DIY WiSA Home Theater system

[somebodyelse]

Cavern Is Windows only, it won’t even run on a pi.

It might, via hangover (uses FEX and Wine to run Windows/intel apps on linux/arm64.) I doubt it's been tried that way though.

 

[voodooless]

It might, via hangover (uses FEX and Wine to run Windows/intel apps on linux/arm64.) I doubt it's been tried that way though.

That might be worth a try...

Though I found this random script:

https://www.avsforum.com/threads/decoding-truehd-atmos-to-16-channel-wav.3265364/

No height channels, though?

@VoidX seems to know a bit about this, can probably tell us more?

 

[somebodyelse]

That's a curious mix of ffmpeg and gstreamer bits - with the gstreamer part using ffmpeg as the decoder for ac3 or truehd. I'm not sure if there's part of it that ffmpeg can't do directly, or people are just doing the bit they need to do with the tool they're more familiar with. Both ffmpeg and gstreamer can use streams instead of files for input and output though, so it should be viable to rework for the necessary inputs and outputs.
https://gstreamer.freedesktop.org/documentation/libav/avdec_truehd.html?gi-language=c
https://gstreamer.freedesktop.org/documentation/libav/avdec_ac3.html?gi-language=c
I wonder which capabilities in cavern aren't implemented in ffmpeg and/or gstreamer yet. For me the interesting bit would be to render the atmos stream for headphones via HRTF as I don't have multichannel speakers.

 

[somebodyelse]

Another thing about ffmpeg, gstreamer and similar - some distros (primarily redhat/fedora/centos) don't include all the plugins in the packages because of patents. I haven't checked to see whether that would apply here.

 

[voodooless]

Another thing about ffmpeg, gstreamer and similar - some distros (primarily redhat/fedora/centos) don't include all the plugins in the packages because of patents. I haven't checked to see whether that would apply here.

Oh, I think it might. I think some of the Dolby patents are about to, or have expired a short while ago. Not the Atmos one, I think.

The Seemingly Incompatible Becomes Reality: Dolby Digital Audio in FFmpeg

Experience seamless Dolby Digital audio integration in FFmpeg with MainConcept's certified AC-3 and E-AC-3 plugins

blog.mainconcept.com

Seems like E-AC-3 (DD+) is still under patent until 2026

 

[somebodyelse]

I hadn't realised there was a paid, Dolby-sanctioned ffmpeg plugin available. Seems to be x86 only though.

 

[Glider95]

I just saw @VoidX open issue where it seems he has succeeded to real-time decode Dolby Atmos: https://github.com/VoidXH/Cavern/issues/201
From what I understand from the thread, he is just missing a player where he could implement CaverPipeline as a codec or similar...

 

[VoidX]

I just saw @VoidX open issue where it seems he has succeeded to real-time decode Dolby Atmos: https://github.com/VoidXH/Cavern/issues/201
From what I understand from the thread, he is just missing a player where he could implement CaverPipeline as a codec or similar...

Yes, on Windows. If you use Cavern embedded, fetching it the bitstream in real time, it can decode Atmos with a 64 sample delay (which is the same in AVRs).

 

[Pete Lee]

@Glider95 I am utterly delighted by this post! Reason? I work for WiSA Technologies (subsidiary of DataVault AI). I literally have a stack of dev boards on my desk that's our next generation technology that will replace the WiSA HT hardware you have (link to WiSA E Product Brief).

I have several decades of experience in tech and support our fantastic hardware engineers*. I don't know how much I can help, but please tag me on your posts and I can ask around and see what I can do!

*I am definitely not an engineer--but recently purchased a Modulus-86 10th Anniversary kit that I'm building (very slowly). For fun, I'll be connecting a pair of Elac UniFi UB5 speakers to some WiSA E dev kits that will be powered by the Modulus amps.

HUGE thanks to @tomchr at Neurochrome for the impressive amounts of documentation and encouraging YouTube videos: only took me an hour to get R1 soldered on, LOOOOL (problem with soldering iron tip, I've already got what I need to fix that). My hardware engineering friends have helped me with soldering in the past, but since the Modulus boards are all thru-hole, they're making me actually do it. SO MEAN!

 

[MCH]

@Glider95 I am utterly delighted by this post! Reason? I work for WiSA Technologies (subsidiary of DataVault AI). I literally have a stack of dev boards on my desk that's our next generation technology that will replace the WiSA HT hardware you have (link to WiSA E Product Brief).

I have several decades of experience in tech and support our fantastic hardware engineers*. I don't know how much I can help, but please tag me on your posts and I can ask around and see what I can do!

*I am definitely not an engineer--but recently purchased a Modulus-86 10th Anniversary kit that I'm building (very slowly). For fun, I'll be connecting a pair of Elac UniFi UB5 speakers to some WiSA E dev kits that will be powered by the Modulus amps.

HUGE thanks to @tomchr at Neurochrome for the impressive amounts of documentation and encouraging YouTube videos: only took me an hour to get R1 soldered on, LOOOOL (problem with soldering iron tip, I've already got what I need to fix that). My hardware engineering friends have helped me with soldering in the past, but since the Modulus boards are all thru-hole, they're making me actually do it. SO MEAN!

Hi Pete, are those modules available for the general public? If yes,where can one place an order? Thanks.

 

[Glider95]

@Glider95 I am utterly delighted by this post! Reason? I work for WiSA Technologies (subsidiary of DataVault AI). I literally have a stack of dev boards on my desk that's our next generation technology that will replace the WiSA HT hardware you have (link to WiSA E Product Brief).

I have several decades of experience in tech and support our fantastic hardware engineers*. I don't know how much I can help, but please tag me on your posts and I can ask around and see what I can do!

*I am definitely not an engineer--but recently purchased a Modulus-86 10th Anniversary kit that I'm building (very slowly). For fun, I'll be connecting a pair of Elac UniFi UB5 speakers to some WiSA E dev kits that will be powered by the Modulus amps.

HUGE thanks to @tomchr at Neurochrome for the impressive amounts of documentation and encouraging YouTube videos: only took me an hour to get R1 soldered on, LOOOOL (problem with soldering iron tip, I've already got what I need to fix that). My hardware engineering friends have helped me with soldering in the past, but since the Modulus boards are all thru-hole, they're making me actually do it. SO MEAN!

Hello @Pete Lee,

Happy to see that this thread has came to your knowledge. I did try to contact your support for some questions when I started but without success aha.

As you can see, I have not progressed much since a few months. However, as requested by @MCH, would you be able to communicate a bit more on WiSA E product ? (Availability, how the final product plan to be sold, etc...)
From my point of view, WiSA HT or particularly the Soundsend device is still of interest because it is Dolby ATMOS certified but as discussed recently in this thread, ATMOS decoding seems to be almost available to all thanks to @VoidX !

 

[Pete Lee]

Hi Pete, are those modules available for the general public? If yes,where can one place an order? Thanks.

Unfortunately, we're not set up to sell to the general public (we're a fairly small company): we work directly with speaker manufacturers and such.

 

[Pete Lee]

Hello @Pete Lee,

Happy to see that this thread has came to your knowledge. I did try to contact your support for some questions when I started but without success aha.

As you can see, I have not progressed much since a few months. However, as requested by @MCH, would you be able to communicate a bit more on WiSA E product ? (Availability, how the final product plan to be sold, etc...)
From my point of view, WiSA HT or particularly the Soundsend device is still of interest because it is Dolby ATMOS certified but as discussed recently in this thread, ATMOS decoding seems to be almost available to all thanks to @VoidX !

We are actively working to get WiSA E integrated with speakers right now, though I don't know the exact timeline.

WiSA | Technologies

WiSA technologies ensure the highest quality audio is transmitted and received—with stunning quality, and perfect synchronization.

www.wisatechnologies.com

 

[Glider95]

Thank you @Pete Lee.
Pitty to see that the technology will be open only to manufacturers :/

With the increasing DIY Speaker projects and the new design appearing every weeks (partially thanks to 3D Printing ! ), an open to public wireless high quality audio transmission would have been great ;D.

On my side, I will try to continue to explore the use of the older WiSA HT (but same issue in term of board procurement, only available from scrapped products) or look into the new ESP32-C5 which onboards 5GhZ WiFI to create an equivalent of what WiSA E is achieving... (but that's huge amount of work).
A combination of these technologies + Cavern + Class D amps would be amazing !

 

[nullpointerninja]

Hi @Pete Lee, it’s nice to have you here!
I’ve been looking all over the internet for more information about a DIY WiSA system but I think @Glider95 is the only one who’s managed to find and document how all this works.

I have a question for you Pete, about the WiSA ready TVs and game systems. The Axiim Link, which is a USB device, presents itself as a generic multichannel DAC when used with a Mac or PC. Do you know what makes a WiSA ready TV to identify it as a WiSA compatible dongle instead of a generic DAC? Is it the vendor/product ID or some identifier, or something completely different?

I wonder if, for example, we could use a Raspberry Pi in gadget mode to trick a TV thinking a WiSA enabled transmitter was connected and then apply DSP correction with CamillaDSP.

Thanks!

 

[Pete Lee]

@nullpointerninja I don't have a ton of knowledge about the Axiim Link, unfortunately because that was way before my time at WISA. I am fairly certain that is based on the WISA HT platform. I don't have a ton of details on how that works - my apologies!

My focus is on the WISA E platform: if I am able to, I'm going to see if I can get approval to send a transmitter and some wireless speakers to Amir to see what he thinks. Fingers crossed!

 

[Glider95]

Curi

@nullpointerninja I don't have a ton of knowledge about the Axiim Link, unfortunately because that was way before my time at WISA. I am fairly certain that is based on the WISA HT platform. I don't have a ton of details on how that works - my apologies!

My focus is on the WISA E platform: if I am able to, I'm going to see if I can get approval to send a transmitter and some wireless speakers to Amir to see what he thinks. Fingers crossed!

Curiosity question @Pete Lee : Does WiSA E will be retro compatible with WiSA HT / WiSA SoundSend ?

 

[Glider95]

I had some time this week-end and I came across really interesting projects/hardware in term of wireless + amplification:

Hardware:
A ready to use ESP32+TAS5805M for only 24 USD: https://sonocotta.com/louder-esp32/. It is really similar to the TAS5828M I tried to sketch around few month ago, but this one does not rely on Texas Instrument proprietary software and everything can be set via I2C from what I understand however it is slightly less powerful in term of outputs.
I will discuss with the maintainer if an update to support ESP32-C5 (5Ghz) could be interesting.

Software:
TAS5805M Driver for Raspberry Pi / Raspbian: https://github.com/sonocotta/tas5805m-driver-for-raspbian
Snapcast - A multiroom streaming player: https://github.com/badaix/snapcast - I think there should be an existing Kodi server addon.
ESP32 Snapcast client: https://github.com/sonocotta/esparagus-snapclient.

I regret not doing more thorough research before starting the WiSA project, I believe almost similar flow could be achieved with a combination of all these projects. Only one points will never beat WiSA is the latency: a combination of the all above will never reach <50ms (more like 200ms) but is it actually important today where streaming services or Media center always include a video delay feature ?

One missing feature is the Dolby ATMOS part but I will get in touch with @VoidX to see how I could support him in either:
- Push Kodi maintainer for integration of his real time ATMOS decoding: https://github.com/VoidXH/Cavern/issues/201. It seems Kodi's Team did not really understood the added value of Cavern, see thread.
- A porting of his work to Linux based system which are on pause due to time constraints I think: Issue 231 & Issue 123. I tried a really dirty porting using Copilot this week-end, but I have no idea of what I am doing so....

Ideas of flows (both include an added video delay idea):
- SmartTV with Kodi + Snapcast server addon + Cavern ATMOS decoding addon ------ WiFi----> ESP32 Snapcast Client + Class D Amp.
- HTPC with Kodi + Cavern ATMOS streaming decoding + Broadcasting (I guess still Snapcast) ------ WiFi----> ESP32 Snapcast Client + Class D Amp.

As well, I managed to get hands on 2 ESP32-C5 Dev boards, should arrive next week !

EDIT:
Snapcast server seems to be only available for Kodi with Libreelec, not on AndroidTV or Kodi Raspbian...

 

[MCH]

I had some time this week-end and I came across really interesting projects/hardware in term of wireless + amplification:

Hardware:
A ready to use ESP32+TAS5805M for only 24 USD: https://sonocotta.com/louder-esp32/. It is really similar to the TAS5828M I tried to sketch around few month ago, but this one does not rely on Texas Instrument proprietary software and everything can be set via I2C from what I understand however it is slightly less powerful in term of outputs.
I will discuss with the maintainer if an update to support ESP32-C5 (5Ghz) could be interesting.

Software:
TAS5805M Driver for Raspberry Pi / Raspbian: https://github.com/sonocotta/tas5805m-driver-for-raspbian
Snapcast - A multiroom streaming player: https://github.com/badaix/snapcast - I think there should be an existing Kodi server addon.
ESP32 Snapcast client: https://github.com/sonocotta/esparagus-snapclient.

I regret not doing more thorough research before starting the WiSA project, I believe almost similar flow could be achieved with a combination of all these projects. Only one points will never beat WiSA is the latency: a combination of the all above will never reach <50ms (more like 200ms) but is it actually important today where streaming services or Media center always include a video delay feature ?

One missing feature is the Dolby ATMOS part but I will get in touch with @VoidX to see how I could support him in either:
- Push Kodi maintainer for integration of his real time ATMOS decoding: https://github.com/VoidXH/Cavern/issues/201. It seems Kodi's Team did not really understood the added value of Cavern, see thread.
- A porting of his work to Linux based system which are on pause due to time constraints I think: Issue 231 & Issue 123. I tried a really dirty porting using Copilot this week-end, but I have no idea of what I am doing so....

Ideas of flows (both include an added video delay idea):
- SmartTV with Kodi + Snapcast server addon + Cavern ATMOS decoding addon ------ WiFi----> ESP32 Snapcast Client + Class D Amp.
- HTPC with Kodi + Cavern ATMOS streaming decoding + Broadcasting (I guess still Snapcast) ------ WiFi----> ESP32 Snapcast Client + Class D Amp.

As well, I managed to get hands on 2 ESP32-C5 Dev boards, should arrive next week !

Do you think you can achieve good synchronization between different channels with the esp32 option? How would that work?

 

[Glider95]

Do you think you can achieve good synchronization between different channels with the esp32 option? How would that work?

It seems Snapcast has a well supported synchronization where client are continuously checking time on server.

Quoting the doc:

Each client does continuous time synchronization with the server, so that the client is always aware of the local server time. Every received chunk is first decoded and added to the client's chunk-buffer. Knowing the server's time, the chunk is played out using a system dependend low level audio API (e.g. ALSA) at the appropriate time. Time deviations are corrected by playing faster/slower, which is done by removing/duplicating single samples (a sample at 48kHz has a duration of ~0.02ms).

Typically the deviation is below 0.2ms.

Do you think it would not be enough ?
Someone also asked recently a similar question for a stereo setup: https://github.com/badaix/snapcast/discussions/1381