[REQUEST] Stereo Pair Support?

[Hedda]

It is not clear to me from reading the Resonate protocol specification if it already supports ”stereo-pairing” or not, so wondering if this Resonate media player client implementation for ESPHome will support setting up a ”stereo-pair” (a.k.a. ”stereo group”) as in pairing two separate clients as dedicated left and right channel speakers to allow a paired set of speakers to be for stereo output together if installed side-by-side?

Use case is example for a proof-of-concept setup would be configuring two Home Assistant Voice Preview Edition satellites as a such stereo-pair and with them paired one of the Home Assistant Voice PE satellite will only output the left channel audio and the other Home Assistant Voice PE satellite only output the right channel audio.

For reference, Squeezekite and a few other Player Providers for Music Assistant do feature "Stereo Pair Support” so wondering if both Resonate Music Provider (server-side) and Resonate Player Provider (client-side) will be supporting a matching function for "Stereo Pair Support” (including UI configuration?)?

• https://www.music-assistant.io/player-support/

Follow-up question is if the player settings for such a "Stereo Pair Support” will allow configurations of stereo pairs of speakers on both the server-side (Music Assistant Server) and the client-side (i.e. client configuration on example ESPHome based device) or both?

These questions is also related to the Resonate Provider PR for Music Assistant and the Resonate synchronized audio component PR for ESPHome:

• Add Resonate Provider music-assistant/server#2348

• WIP: Resonate synchronized audio component esphome/esphome#10183

[maximmaxim345]

Stereo pairs can already be created with the current protocol spec, since:

• Every player has its own audio stream.
• There is no requirement that the streams need to be exactly the same between players. (this is actually one of the main reasons why resonate is not broadcasting audio!)

However, creating a pair cannot be done yet (but will be later), since Music Assistant can only send a single audio stream to the aioresonate library, which then distributes it across all players in the group. This approach is easier for prototyping for now.

An idea I've had from the start is to have a standardized way for the player to tell the server what kind of speaker is connected, for example:

• This is a small speaker with a frequency range of 300 Hz to 20 kHz, like a Voice Preview edition.
• This is a small, full-range speaker with a frequency range of 80 Hz to 20 kHz. Even though it is stereo-capable, one should be able to include it in a stereo pair (since the left and right channels are very close to each other).
• This is a subwoofer with a frequency range of 20 Hz to 200 Hz.

Then, you would be able to group not only a stereo pair but a whole 2.1 setup, with the subwoofer's crossover adjusted to the capabilities of the two main speakers.

Also, Resonate is not limited to just two channels, so technically, surround sound should also be possible. However, this is a very niche use case for music and is very difficult to support for movies for many reasons.

Another question is if the distinction between 2 grouped players and a stereo pair should be done in the application (so Music Assistant), the server implementation (so aioresonate), or somehow be integrated in the protocol (closely related to #5)

[Hedda]

Stereo pairs can already be created with the current protocol spec, since:

• Every player has its own audio stream.
• There is no requirement that the streams need to be exactly the same between players. (this is actually one of the main reasons why resonate is not broadcasting audio!)

However, creating a pair cannot be done yet (but will be later), since Music Assistant can only send a single audio stream to the aioresonate library, which then distributes it across all players in the group. This approach is easier for prototyping for now.

An idea I've had from the start is to have a standardized way for the player to tell the server what kind of speaker is connected, for example:

• This is a small speaker with a frequency range of 300 Hz to 20 kHz, like a Voice Preview edition.
• This is a small, full-range speaker with a frequency range of 80 Hz to 20 kHz. Even though it is stereo-capable, one should be able to include it in a stereo pair (since the left and right channels are very close to each other).
• This is a subwoofer with a frequency range of 20 Hz to 200 Hz.

Then, you would be able to group not only a stereo pair but a whole 2.1 setup, with the subwoofer's crossover adjusted to the capabilities of the two main speakers.

Also, Resonate is not limited to just two channels, so technically, surround sound should also be possible. However, this is a very niche use case for music and is very difficult to support for movies for many reasons.

Another question is if the distinction between 2 grouped players and a stereo pair should be done in the application (so Music Assistant), the server implementation (so aioresonate), or somehow be integrated in the protocol (closely related to #5)

FYI, @sonocotta has posted a crowdfunding page on CrowdSupply for their upcoming ”Esparagus Audio Brick” (a new ESP32 based Hi-Fi D-class amplifier as open-source hardware) also mention future ideas 2.1 stereo setup with subwoofer or bi-amp concept connecting separate devices which sound like awesome concepts if could be achieved with the Resonate protocol:

https://www.crowdsupply.com/sonocotta/esparagus-audio-brick

Sonocotta’s PCBs are design by @anabolyc as open-source hardware so could maybe be a good reference hardware for an ESPHome-based Resonate client to Music Assistent to test different setups with stereo pairs and separate subwoofer(?):

• https://youtu.be/Ft5_anhSEE8

He has by the way has previously also designed similar ESP32-based audio hardware in the ”Esparagus” and ”ESP Louder” series lineups under his Sonocotta brand, including ”ESParagus Media Center” and ”Loud ESP”, as well as some other ESP32-based HiFi DAC development boards:

https://www.crowdsupply.com/sonocotta/esparagus-media-center

• https://github.com/sonocotta/esparagus-media-center

• https://www.crowdsupply.com/sonocotta/loud-esp

  • https://github.com/sonocotta/loud-esp

  • https://github.com/sonocotta/esp32-audio-dock

PS: Another off-topic but cool thing that he and @mrtoy-me got working in ESPHome is an external TAS5805M DAC component for ESPHome that supports advanced device configuration for DSP equalizer control:

• https://www.crowdsupply.com/sonocotta/esparagus-media-center/updates/esphome-driver-changes-to-support-home-assistant-integration-of-louder

  • https://github.com/mrtoy-me/esphome-tas5805m

[balloob]

Let's keep this issue on-topic about stereo pairing.

Maxim is right that the Resonate spec can do stereo pairing today. However, it would then need to be fully configured inside the server. I think that's okay for the start.

In the long run, and especially if more Resonate servers come available, it would feel more natural if Resonate clients can somehow tell a server that they are a stereo or 2.1 set up with other clients, so each server doesn't need to be configured separately. One of the cool things about music is that you can bring it with you, and having to ask your guests to configure the speakers in their app is a barrier to sharing.

[Hedda]

In the long run, and especially if more Resonate servers come available, it would feel more natural if Resonate clients can somehow tell a server that they are a stereo or 2.1 set up with other clients, so each server doesn't need to be configured separately. One of the cool things about music is that you can bring it with you, and having to ask your guests to configure the speakers in their app is a barrier to sharing.

Having support for multiple Resonate servers is on my wishlist for a other reason as well, so I think that is another good argument for moving configuration of stereo-pair to Resonate clients in the long run.

[Hedda]

Does the work-in-progress Sendspin component for ESPHome fully support configurations stereo-pairing of Sendspin clients now?

• [sendspin] (WIP) Add Sendspin synced audio protocol support esphome/esphome#12284

Follow-up question is if configuration of stereo-pairing be added as settings per device in Music Assistant and Home Assistant's UI?

That it, wish for the ability to configure the setup stereo-pair, change/adjust, and break-up/remove a stereo-pair be easy in Music Assistant and Home Assistant (including official companion apps), i.e. it should preferably be as simple (or easier) as it is to configure "Stereo Groups" of Google Nest smart speaker devices and "Speaker Sets" Amazon Echo smart speaker devices in their respecive propriatory Google Home and Amazon Alexa companions apps for mobile platforms. For reference, see:

• https://support.google.com/googlenest/answer/7174267?hl=en

and

• https://www.amazon.se/gp/help/customer/display.html?nodeId=GCJUDLBM6WF79QW2

Btw, noted that Google calls combining speakers a "stereo group", while Amazon refer two combined speakers as a "Speaker Set".

• https://www.youtube.com/watch?v=RXQnJNhh5Ig

As heard there; streaming music is the number-1 use-case for smart speakers, so using two-and-two as a stereo-pairs is a key feature:

Create Speaker Sets with Compatible (Amazon) Echo Devices

For stereo sound and improved music playback in the left and right channels, pair two supported Echo devices:

1. Open the Alexa app.
2. Select Devices.
3. Select plus.
4. Select Combine Speakers.
5. Select Stereo Pair / subwoofer, and then follow the on-screen instructions.

Create and manage speaker groups in the Google Home app

Important prerequisites:

• Speaker group creation is only available on a phone or tablet. This feature is not supported on the Chrome browser. However, you can still cast audio from Chrome.
• Make sure your phone or tablet is connected to the same Wi-Fi or Google Account as your speaker, display, Google streaming devices, or Pixel Tablet.
• You can add or remove speakers from playback or move media from one speaker to another without creating a pre-set speaker group.
• If you make edits to a group that's currently casting media, group playback stops and all audio devices in the group also stop playing. You need to return to the Cast-enabled app or say a voice command to start casting media again.
• If you delete a group that's currently casting media, group playback stops and all audio devices in the group also stop playing.
• Check if your devices are on the same Wi-Fi network.
• Find out which Google Account is linked to your speaker or display.
• Tip: You can only add a Pixel Tablet to a speaker group when it's attached to the Charging Speaker Dock.

Create a speaker group

1. Open the Google Home app Google Home app.
2. At the top, tap Add Add Icon and then Speaker group.
3. Choose the devices that you want to add to the group. A check Check will appear next to each selected device.
4. Tap Next, then enter a name for your group.
5. Tap Save.

Edit a speaker group

1. _ Open the Google Home app Google Home app._
2. _ Tap Home and then All devices and then the group you want to edit._
3. _ Tap Settings Choose devices._
4. _ Choose the devices that you want to add or remove. All devices in the group will have a check Check._
5. _ Tap Save._

Delete a speaker group

1. Open the Google Home app Google Home app.
2. Tap Home and then All devices and then the group you want to delete.
3. Tap Settings .
4. Tap Delete group Remove or Delete.

Add, remove, or move speakers while playing media

1. Open the Google Home app Google Home app.
2. Tap Home and then Favorites .
3. On the mini-player, tap Cast .
4. Tap each device that you want to add or remove. All selected devices will have a check Check.

How to correct group playback delay

1. Cast music to the group that's playing out of sync.
2. Stand between the out of sync speakers (or as close to in between as you can) and adjust their volume levels so you hear them both playing at about the same level.
3. Open the Google Home app Google Home app.
4. Tap Home and then All devices , then touch and hold your device's tile.
5. Tap Settings settings icon and then Audio and then Group delay correction.
6. Move the slider to the left or right until the audio sounds in sync.

Note: For many devices, the inherent delay is extremely short, so group playback will be in sync without any additional adjustments. Don’t adjust the group delay correction unless you notice a consistent, significant delay. The group delay correction setting only affects group playback. This setting will have no effect when casting to a single speaker or TV. A device’s group delay correction setting will be applied to all the groups that the device is a part of. If you connect a Google streaming device to another speaker or TV, the previous group delay correction setting might be incorrect. Make sure to reset it to "0" delay again and check if additional correction is required.

• https://support.google.com/chromecast/answer/6318642

[maximmaxim345]

Does the work-in-progress Sendspin component for ESPHome fully support configurations stereo-pairing of Sendspin clients now?

Technically, stereo groups should already be possible with the current beta.
Its not intuitive/user friendly yet, but you can:

• Change the Output Channel Mode under the the player from Stereo to just Left/Right
• Enable DSP (you don't need to change anything in there, just enable the DSP) (no longer required)
• And group both together (or create a permanent sync group with both left and right speakers)

Haven't tested it myself, but it should work.
Edit: It works!

But there is still a long way in terms of making it simple and easy.
From the protocol itself. Stereo groups are fully transparent to the players.
So adding proper stereo pair/group/set support to Music Assistant and the protocol don't need changes to existing speaker implementations.

[timotoots]

Why stop with stereo. I would be interested in using surround sound part if it's possible to achieve somehow. Sending several different sound sources (channels) to several different clients, but in sync. There are also multi-channel music made and played. Also multi-channel recordings of concerts. If thinking long-term, it makes sense to have support also for this, even if niche. Great project by the way!

[Hedda]

But there is still a long way in terms of making it simple and easy.
From the protocol itself. Stereo groups are fully transparent to the players.
So adding proper stereo pair/group/set support to Music Assistant and the protocol don't need changes to existing speaker implementations.

Indeed, for refence: From a end-user-perspective to explain why configuring stereo-pairs will need to made more user-friendly in implementations such as Music Assistant, just take a look at this video from FutureProofHomes that show what they currently must do in Music Assistant UI in order to setup a Snapcast stereo-pair:

• https://youtu.be/n5yNEhcAjWU?si=h96CfVtdpn_1UgWf&t=966

So today, using their ESPHome-based Satellite1 speakers) and the Snapcast player provider, they first need to add both speakers to a group player and then they have to go into the ESPHome integration and from inside there set the speaker channel out put to left or right channel only on each ESPHome-based speaker device.

PS: I know this exampel is showing using Snapcast and not Sendspin, but only wanted to note that the integration implemntation for Music Assistant player provider for both really needs to be made more user-friendly so it gets much easier for end-users. In my opinion the ultimate goal should be that end-user will never need to go into ESPHome settings.

[eugenio]

+1 on stereo pair support — and I'd like to highlight that this is the foundation for a much bigger feature: discrete surround sound.

I'm planning a 9.2.1 Dolby Atmos setup using Louder-ESP32S3 boards as wireless surround/height speakers (front stage wired to AVR). Sendspin's 50μs sync is already better than WiSA for this use case — what's missing is the channel routing.

Stereo pair (L/R assignment) is the 2-channel case. Surround is the N-channel generalization. I've opened two companion issues to track the full surround path:

• [REQUEST] Server-side multichannel audio splitting for surround sound (5.1/7.1/9.x Atmos) use cases #74 — Server-side multichannel audio splitting
• [REQUEST] UI/protocol support for surround channel assignment per client (speaker role mapping) #75 — UI/protocol support for surround channel assignment per client

Would love to see stereo pair land first as it unblocks the surround workflow.