default AVRCP raises volume instead of lowering if volume is below ~128

[modpotato]

Problem Description

when pressing the volume down button on a bluetooth speaker (using a2dp source) if the volume is greater than 127 (approx) it raises the volume instead.

Device Description

esp32-wroom-32d

Sketch

#include "BluetoothA2DPSource.h"
#include <Arduino.h>  // for random()

BluetoothA2DPSource a2dp_source;

// callback
int32_t get_sound_data(Frame *data, int32_t frameCount) {
  // generate white noise
  for (int i = 0; i < frameCount * 2; i++) {
    int16_t sample = random(-32768, 32767);
    data[i] = sample * 0.01;  // apply volume level
  }

  return frameCount;
}

void setup() {
  randomSeed(analogRead(A0));  // seed the random number generator
  a2dp_source.start("SRS-XB23", get_sound_data);
}

void loop() {
}

Other Steps to Reproduce

specifically using a sony SRS-XB23 bluetooth speaker

Provide your Version of the EP32 Arduino Core (or the IDF Version)

2.0.13

I have checked existing issues, discussions and online documentation

• [X] I confirm I have checked existing issues, discussions and online documentation

[pschatzmann]

Strange that your speaker is even sending values > 127. I committed a correction, but I am not able to test this myself since I don't have any speaker with volume control.

[modpotato]

sorry for the late reply, it seems to have caused a different problem. when connecting the volume is set to 0 and from there + and - both add more volume until it hits the limit, which then the + button does nothing and the - button resets it to 0 (or 2 for some reason)

the following logs went as follows: (speaker) power on -> (speaker) bluetooth pairing -> (esp) connect -> (speaker) volume down x10 ->(speaker) volume up x20

https://logpaste.com/PG0DSl7e

[pschatzmann]

The logic was wrong indeed: Second trial

[modpotato]

thanks, thats solved the volume behaving oddly at high levels but the volume down button still doesnt work, these are the logs for when pressing volume down: https://logpaste.com/B1yOlSk7

would it be possible to add a callback any time an avrcp request is made for a volume change?

[pschatzmann]

I am printing all events that have been received. I can't update the volume if I don't get any reasonable updates from the source...

[modpotato]

the speaker works fine with changing the volume of my phone (ios 16) so i assume its not at fault, my only other idea is something with these lines

15:24:10.894 -> [ 73759][I][BluetoothA2DPSource.cpp:928] bt_av_notify_evt_handler(): [RCCT] Volume changed: 125 15:24:10.894 -> [ 73769][I][BluetoothA2DPSource.cpp:931] bt_av_notify_evt_handler(): [RCCT] Set absolute volume: volume 127

it seems as though its being notified to lower the volume, yet sets the volume to another value though i could be wrong as i havent really looked at the source

[pschatzmann]

This logic has been taken from the Espressif example code: For some reason they increase the received value always by 5. If you press the volume down, I would expect that we should receive decreased [RCCT] Volume changed: values.

There is no logic in place to increase or decrease values: we just take the values that have been received. You can try to play around with the code to figure out what is going on...

[modpotato]

it appears it increases the volume by 5 for literally no reason. https://github.com/pschatzmann/ESP32-A2DP/blob/aeb07ef844d6c307ea6fb70610bd0abee6690dad/src/BluetoothA2DPSource.cpp#L930

my speaker goes changed values 3 at a time already so it doesnt seem to need to be multiplied, but upon removing the +5 (rather replacing +5 for + 0 because i dont remember much c++) it works perfectly fine, if you could implement this in a way that isnt incredibly scuffed like mine id greatly appreciate it!

[pschatzmann]

I removed the addition of 5

[modpotato]

my reason for doing + 0 was because removing it threw this same error

[pschatzmann]

Ok, I see. It is compiling now again...

My guess the +5 was supposed to prevent a user from setting the volume to 0. A 0 value has the risk that the user might forget that the device is still switched on

[modpotato]

thanks