Article: How To Decode and Play Dolby TrueHD Atmos on Windows and macOS

[El Guapo]

1 hour ago, retro said:

I assume the original ripped mkv is enough for Dolby Reference Player to play, or am I wrong?

Hi retro,

For movies with ATMOS audio, only three containers can be accepted by DRP which are .m2ts, .mp4 and .mov. The supported video codec are H.264, H.265 and AppleProRes. The ATMOS audio are .mlp and .eac3 mainly.

If you already ripped your BD in .mkv, you have to change the container and video codec to the one of the accepted format (assumed the audio is original format when ripped).

 

1 hour ago, retro said:

what about the video..??

DRP doesn’t support full screen video playback and also doesn’t support audio delay adjustment. What I did was I use Audacity to record the DRP output audio then use video editing software to build an audio delayed (audio before video about 1.5s on my system) movie.
I know it’s painful but that’s the only workaround for movies so far.

Guapo

[Jivatma]

thecomputeratmosphile.com ? 🤣

[jrobbins50]

@El Guapo, I actually don’t know how to play the output of the Lynx through my system, even for vinyl, other than through HQP.  The Lynx ASIO shows up as a Roon endpoint, but there’s no track in Roon to play, so obviously that doesn’t work. The vinyl setting in HQP is the same as the preferred method for DRP, with ASIO Lynx as the input backend and then starting a stream. I can’t direct switch from the vinyl stream to DRP, in 2-channel, because the vinyl input comes through the XLR inputs for channels 1&2 and they override the internal digital input from the DRP. I have to physically detach the XLR vinyl input for Lynx to see the DRP. 
 

What does work is a direct USB connection from the DRP server to the Okto DAC, bypassing both Lynx and HQP.  Both the ASIO and the WASAPI DIYNHK driver notations work as DRP outputs in this fashion, for eight channels. See screenshot.  I have to use the physical volume control on the Okto since there is no control software other than the DRP itself. I can hear the 5.1 output in its nicely lossless format — at the moment, my height speakers are directly attached to my Marantz prepro, where I’ve been listening to Apple spacial audio through an Apple TV 4K. I’m certain that the height speakers would work as well with the direct USB connection. I have to combine the two subs through a Y connection for 5.1.2, but that is trivial if everything else runs correctly. 
 

But I need volume control and convolution, which HQP provides. As soon as I plug the Okto back into the microRendu, which, going backwards, runs through an EtherRegen, then to an OpticalModule and then Ethernet back to the server, the DIYNHK outputs both disappear from DRP. 
 

So what gives here?  JCR 

[El Guapo]

1 hour ago, jrobbins50 said:

I have to physically detach the XLR vinyl input for Lynx to see the DRP. 

After quick look at Lynx’s user’s manual I think the meters indicates “digital output” of Lynx’s AES/EBU, not a route to HQP.

 

So you still need another virtual interface to route DRP output to HQP.

 

After installed VAC did you see the “Virtual Audio Cable” options?

Please try to set DRP output to one of the virtual cable, then set such output as HQP’s input. (I saw your previous post you selected Windows’ internal audio port which was incorrect source for this case).

 

[jrobbins50]

@El Guapo, when I play vinyl through the Lynx ASIO, the meter function is solely at the top in channels 1&2 of the record section. I’ve been trying to figure out how to link the digital out at the bottom to the record at the top, figuring then that HQP and Lynx would work together just like they do for vinyl.  I’ve not yet heard back from Lynx technical support on this one. Maybe you have an idea?

 

Meanwhile, what I have is VB-Cable, which may be the same thing you show. I can select it in DRP as an 8 channel virtual input to the DRP’s output, and I can select the Cable Output in HQP.  The problem is that in Windows sound properties, it limits to 2-channel. Even so, this otherwise logical arrangement produces no input for the Okto, not even 2-channel. I tried both with and without Exclusive Mode enabled, this virtual cable showing up only as WASAPI in HQP.  JCR 

[El Guapo]

35 minutes ago, jrobbins50 said:

how to link the digital out at the bottom to the record at the top

Worth a try using physical cable connection (loopback the digital AES/EBU output to the input for vinyl.)

 

41 minutes ago, jrobbins50 said:

Even so, this otherwise logical arrangement produces no input for the Okto, not even 2-channel.

VB-cable has a small control panel, how does the status look like?

 

[jrobbins50]

I must locate appropriate digital XLR patch cable to try the physical connecting of the Lynx outputs back into the inputs. 
 

I cannot find that VB Cable control panel in my installation. But given that it shows only 2 channel anyway in Windows sound properties, I have installed another virtual cable, Virtual Audio Cable https://vac.muzychenko.net/en/. As you can see in the screenshots, it supports 8 channels for playback and recording in Windows sound, appears in DRP as Line 1 (8 channels) for output, is found by HQP as a WASAPI input backend and, when playing a track on the DRP, shows that audio is passing through in this product’s control panel. Yet still, nothing is getting to the DAC. JCR 

[The Computer Audiophile]

Just received this Gordon Goodwin's Big Phat Band, The Gordian knot. Mastered for Dolby Atmos, and I'm ripping it now. 

 

[El Guapo]

1 hour ago, jrobbins50 said:

shows that audio is passing through in this product’s control panel. Yet still, nothing is getting to the DAC

Hmmm… I could use VAC to route audio into HQPlayer Desktop.

Did you try PCM?

I suggest you set the bit range fix at 32bit.

[bbosler]

Quote

Why would someone want to decode and play Dolby TrueHD Atmos content on a computer rather than an AVR or processor? Mainly it's about cost, flexibility, and performance. The cost of a high end processor can set you back $10,000 or much more. The flexibility of a computer based system is endless (for better or worse). With respect to performance, a computer based system can use several types of room correction, 65,000+ taps, upsample to high rate PCM or DSD, and output to high end DACs with interfaces much better than HDMI.

 

I  suspect I may be coming off as a bit of a Trinnov fan boy, but I would like to speak to the final point above regarding performance. I get the point about being able to apply room correction through HQP  What I don't know is how much better if any that correction will be versus Trinnov's built in correction using their microphone and software? Using HQP you can certainly correct for many things, but it will not do what Trinnov does as far as precisely mapping the location of each speaker in the room (azimuth, distance, and height) and mapping objects to them. This allows you to place speakers in positions that deviate from the ideal locations recommended by each CODEC and compensate for these deviations. It also allows you to use the same speaker locations for Auro and Dolby by mapping sounds to multiple speakers. Say due to room constraints  your L-R speakers are further apart than ideal. it will mix a bit of each into the center to move the image toward the middle. It also allows you to take source files and up mix them to Dolby surround, DTS, or Auro 3D. 

 

And unless you are willing to take the deep dive and learn how to use DSP software like Audiolense to generate the required filter files,  you are dependent on services like Mitch to generate the files. If you move or add speakers,  you have  to take measurements, send them Mitch, and wait for the files to come back. With Trinnov, move a speaker, run calibration, save it to a preset. And I will stress again, Trinnov not only knows the frequency and phase response along with how far away the speaker it is, it also knows the azimuth and how high it is. I don't think anybody else can do this

 

Another factor is support. Last night a Trinnov tech from Paris logged onto my AL16 and went through and set up 3 different presets based on my calibration that he likes that set target curves for frequency response, phase, amount of boost and cut, and a bunch of other stuff I'm not sure about. I can select a different preset with a click and try them until I get one that I prefer. 

 

So even if you can come up with software to do what Trinnov does (doubtful) you are faced with the complexity of implementing it. As you say, the flexibility is for better and worse. And yes, I realize that the challenge is a great deal of the fun, so have fun. I am.

 

Here's how it mapped out my 9.3.6 speakers. Green dots are ideal and blue are where mine are. FYI just like us, it can't locate and map sub position, only response

 

 

 

 

 

[The Computer Audiophile]

5 minutes ago, bbosler said:

 

I  suspect I may be coming off as a bit of a Trinnov fan boy, but I would like to speak to the final point above regarding performance. I get the point about being able to apply room correction through HQP  What I don't know is how much better if any that correction will be versus Trinnov's built in correction using their microphone and software? Using HQP you can certainly correct for many things, but it will not do what Trinnov does as far as precisely mapping the location of each speaker in the room (azimuth, distance, and height) and mapping objects to them. This allows you to place speakers in positions that deviate from the ideal locations recommended by each CODEC and compensate for these deviations. It also allows you to use the same speaker locations for Auro and Dolby by mapping sounds to multiple speakers. Say due to room constraints  your L-R speakers are further apart than ideal. it will mix a bit of each into the center to move the image toward the middle. It also allows you to take source files and up mix them to Dolby surround, DTS, or Auro 3D. 

 

And unless you are willing to take the deep dive and learn how to use DSP software like Audiolense to generate the required filter files,  you are dependent on services like Mitch to generate the files. If you move or add speakers,  you have  to take measurements, send them Mitch, and wait for the files to come back. With Trinnov, move a speaker, run calibration, save it to a preset. And I will stress again, Trinnov not only knows the frequency and phase response along with how far away the speaker it is, it also knows the azimuth and how high it is. I don't think anybody else can do this

 

Another factor is support. Last night a Trinnov tech from Paris logged onto my AL16 and went through and set up 3 different presets based on my calibration that he likes that set target curves for frequency response, phase, amount of boost and cut, and a bunch of other stuff I'm not sure about. I can select a different preset with a click and try them until I get one that I prefer. 

 

So even if you can come up with software to do what Trinnov does (doubtful) you are faced with the complexity of implementing it. As you say, the flexibility is for better and worse. And yes, I realize that the challenge is a great deal of the fun, so have fun. I am.

 

Here's how it mapped out my 9.3.6 speakers. Green dots are ideal and blue are where mine are. FYI just like us, it can't locate and map sub position, only response

 

 

 

 

 

 

 

Great post with many excellent points. We have amazing options now days. It's truly a great time to be into this stuff. 

 

I'm a huge fan of Trinnov and think your opinion of the product / service is spot on. I'll call myself a Trinnov fanboy as well. I don't own one, but have used them and have talked to the company a couple times. Everything about Trinnov is first class. I plan to get an Altitude16 in for review. I really look forward to hearing what it can do in my room and to write about it all. Greta stuff.

 

Immersive sound from a computer, using Audiolense based convolution filters is new territory for me. I'm excited to see what magic Mitch can do in this area. 

 

All fun stuff.

[jrobbins50]

@bbosler, I think that to do it in Trinnov is great.  But, I’m really constrained to 8 channels and my Okto dac can handle that (if I can get it all to work), without having to invest, at this time, in the Trinnov. So, it is a cost factor, at some juncture, for everyone as to how to proceed. 
 

@El Guapo, the physical loopback does work. While I don’t have my height channels connected yet this way, I’m getting 5.2 through HQP.  
 

Why you can get audio out using the VAC — which I far prefer — and I can’t, is a mystery to me. The physical arrangement is annoying because if I want to switch from DRP to vinyl, I have to go back into my storage area where the server is located and physically change the XLR plug-ins. I suppose there must be product that could automate switching XLR inputs, if I want to track that down. 
 

Could WS2019 be the culprit in this all?  I wonder if I should just image the server for backup and start over fresh with a W10 install. 

 

Meanwhile. I think the channels out of the DRP are routed incorrectly — I think the height info is coming out of the surrounds, for example — as are volume levels among channels.  I bet that can be reassigned in the Lynx mixer or by physically rearranging the loopback.  I wish I could play my 8 channel test track through the DRP to confirm L R C LS RS LH RH routing. Do you think a FLAC file could be converted to play out through the DRP?

 

Progress, anyway.  But,@The Computer Audiophile, I’m getting tired of Come Together as the only playable track from Abbey Road. Any further word on getting the right metadata ripped into .mlp files? JCR 

[jrobbins50]

@El Guapo, 32 bit fixed did not awaken the DAC. JCR 

[El Guapo]

6 minutes ago, jrobbins50 said:

I suppose there must be product that could automate switching XLR inputs,

Worth a try:

Input ch 1 and 2 for vinyl

Input ch 3~10 for 5.1.2

In HQPlayer when you play ATMOS you can set input “channel offset” to 2 and HQPlayer should see ch3 as the 5.1.2’s ch1.

DRP will put the 5.1.2’s 0.0.2 to 7th and 8th output channel. You have height channel for sure. Guapo

[mitchco]

Great topic Chris! Folks should know there are Atmos specifications for both the production (i.e. creation) and reproduction of Atmos material. Includes speaker arrangements, distances, angles, azimuth, etc., including tolerances.

Here is one such specification: Multichannel sound technology in home and broadcasting applications.

As long as one is within the tolerances specified, pretty much guaranteed excellent spatial sound.

 

I have been working with a few recording studios on the Atmos production side providing DSP/DRC services for their 7.1.4 setups. These folks are using Dolby Atmos Production Suite on the Mac and DAW's. It has been an interesting experience and I will see if I can get permission to show the room correction results of a 7.1.4 system, it is pretty amazing.

While not strictly Atmos speaking, I have a few clients that have compared (extensively) digital room corrections from software based DSP/DRC to Trinnov Altitude. The main differences are that the Trinnov, like most h/w based DRC solutions, are limited by the DSP chipsets used and how many FIR filter taps can be processed. The Trinnov uses 2048 FIR filter taps per channel but only down to 150 Hz and then switches to IIR filters, which means no excess phase correction at low frequencies (think back wall bounce of low frequencies not taken care of). Conversely with software based DSP solutions, 65,536 or 131,072 tap filters are possible for the ultimate low frequency response control

 

Here is a comparison of a system using Trinnov and Audiolense for room correction on the same system. Trinnov on top and Audiolense on bottom:

 

 

There still is a performance gap between the best h/w versus s/w based DSP/DRC systems. Don't get me wrong, pretty much any DRC system is better than nothing, but not all DSP/DRC systems have the same performance.

 

On a side note, I am working on the multi-channel version of Hang Loose Convolver, so hope have something for you to test when your system arrives Chris.

 

[jrobbins50]

@El Guapo— A good experiment, thank you. Problem I think is that since I am using physical loopback, and the ins and outs of the AES16e are in pairs, I am using all physical inputs for the DRP, leaving none for the vinyl. 
 

This file https://www.youtube.com/redirect?event=video_description&redir_token=QUFFLUhqbjFtVjh4VldJVEliYVlLUzk3Q2dNel9hRzdLZ3xBQ3Jtc0trN0t2TkZybGpqeWhtR2JFcHY1dnFoLVRBTkZodXMyQUk1aFNrZWVZY3Exc1JqZzZyQTFGQTBPZ25MeWZ4SEQ2ckZDaUVkOHNUd1liRHo1R2ZNbWh5S0QyQkYyQndHamx0VWtoQVpJMVdRRGdTZjZuUQ&q=https%3A%2F%2Fbit.ly%2FDownload-5-1-2-MP4&v=3MC-Y0vvwx4 is recognized by the DRP as a Dolby Atmos .mp4 for 5.1.2 test tones with a little video, and audio, to show speaker setup. My left surround is coming out of the center and left main, as is the left height, and the opposite for the right side. So, it’s very screwed up and I’m not sure how to map this to work right. Further thoughts?

 

@mitchco, I see you have now let the cat out of the bag on your HLC MCH. Great!  Maybe you have some thoughts on how to fix the channel mismatching now that I actually have DRP producing audio. JCR 

[The Computer Audiophile]

10 minutes ago, mitchco said:

On a side note, I am working on the multi-channel version of Hang Loose Convolver, so hope have something for you to test when your system arrives Chris.

@mitchco don’t bury the lede!

 

This will be so great!

[The Computer Audiophile]

11 minutes ago, jrobbins50 said:

@El Guapo— A good experiment, thank you. Problem I think is that since I am using physical loopback, and the ins and outs of the AES16e are in pairs, I am using all physical inputs for the DRP, leaving none for the vinyl. 
 

This file https://www.youtube.com/redirect?event=video_description&redir_token=QUFFLUhqbjFtVjh4VldJVEliYVlLUzk3Q2dNel9hRzdLZ3xBQ3Jtc0trN0t2TkZybGpqeWhtR2JFcHY1dnFoLVRBTkZodXMyQUk1aFNrZWVZY3Exc1JqZzZyQTFGQTBPZ25MeWZ4SEQ2ckZDaUVkOHNUd1liRHo1R2ZNbWh5S0QyQkYyQndHamx0VWtoQVpJMVdRRGdTZjZuUQ&q=https%3A%2F%2Fbit.ly%2FDownload-5-1-2-MP4&v=3MC-Y0vvwx4 is recognized by the DRP as a Dolby Atmos .mp4 for 5.1.2 test tones with a little video, and audio, to show speaker setup. My left surround is coming out of the center and left main, as is the left height, and the opposite for the right side. So, it’s very screwed up and I’m not sure how to map this to work right. Further thoughts?

 

@mitchco, I see you have now let the cat out of the bag on your HLC MCH. Great!  Maybe you have some thoughts on how to fix the channel mismatching now that I actually have DRP producing audio. JCR 

DRP has mapping. 

[jrobbins50]

Ok, Chris, and for the uninitiated, where and how does that work within the DRP?  JCR 

[The Computer Audiophile]

1 hour ago, jrobbins50 said:

Ok, Chris, and for the uninitiated, where and how does that work within the DRP?  JCR 

[retro]

23 hours ago, The Computer Audiophile said:

 

Hi Retro, I'm glad you're as excited about this as we are. This is really something, and it's only going to get better. I was playing around with the command line interface of the Dolby Reference Player, and think ease of use can be taken to another level. But anyway, I'll leave it to @El Guapo to talk about video. He has video working in his system. You need to convert to M2TS files because the Dolby Reference Player only accepts a few file formats.

 

 

Hello and thanks for the reply,

 

Yes, for sure it's early days. 

One worry I have is that this is relying on Dolby's software, and since it's used in a way they obviously not intended, they could stop selling it to us, or change the way it works. It's possible, don't you think..?

Yes, I worry about everything. It's a problem..😉

[retro]

21 hours ago, El Guapo said:

Hi retro,

For movies with ATMOS audio, only three containers can be accepted by DRP which are .m2ts, .mp4 and .mov. The supported video codec are H.264, H.265 and AppleProRes. The ATMOS audio are .mlp and .eac3 mainly.

If you already ripped your BD in .mkv, you have to change the container and video codec to the one of the accepted format (assumed the audio is original format when ripped).

 

DRP doesn’t support full screen video playback and also doesn’t support audio delay adjustment. What I did was I use Audacity to record the DRP output audio then use video editing software to build an audio delayed (audio before video about 1.5s on my system) movie.
I know it’s painful but that’s the only workaround for movies so far.

Guapo

Hello Guapo and thanks for taking the time.

 

Another container is of course no problem. But you still would need to have them split then, right? Like m2ts for video and mlp for audio?

 

The video playback is a "thing" though..🙂 I play all my audio and video with JRiver, guess at this point it's not possible then..?!?

But hey..this is early days..I'm sure solutions will come.

 

 

[The Computer Audiophile]

4 minutes ago, retro said:

Hello and thanks for the reply,

 

Yes, for sure it's early days. 

One worry I have is that this is relying on Dolby's software, and since it's used in a way they obviously not intended, they could stop selling it to us, or change the way it works. It's possible, don't you think..?

Yes, I worry about everything. It's a problem..😉

That’s the sole reason why I hesitated to release this article. 

[retro]

1 hour ago, The Computer Audiophile said:

That’s the sole reason why I hesitated to release this article. 

Understood.

And for the same reason, I have not linked this on the JRiver forum. But there are some excellent software guys there that I'm sure could help to get this going further. 

[El Guapo]

2 hours ago, retro said:

still would need to have them split then, right? Like m2ts for video and mlp for audio?

No need to split. For example you can put H.264 video and .mlp audio into .mp4 container. 

 

2 hours ago, retro said:

guess at this point it's not possible then..?!?

Full screen playback for movies it’s not possible ATM.🥲