Muse Radio Squeezelite

42

[quote=“Manul, post:41, topic:1099”]

  • The device is beautiful and very well built. I was a bit unsure when ordering it as it is listed as “wood color” and I feared some cheesy plasticky fake wood surface. The front is also very nice looking and both dials turn smoothly. The knobs and buttons feel solid.

Thanks

  • Sound quality is a bit of a mixed bag. Overall it’s quite good, but the noise / static level is a bit higher than I would have hoped or anticipated. I’ll find out how much an annoyance that is for my intended use case (bathroom radio) in the long term, could well be one of the things one notices when the device is new but forgets in the long run. Time will tell.

If you have the time to do it I will be curious for you to try the default standalone radio app from apps.rasipaudio.com to check if you have the same static noise? If you have not it will mean that we need to check the codec commands in Squeezelite

  • @Raspiaudio, you mentioned that the display blanking is something that can be set in NVS. I didn’t find the corresponding setting, how would I change that? And is it possible to have it behave differently when it’s plugged in vs. when it’s running on battery?

it’s I think in the LMS pluging SqueezeESP32 settings

  • I understand that the buttons can currently not be used in squeezelite. That is a bit of a disappointment and I think something that should be mentioned on the store page. Out of curiosity, what was the reason to implement them with ADC/resistors? Were there not enough GPIOs?
    Yes it was gpio numbers related, no worry I believe that it will be done one day
43

I did that and it does indeed seem that the noise is much lower. I tried (without success) to find a stream of silence to be sure, but even with the radio constantly playing, noise seemed much better. Could you try whether you can reproduce the high noise level with squeezelite and LMS? What would be needed to reach better sound quality? Can I help in any way? Or do you have some tips how I could experiment on my own?

I did not find it there - nor in the NVS configs as mentioned before. Any more ideas?

I do not worry. I just would have liked to know in advance. I would probably have ordered the radio anyway but I would have know what to expect and not been disappointed.

Any thoughts on or answers to my other questions? Although I’d say the most important right now would be to fix the noise issue - it’s really a bit annoying. Is anybody else here using the device as an LMS player and do you experience the same issues?

44

I use the device as a Lyrion player. There’s a slight noise, especially with quiet music, but it doesn’t bother me. If the device doesn’t play any sound, the noise disappears.

45

Thanks for the feedback. At least that probably means I don’t have a defective device. Now the question remains whether it’s a software issue with squeezelite or an actual limitation of the hardware.

46

Issue Summary:
When playing a stream on LMS, there is small but noticeable background noise at low level. The noise does not appear to depend on the volume level, making it easier to hear at lower volume levels, particularly with music containing silent passages, such as classical music.

Observation:
The noise appears to be significantly lower when using the Radio demo app under the same conditions, reaching an acceptable level (only audible when listening very close to the speaker).

To-Do:

  • Compare the codec initialization sequence between Squeezelite and the Radio app.
  • Test the differences.
  • Submit a pull request if improvements are identified.
  • It is necessary to measure the background noise using a calibrated microphone.
47

I think I’ve figured out the issue, even if I haven’t managed to solve it just yet:

In the dac_controlset, the I2C sequence for volume control is not configured. As a result, Squeezelite defaults to using I2S software volume control on top of the codec forced to max gain. This explains why the noise level is not proportional to the volume. The good news is that if we fix this, there should be a noticeable improvement in audio quality, with less quantization and reduced noise at lower volumes.

However, my current volume sequence doesn’t seem to have any effect on registers 46 and 47 when I test it using the I2C debug tools available in the web interface.

Here’s my current configuration:

{
“init”: [
{“reg”:3,“val”:24}, {“reg”:24,“val”:2}, {“reg”:38,“val”:9}, {“reg”:39,“val”:144},
{“reg”:42,“val”:144}, {“reg”:43,“val”:128}, {“reg”:45,“val”:128}, {“reg”:27,“val”:0},
{“reg”:26,“val”:0}, {“reg”:2,“val”:240}, {“reg”:2,“val”:0}, {“reg”:28,“val”:32},
{“reg”:29,“val”:32}, {“reg”:4,“val”:60}, {“reg”:25,“val”:0}, {“reg”:46,“val”:33},
{“reg”:47,“val”:33}, {“reg”:48,“val”:33}, {“reg”:49,“val”:33}
],
“volume”: [
{“reg”:46,“val”:“volume * 33 / 100”}, {“reg”:47,“val”:“volume * 33 / 100”}
]
}

I have opened a github issue hopping that the 2 gurus will answer.