Audio: earphones give a horrible noise


Haraldur

Very Active Member
Joined
Mar 12, 2015
Messages
301
TL;DR: Horrible noise in earphones (apparently not speakers, except, perhaps, if I put my ear right against them). I suppose I should note that these are wired, not wireless, connected to the 3.5 mm jack, with no microphone (TRS). Am I alone to suffer this? Hardware or software problem? Can I fix it? Would an OS update fix it? Should I update my OS anyway (I have less motivation to do this if the audio is not fixed)? I just want to be able to listen to music through wired earphones on the bus etc., without unpleasant noise.

uname -a
Linux my-pyra 5.6.19-daveiii-pyradef-aufs #3 SMP PREEMPT Mon Dec 14 18:23:52 CET 2020 armv7l GNU/Linux

My Pyra is EU-LTE 4GB RAM.

-----

I got my Pyra in October 2024. I have not used it much as the headphone audio is ruined by noise, rather continuing to use my Pandora (which has a malfunctioning LCD cable and a broken hinge -- wear and tear! ;-) ). I have been pre-occupied with other matters so have not given much thought to it since.

I would like to know whether this is likely a hardware or software problem and, if the latter, what I could do to fix it. I would also be interested to know whether I alone have this trouble, or if it rarely occurs or if it is the common case. I have not used my Pandora for gaming for the last few years, mainly using it instead as an audio player (!), note taker and book reader. I also did some programming on it in the past. Anyway, I cannot use the Pyra as a replacement for the Pandora with the noise from the headphones as it currently is.

One interesting thing about the noise is that it correlates strongly with what is displayed on screen: if the displayed graphics change, the sound changes. Opening a window causes a 'whoom'-ish noise. I initially wondered if there might be some cross-talk from the LCD cable to the jack, but no amount of fiddling and detachment and re-attachment made any difference (incidentally, I first detached and re-attached it when I first got the Pyra as I had trouble turning it on and wondered whether the cable was attached properly -- I lack patience -- and am slightly worried that I may have made a mess of it as there is a component nearby that the cable seems to rest upon, and it bends it, which is worrying). Eventually all output to the earphones stops if nothing is played for a while, which likely rules out the cross-talk possibility. Output can apparently be forced, however, by going to the audio settings via the taskbar volume applet and activating the 'input' tab -- which is interesting behaviour in itself. This output is, of course, this horrible noise.

Changing the volume does nothing to the noise in my testing/fiddling. If the screen is closed (even if already off), a high-pitched whine is added (or a pre-existing whine lowers to a pitch that I can hear). The base noise sound sounds like sputtering, or like a muffled sound of high winds outside with flapping fabric. The volume of the noise with my headphones seems roughly equivalent to a sample of voice audio in the range -40 to -60 decibels -- but this is very rough, from the tooltip on the taskbar volume indicator and aural judgement while adjusting the voice audio volume/master volume.

AlsaMixer (v1.1.8): If I go to Letux Cortex 15, as sound card, I can remove the noise by setting Headset Left Playback, and Right, to Off, but then there is no sound at all. HS DAC gives sound + noise, Line-In only noise. Headset volume has the effect of adding to volume (but not reducing to nothing). The Earphone setting seems to do nothing. Incidentally, I am using simple earphones with no microphone, so my actions likely have nothing to do with why Headset stuff is relevant and Earphone apparently not. My fiddling also seems to have stopped speaker output... Oh, well.

I found a thread here once, which I struggle to find again but is probably this one from 2021: https://pyra-handheld.com/boards/threads/audio-via-headset.99601/ , where someone else complained of a similar-sounding problem and someone (I think EvilDragon) said something about the DSP and its registers for volumes from various sources and that if a certain hardware volume register value were decreased and a software value from later in the chain increased then this would be fixed. I have not found the time or motivation to investigate that myself. It is an interesting possibility, though. I have no kernel development experience, so if there were some way to prod those registers from the CLI then that would be much more convenient.

It would also be nice to know if power draw will decrease with software updates, as with my (weak?) mobile phone charger it can only charge (slowly) if the screen is off (and WiFi and Bluetooth as well).
 
Back
Top