ANNOUNCE: Squeezelite-ESP32 (dedicated thread)

[ElFishi]

I have the hunch that the difference has something to do with encoding ...

Nope, it was the equaliser. I am quite sure I never touched it, but somehow it had non-zero entries.

[sle118]

All,

I just released build dev.526, which contains a fix that will reduce the burden on the system when refrehing the web UI. I am still working on some additional changes to the menus and firmware update page so they look nicer on a mobile... and also hopefully work on Apple devices.

Stay tuned!

[mausginh]

ESP32-A1S Display

Hi,
I worked with the Thinker A1S Board. Other than mentioned by me, there is a way to work with another SCL than GPIO21 (which conflicts with the amp). SDA at 12 and SCL at 14. Works fine with SSD1306.

[mausginh]

ESP32-A1S Display

Hi,
I worked with the Thinker A1S Board. Other than mentioned by me, there is a way to work with another SCL than GPIO21 (which conflicts with the amp). SDA at 12 and SCL at 14. Works fine with SSD1306.

[mark wollschlager]

All,

I just released build dev.526, which contains a fix that will reduce the burden on the system when refrehing the web UI. I am still working on some additional changes to the menus and firmware update page so they look nicer on a mobile... and also hopefully work on Apple devices.

Stay tuned!

Seems to be pretty stable for me.
Thanks.

[MumPf]

Messages in Status

Hi
great Project... thanks for all !

I don't know if it's a problem? I have a lot of messages in status... a new entry every second with version 526.. see picture:



with version 148 it's ok:



I've a AI-Thinker-ESP32AudioKit-V2.2 and nothing in jack, only 2 speekers on amp.

I can clear the messages, but after some minutes I have a lot of messages again. Is this a problem? Maybe problem with memory after some hours?

Thank you very much for all your work. Realy great...

[mboxler]

I assume you've set dac_config with bck, ws, do, model, sda, scl and i2c properly. Then, "dac_controlset" should look like

{ init: [ {"reg":54,"val":17} ] }

PS: please verify the value. You might have to set register 0x35 to 0x00, but I'm not sure

Finally getting back to this.

I think I have things set up properly, but when I execute the "Scan I2C bus for devices" command, I get nothing...



Does this mean the ESP32 isn't detecting the evaluation board?

Mike

Edit...Just hooked up an I2C display and got the same results. This command must be used for something else. Sorry

[funkybassman]

I've a AI-Thinker-ESP32AudioKit-V2.2 and nothing in jack, only 2 speekers on amp.

I can clear the messages, but after some minutes I have a lot of messages again. Is this a problem? Maybe problem with memory after some hours?

Thank you very much for all your work. Realy great...

I have the same board and use the on board amp only. I stopped these messages by removing the jack configuration from 'set_GPIO' tab on NVS editor.
I hope that helps?

[sle118]

Hi
great Project... thanks for all !

I don't know if it's a problem?

with version 148 it's ok:

[ATTACH]32262[/ATTACH]

I've a AI-Thinker-ESP32AudioKit-V2.2 and nothing in jack, only 2 speekers on amp.

I can clear the messages, but after some minutes I have a lot of messages again. Is this a problem? Maybe problem with memory after some hours?

Thank you very much for all your work. Realy great...

I'll come up with a fix soon. Thank you for reporting!

[KeBul]

I tried 526 on my A1S Audio Developer board, but it was a real struggle, OTA worked but device went from slow to very slow, lost rotary encoder and buttons despite them being configured correctly, in the end had to recover with download tool re-flash because the device became completely unresponsive.

Tried erase and then fresh download of 526, still no encoder or buttons... eventually traced this to "too many CLI instances in LMS" reboot of LMS fixed it. Also playback was less reliable, lots of pauses, stops, loss of connection and skips to next track... wifi disconnected, wifi connected.
Checked I had changed wifi mode to Comprehensive (a in NVM editor) and yes that was set, but checked my access points and it was connected to the one furthest away that just happened to be Channel 1, repeated this behaviour with a brand new previously unused A1S, on 526 both pretty much always connected to my Router on Channel 1, two floors and two walls away, rather than the access point in the same room on channel 11.

Went back to the trusty 142 on both devices and all is well again, they both connect to the nearby access point every time, although I hit the "too many CLI instances in LMS" again and had to once more restart LMS to get encoder and buttons working.

...and now just noticed on the current one powered up the equaliser doesn't work... getting this in the LMS logs:

[20-11-21 21:06:55.7461] Slim::Web::JSONRPC::requestMethod (455) Request failed with error: Bad dispatch!
[20-11-21 21:06:56.7094] Slim::Control::Request::execute (1888) Error: While trying to run function coderef [Plugins::SqueezeESP32::Plugin::setEQ]: [Can't locate object method "send_equalizer" via package "Slim::Player::SqueezePlay" at /usr/local/slimserver/Cache/InstalledPlugins/Plugins/SqueezeESP32/Plugin.pm line 98.
]

Time for a cleanup in LMS caches and remove/re-install SqueezeESP32 plugin I think, must admit I'd had enough after messing around with them for many hours, I'll take one back up to 526 again and re-check the wifi behaviour when I feel like having a play again.

Edit:

One more thing I found with 526 - I could not get it to fully populate the squeezelite startup line for bluetooth, even though I'd put the sink device name in at the bottom of the page, and then save and apply
I always got:
squeezelite -b 500:2000 -d all=info -C 30 -W -n "squeezelite" -o "BT" -R -Z 192000"
I had to edit in a name manually in NVS:
squeezelite -b 500:2000 -d all=info -C 30 -W -n "squeezelite" -o "BT -n 'SoundCore 2'" -R -Z 192000"

I've not got screenshots of what I was entering where - I'll double check it when I try 526 again.


Kev

[sle118]

I tried 526 on my A1S Audio Developer board, but it was a real struggle, OTA worked but device went from slow to very slow, lost rotary encoder and buttons despite them being configured correctly, in the end had to recover with download tool re-flash because the device became completely unresponsive.

Tried erase and then fresh download of 526, still no encoder or buttons... eventually traced this to "too many CLI instances in LMS" reboot of LMS fixed it. Also playback was less reliable, lots of pauses, stops, loss of connection and skips to next track... wifi disconnected, wifi connected.
Checked I had changed wifi mode to Comprehensive (a in NVM editor) and yes that was set, but checked my access points and it was connected to the one furthest away that just happened to be Channel 1, repeated this behaviour with a brand new previously unused A1S, on 526 both pretty much always connected to my Router on Channel 1, two floors and two walls away, rather than the access point in the same room on channel 11.

Went back to the trusty 142 on both devices and all is well again, they both connect to the nearby access point every time, although I hit the "too many CLI instances in LMS" again and had to once more restart LMS to get encoder and buttons working.

...and now just noticed on the current one powered up the equaliser doesn't work... getting this in the LMS logs:

[20-11-21 21:06:55.7461] Slim::Web::JSONRPC::requestMethod (455) Request failed with error: Bad dispatch!
[20-11-21 21:06:56.7094] Slim::Control::Request::execute (1888) Error: While trying to run function coderef [Plugins::SqueezeESP32::Plugin::setEQ]: [Can't locate object method "send_equalizer" via package "Slim::Player::SqueezePlay" at /usr/local/slimserver/Cache/InstalledPlugins/Plugins/SqueezeESP32/Plugin.pm line 98.
]

Time for a cleanup in LMS caches and remove/re-install SqueezeESP32 plugin I think, must admit I'd had enough after messing around with them for many hours, I'll take one back up to 526 again and re-check the wifi behaviour when I feel like having a play again.

Edit:

One more thing I found with 526 - I could not get it to fully populate the squeezelite startup line for bluetooth, even though I'd put the sink device name in at the bottom of the page, and then save and apply
I always got:
squeezelite -b 500:2000 -d all=info -C 30 -W -n "squeezelite" -o "BT" -R -Z 192000"
I had to edit in a name manually in NVS:
squeezelite -b 500:2000 -d all=info -C 30 -W -n "squeezelite" -o "BT -n 'SoundCore 2'" -R -Z 192000"

I've not got screenshots of what I was entering where - I'll double check it when I try 526 again.


Kev

If possible, please log some defects in GitHub. It is currently very difficult for me to commit to the project and do proper regression testing. If 142 works, then please stick to it. Is anyone else using 526 or 529? If so, is it any stable? I'd like to know.

[KeBul]

If possible, please log some defects in GitHub. It is currently very difficult for me to commit to the project and do proper regression testing. If 142 works, then please stick to it. Is another else using 526 or 529? If so, is it any stable? I'd like to know.

Sure, will do once I've clarified and proved the issues are 526 - quite happy to stay on 142, but thought I'd try 526 hoping to report back some good news on it, but to be honest I had such a torrid time with it, in the end clear thinking and troubleshooting became difficult.

Kev

[ElFishi]

Is anyone else using 526 or 529? If so, is it any stable? I'd like to know.

fw: I2S-4MFlash.526.master-cmake
config as attached

just experienced it stop in mid-play (synchronised with another ESP). no automatic reboot, no response to telnet or web interface, just stuck.
power cycle brought it back on.

[KeBul]

Sure, will do once I've clarified and proved the issues are 526 - quite happy to stay on 142, but thought I'd try 526 hoping to report back some good news on it, but to be honest I had such a torrid time with it, in the end clear thinking and troubleshooting became difficult.

Kev

Hopefully final update on this here...

Last night I started again, OTA from 142 to 529 (thought i might as well test the latest). As before OTA worked fine and upgrade completed.

This time I was more organised and had putty connected to the device so I could see what was going on...

1. The poor response from the web interface is because the device is continuously looping through a crash and core dump... ESP_ERROR_INVALID_STATE in bt_app_sink.c I think it does eventually recover from this and starts responding normally
2. The sound pauses and skipping to next track is caused by wifi disconnects and subsequent underflow error
3. The wifi mode seems to be working correctly, testing last night and this morning showed it was connecting to the nearest AP the majority of the time, but with so many disconnects sometimes it would connect to an AP further away - but that's wifi! Having said that I'm currently listening on i2S with wired headphones and it's still disconnecting but seems to be favouring the AP furthest away on Channel 1 again
4. Both SqueezeESP32 and LMS report the CLI connection refused too many connections error when it happens, Reboot LMS clears, assume the excess connections are due to the "dirty" wifi disconnects
5. Filling in the Bluetooth sink name in the web interface and saving changes the name in NVS but it doesn't put that name in the Squeezelite startup line. Easy to fix in NVS editor by adding -n 'your sink name'" after "BT in the squeezelite startup line in autoexec1

I've got output and screenshots if relevant for each issue, will head over to GitHub and log them there.

Kev

[sle118]

Hopefully final update on this here...

Last night I started again, OTA from 142 to 529 (thought i might as well test the latest). As before OTA worked fine and upgrade completed.

Kev

I have installed rev 529 and it has been stable for me. The main difference, however, seems to be the hardware platform; you are running on the ESP32-A1S build and I'm running the I2S 4Mb build on a DAC that does not have i2c. I also noticed that you are sharing the i2c bus between the amp board and the display and I am wondering if this could be a source of conflict, as I do believe our core assumes a separation between i2c for amp and display (I'd have to check that).

Could you please try to deactivate your display and see if this helps? Something else you could try is attaching your display's sda/scl to different GPIOs.

I think it's time I order one of these boards so I can include them in my tests going forward.

Thank you