ANNOUNCE: Squeezelite-ESP32 (dedicated thread)

[sle118]

THX for the reply. I searched a lot here and in the old thread. I dont know what I do wrong atm in my config.
Your mentioned dac_config is one of the 2 mentined on github, see here:

https://github.com/sle118/squeezelite-esp32#esp32-a1s

I am a bit confused what firmware I should flash exactly?
As I mentioned I flashed mule with the webinstaller. Is that right? Or should I try another base-image or another flash method?

You can get started by flashing the I2S firmware from here

https://github.com/balloob/squeezelite-esp32-install

This will get you the recovery partition and a working platform. Future firmware installations can be initiated from the web interface, and this method has the benefit of not requiring any extra download of flashing tools, etc

[philippe_44]

Good Morning

I have 3 ESP32-A1S in use, 1 with the AC101 DAC, the others ES388. They are running the I2S-4MFlash.16.1036.master-cmak firmware.

All 3 work fine in BT and Squeezelite mode, other than the fact that I have to rotate the encoder shaft more than I'd like to effect a volume increase. It takes between 4 and 5 full rotations to go from 0 to 100% volume. Is this a physical or software issue? I've had a look in the NVS and cannot see where the rate of volume increase can be adjusted.

It probably means some steps are missed, although I don’t remember what’s the case on a real SB

[philippe_44]

Yes.
I have flashed your last compilation again, 1.036, and still happens the same.
fw: I2S-4MFlash.16.1036.master-cmak, mode: Squeezelite-ESP32

As often ? If it does, I might try to send you my local build to check.

[yateng]

As often ? If it does, I might try to send you my local build to check.

Hi, thanks for your dedication.
I have flashed the new 1.1038, and the problem continues, in both of my reproducers...
Same frecuency

[yateng]

Hi. Flashed the 1.1040, and same problem with the art cover...

[philippe_44]

Hi, thanks for your dedication.
I have flashed the new 1.1038, and the problem continues, in both of my reproducers...
Same frecuency

How often (ballpark)

[yateng]

Hi.
If I push the fast-forward "button" in the LMS and force to skip to the next song, It's not happening again, so fixed?
But if I wait until the end of the song, so the song finishes and start the next song (or skip the current song to nearly the end of the song so it finishes and starts the next song), then like one third of the times the image is not fully displayed.
Thanks again for your wonderful job!

[philippe_44]

Hi.
If I push the fast-forward "button" in the LMS and force to skip to the next song, It's not happening again, so fixed?
But if I wait until the end of the song, so the song finishes and start the next song (or skip the current song to nearly the end of the song so it finishes and starts the next song), then like one third of the times the image is not fully displayed.
Thanks again for your wonderful job!

oh, 1/3 is a lot. Can you remind me what you typically play when it happens? Local or streaming?

[yateng]

oh, 1/3 is a lot. Can you remind me what you typically play when it happens? Local or streaming?

Hi.
Streaming, Spotify usually

[AngloCuencano]

Originally Posted by Yatsushiro

Good Morning

I have 3 ESP32-A1S in use, 1 with the AC101 DAC, the others ES388. They are running the I2S-4MFlash.16.1036.master-cmak firmware.

All 3 work fine in BT and Squeezelite mode, other than the fact that I have to rotate the encoder shaft more than I'd like to effect a volume increase. It takes between 4 and 5 full rotations to go from 0 to 100% volume. Is this a physical or software issue? I've had a look in the NVS and cannot see where the rate of volume increase can be adjusted.

It probably means some steps are missed, although I don’t remember what’s the case on a real SB

I built my first ESP32 player this weekend. I'm very impressed with the performance, so congratulations to everyone who contributed to this project.

But I do agree with Yatsuhiro that volume response from the rotary encoder is a bit lacking. I am using a Bourns encoder with 24 pulses/rotation, and I added some discrete components for debouncing. Each pulse is detected, and I can see in the LMS interface that one pulse changes the volume by 1%. So like Yatsuhiro, it would take more than 4 complete rotations to go from 0 to 100% volume.

If it were possible to increase the sensitivity to 3 or 4% per pulse, I think this would make the volume control more useful.

[philippe_44]

Hi.
If I push the fast-forward "button" in the LMS and force to skip to the next song, It's not happening again, so fixed?
But if I wait until the end of the song, so the song finishes and start the next song (or skip the current song to nearly the end of the song so it finishes and starts the next song), then like one third of the times the image is not fully displayed.
Thanks again for your wonderful job!

I've pushed a potential workaround. Let me know if it helps

[yateng]

I've pushed a potential workaround. Let me know if it helps

Hi, here it happens again, with the same frequency...

[philippe_44]

Hi, here it happens again, with the same frequency...

Grrr ... for a reason that I really cannot understand, the JPEG decoder fails silently from time to time. The data is not corrupted, it's just the decoder that ends. I've pushed a new release with a more efficient (I hope) workaround. Let me know

[edit]: it's not released, you just need to use the art(i/e)facts of the build. There is also an extra log that I'd really like to have

[yateng]

Grrr ... for a reason that I really cannot understand, the JPEG decoder fails silently from time to time. The data is not corrupted, it's just the decoder that ends. I've pushed a new release with a more efficient (I hope) workaround. Let me know

[edit]: it's not release, you just need to use the art(i/e)facts of the build. There is also an extra log that I'd really like to have

Hi, it has improved. Only fails 3 or 4 times every 20 tries, so it's much better.
The point is that when you force to skip to the next song using the fast forward, it doesn't fail.
Thanks for your work!

[philippe_44]

Hi, it has improved. Only fails 3 or 4 times every 20 tries, so it's much better.
The point is that when you force to skip to the next song using the fast forward, it doesn't fail.
Thanks for your work!

Could you get me the log (using telnet or a serial local connection)?