmos play back using a computer audio interface (with sufficient output channels) and no dedicated hardware decoder (like an AVR).

 

[mkt]

Yes, but it can't be decoded by most operating systems natively, so it's no mean feat to have Atmos play back using a computer audio interface (with sufficient output channels) and no dedicated hardware decoder (like an AVR).

And without the $400 dolby software decoder. Cavern seems to have an upmixer and room eq.

 

[VoidX]

Atmos can already be played back on any cheap notebook PC

That title was meant to show the decoder's performance. Dolby is using dual Xeon processors in their CP850, I did the same render (DCP Atmos) in a single entry-level notebook core (Intel 5200U). While Dolby made great efforts to make DD+ really heavy on a CPU, pushing users to dedicated hardware, it's also running really well, about 8 seconds of DD+ can be decoded per second on a high-end PC. This is really far from the regular DD's 40x performance, and on cheaper PCs, real-time Atmos needs more optimizations to run next to a video, it's well within reach.

I guess they mean this technology will give you Atmos-y effects on your two laptop speakers alone?

This video has Dolby Atmos for Headphones audio. While I also have my own virtualizer, I'm not advertising that since it was made only for my ears. YouTube can't have Atmos audio, so this was the closest. Even if it could use any home codec, this is 8x the object count they can do.

 

[zeerround]

I think people should be aware of the current state of Cavern's limitations for decoding atmos:

1) It can decode lossy atmos, typically used for streaming, but not lossless Atmos with TrueHD, as on BlueRay, etc.​

​

2) Tests on music decode, to date, have produced distorted output. Bug filed and in open state.​

​

3) Live decode is limited to 8 channels (Think 5.1.2) or the above mentioned headphone virtualization targeting the author's ears.​

​

​

 

[himey]

I think people should be aware of the current state of Cavern's limitations for decoding atmos:

1) It can decode lossy atmos, typically used for streaming, but not lossless Atmos with TrueHD, as on BlueRay, etc.​

​

2) Tests on music decode, to date, have produced distorted output. Bug filed and in open state.​

​

3) Live decode is limited to 8 channels (Think 5.1.2) or the above mentioned headphone virtualization targeting the author's ears.​

​

​

Do you have an opinion on Cavern?

 

[VoidX]

I think people should be aware of the current state of Cavern's limitations for decoding atmos:

1) It can decode lossy atmos, typically used for streaming, but not lossless Atmos with TrueHD, as on BlueRay, etc.​

​

2) Tests on music decode, to date, have produced distorted output. Bug filed and in open state.​

​

3) Live decode is limited to 8 channels (Think 5.1.2) or the above mentioned headphone virtualization targeting the author's ears.​

​

​

Yes, these are true. A bit more info on these: TrueHD's documentation is kept as a secret by Dolby, and DD+ has a sparse and coarse mode, one of which is documented in a wrong way. This made every third party to only support one of the DD+ modes, including the streaming encoders and Cavern. Only official Dolby encoders know both modes, and content created with them cannot be decoded with any third party software - yet. However, these are super rare, commercial movies work without issues.

 

[himey]

Yes, these are true. A bit more info on these: TrueHD's documentation is kept as a secret by Dolby, and DD+ has a sparse and coarse mode, one of which is documented in a wrong way. This made every third party to only support one of the DD+ modes, including the streaming encoders and Cavern. Only official Dolby encoders know both modes, and content created with them cannot be decoded with any third party software - yet. However, these are super rare, commercial movies work without issues.

Can you comment on the distorted output? I think I know about what zeerround mentioned, but I would like to hear your thoughts first.

 

[VoidX]

Can you comment on the distorted output? I think I know about what zeerround mentioned, but I would like to hear your thoughts first.

Yes, the root cause is only one of the DD+ methods are documented correctly, and content that use both can't be played. This includes content exported with Dolby tools, but nothing else, as other encoders are made with the same wrong documentation (and skip this method completely). There were issues with the doc I could just try again with different parameters, but this one is so complex that it's not possible. Dolby's help is needed, they said they'll update the doc, but I have no idea when it will happen.

 

[zeerround]

What I have seen is music from Tidal has the issue, and music that I have encoded with Dolby tools has the issue.

I have not tested beyond that to date.

I guess another case that I could test would be music that I have encoded on AWS Elemental. I will try that, and report back.

 

[himey]

What I have seen is music from Tidal has the issue, and music that I have encoded with Dolby tools has the issue.

I have not tested beyond that to date.

I guess another case that I could test would be music that I have encoded on AWS Elemental. I will try that, and report back.

Do they show up as spikes in Audition? That is what I noticed when I look at the four top channels in the waveform.

 

[zeerround]

It's been a while, but when I quickly looked in audition, I checked for clipping, with the default settings, and it didn't find any, yet the distortion was clearly audible. I didn't actually look at the waveforms.

Re: testing decodes of atmos music encoded with AWS elemental, I am unable to get (today's build of, or the released version of) cavernize, to render them at all.

AWS gives, mp4, which I tried both just renaming the extension to mkv, and mka, as well as converting to mkv, mka, and ac3 with ffmpeg, e.g.:

>ffmpeg -i "E:\Google Drive\7.1.4 Atmos up-remix samples\Paula Cole - The Hard Way 7.1.4 up-remix 768 Music Light Dialog Disabled.mp4" -map 0:a -c:a copy "E:\Google Drive\7.1.4 Atmos up-remix samples\Paula Cole - The Hard Way 7.1.4 up-remix 768 Music Light Dialog Disabled.ac3"

FYI the ac3 versions played correctly in Dolby Refernce player, rendering to 9.1.6.

Cavernize gave:

Here is how that ac3 file looks in media info:

The original mp4:

 

[himey]

It's been a while, but when I quickly looked in audition, I checked for clipping, with the default settings, and it didn't find any, yet the distortion was clearly audible. I didn't actually look at the waveforms.

The spikes aren't causing clipping, but they look out of place in the waveform. Maybe we are not talking about the same thing. I only tested one album to any extent, the new Bjork, which has so much unique added distortion, that I didn't notice any "clearly audible" distortion.
Does it show up in every channel after decoding, or just the tops?

 

[zeerround]

Clipping (flat tops) is just one kind of distortion. Spikes / Transients being another.

Like I said, I didn't spend a lot of time analyzing what I was hearing, I just reported it and provided samples, in my bug report. It was accepted as a bug, so no need for me to look deeper.

 

[himey]

Clipping (flat tops) is just one kind of distortion. Spikes / Transients being another.

Like I said, I didn't spend a lot of time analyzing what I was hearing, I just reported it and provided samples, in my bug report. It was accepted as a bug, so no need for me to look deeper.

Haha. Sorry for asking.

 

[zeerround]

The latest cavernize download can render my examples from AWS elemental (thanks for the update!) however I do hear distortion in my examples (compared to the dolby reference player). I guess we don't know if AWS is using Dolby encoders or their own code?

FYI .riff output has a bit depth of 16. Is that what was intended?

 

[VoidX]

I guess we don't know if AWS is using Dolby encoders or their own code?

If there's distortion, then it's the Dolby encoder, which can only be played back with Dolby-supported stuff.

FYI .riff output has a bit depth of 16. Is that what was intended?

Yes, there's 16 (int) and 32 (float).

 

[himey]

Do any of the pro, usb DAC's, designed for mixing, which are 10+ channels, work with standard media players like Kodi, Foobar ect., or do they only playback in DAW's?