Bad Sound Quality

[HackModford]

I'll try that too. I have a microSD that's a Sandisk and a sandisk usb reader. I didn't experience any improvement. So I wasn't running the pand or the rom through the SD slots but through USB

Edit: I can't play mp3s? I've got the codec pack.

Edit: I tried playing off the NAND and no difference. INteresting thing is Mplayer/Exaile explicitly follow the rules in the .asoundrc file. So I think we should use those programs to test the sound issue instead of emulators. When I used Ed's beta settings you get no audio because they're not able to open the audio device. My settings worked.

 

[Spirit]

I've had no glitches with "mplayer -ao sdl" so far if that helps.
edit: Actually, that's not true anymore.

 

[God Ginrai]

Heh, the patch seems to break some sound stuff.
PSX4ALL (now already with crystal clear sound) doesn't work with the new settings, only with the old ones.

ALSA is haaaard...

Crystal Clear sound? What are you talking about? The sound in PSX4ALL is scratchy as hell for me, and I don't have any sound problems with my Pandora.

-God Ginrai

You don't have the latest version I always get testing versions before they are up. Expect that one soon

You have me so jealous right now. T_T

-God Ginrai

 

[fusion_power]

I already would be happy if this really is only a Software issue and not a Hardware Problem.

Is there anything you DON'T think it's a hardware problem?

I mean, I've seen you commenting on EVERY thread on nearly every issue that you hope it's not a hardware problem (even if it clearly IS a software issue).
Next thing would be... "Hm, we don't have a Pong Emulator yet" - "I really hope it's no hardware problem!"

Geeez.
We're working on it, the unit to notaz is on the way, but there really is no need commenting each thread with thinking about hardware problems...
We'll see when he debugs a bit.

Hm, I don't remember this at all. ^^
Wel, I even don't remember what I've eaten yesterday at the evening meal but with my Swiss Cheese brain this is nothing unusual. :lol:
I try to not repeat me but I often forget what I wrote here and there so it can be that I repeating Stuff. ^^"""
And hope dies at last, so I continue hoping here and there...I also hope to win in the lottery and if you a not_so_evil_dragon I maybe spent some Millions to the Pandora Team.

^ Hehe, I once compared him on the German boards to the two dudes from the Muppet Show
(I don't mean this offensive, as it often enough fits well to my dark sense of humor )

Aha, You are EVIL_Mali!

 

[mali]

^ Don't worry, fusion The team is a bit stessed at the moment and I know you are a supporter, seing how much time you spend on the boards

 

[EvilDragon]

Okay I have investigated this a little bit more and you're not going to be happy with what I have discovered, I think:

1. The *type* of SD card you have in your Pandora effects how strange the audio gets. Some blocking operation on SD i/o - on *some types of card* - causes massive underruns of ALSA. I have gone through 3 SD cards - Fujitsu, NONAMEBRAND, and SANDISK Ultra II .. only the Sandisk card provides superlative audio. The others, on some i/o operations it seems (forced by sync;sync;sync while playing audio with Exaile or games with Picodrive), stall the audio immensely and produces glitches/noise during playback. The noise is really massive spurts of buffer underruns happening concurrently during the i/o op on the SD card, occurring so fast that it appears to sound like noise/overdrive.

2. I/O on the WLAN *also* causes audio glitches. On my two Pandoras, if I turn off wlan and put the SANDISK ultra II card in, I always get perfect Exaile playback. With one, if the WLAN is active (wget in the background), it glitches occasionally. With the other, it glitches often and more frequently for some reason. However with WLAN off, both play nicely.

Technically, I predict that in some driver - either WLAN/SD or ALSA - there is a lock that should not be being taken .. or such lock usage is simply broken. Could it be the display driver, also? Maybe I'm misinterpreting the initial stalled i/o op report as the first in a chain of stalled i/o ops that occur during the *scroll of my Terminal window*. A bit of debugging of the interaction of the WLAN/Audio/SD/USB/Display drivers is going to be needed to fully answer this question, I think.

Anyway I do not think this is hardware related, but bad-quality hacks in some driver around these parts, coming to light through more testing than the driver-developer had the luxury to afford themselves during development ..

Wow, actually, THANKS torpor!

I AM happy, since notaz is already working on buffer underrun problems on the ALSA kernel driver and had some success!

This means we're on the right way! Cool

 

[Monk]

The Muppets ws a brilliant show. Truly inspired

And I like the 2 old geezers!

 

[Bosbeetle]

statler and waldorf

 

[x68000]

^ or Waldorf and Statler

 

[notaz]

Technically, I predict that in some driver - either WLAN/SD or ALSA - there is a lock that should not be being taken .. or such lock usage is simply broken.

All this could be scheduler related too, maybe it gives too much priority to I/O tasks and doesn't preempt enough. Need to test various things but I'm away from home now.

My current progress improves underrun consequesnces but underruns still happen just like they did, still much work to do.

 

[GuSec]

Extremely interesting issue, clearly seems to be a software conflict with ALSA (confusing stuff)! This is the prize for running this young hardware/software (now)!
I hope a solution will be found soon, it don't believe it to difficult to fix, once the problem has been located. (Hotfix 3?)

 

[torpor]

Also might be worth upgrading to the latest ALSA, because I seem to recall there being a recent bug related to underruns in ALSA, but I don't have the fu to provide a link right now. But, perhaps, a radical leap to the latest version would be in order ..

 

[fusion_power]

Extremely interesting issue, clearly seems to be a software conflict with ALSA (confusing stuff)! This is the prize for running this young hardware/software (now)!
I hope a solution will be found soon, it don't believe it to difficult to fix, once the problem has been located. (Hotfix 3?)

Yupp, I also find this interesting. SD and WIFI can cause bad sound Quality, this wouldn't come in my mind when I would try to find the reason of bad sound quality. But I'm not such a tech guy so it's cool that we have at least a direction for looking at the cause of the Problem.

For me as noob it sounds like this ALSA never was made to work with a OMAP ARM Chipset well, what had Texas Instruments in mind which Sound driver should be used to work 100%?
And if this ALSA problem is known also from the PC, why is it still "unsolved" even if ALSA exists since 1998? Is there an "bullet proof" and always working method to prevent any buffer underruns 'n stuff?
What would Brian Boitano do?

 

[mali]

ALSA is used on the N900 too, which has an OMAP3. Exophase has always been against ALSA, cause it's a resource hog. Whatever, ALSA works on N900 and will be made to work properly on Pandora sooner or later, I'm sure.

 

[fusion_power]

ALSA is used on the N900 too, which has an OMAP3. Exophase has always been against ALSA, cause it's a resource hog. Whatever, ALSA works on N900 and will be made to work properly on Pandora sooner or later, I'm sure.

Sounds good. I guess the Nokia Guys are no magicians so if they were able to solve this problem, "we" can do this too.
Yes, some time ago I read somewhere, that Exophase is not a friend of ALSA, and now I know why. :lol: He is a coder, he could write something new and efficiant.
But I guess there is no real alternative to ALSA for the Pandora at the Moment. ^^"

 

[EvilDragon]

I've also never been a friend of ALSA, as it caused all kind of weird issues on my desktop PCs (random order of two soundcards, even if I stated them correctly, crackling sound (yay!), etc.)

Using OSS4 on my PC, but as ALSA exists for our OMAP, fixing it seems easier than using a new one that doesn't support it yet

 

[HackModford]

There was an update available for ALSA-state so I upgraded...

Still the same...

Edit: I'm gonna mess around with Pulseaudio and see what happens. (I really have no idea what I'm doing but never underestimate the ingenuity of fools)

Edit: Does anyone have any idea how to get Pulse audio setup? I've installed pulseaudio, pulse audio mixer thingy, a now pulse audio preferences

Edit: When I go to pulse audio mixer it says "connection failed" I'm gonna reboot to see if that makes a difference. Also after updating Alsa-state my asoundrc is blank

Edit: bahhh... I can't figure it out. even though I've installed pulseaudio there are no configs where there should be.... something I'd be very interested in seeing/trying is if another distro has this sound problem. i.e. can I try running ubuntu netbook remix of an sd and seeing if I still have a sound issue?

 

[notaz]

Pulse can't replace ALSA, it's just another layer on top of it, so you can only get more problems with this.

 

[torpor]

There is absolutely no point in looking at Pulseaudio until we get ALSA sorted.

I'm no fan of Pulseaudio (I use firewire on my home Linux box, Firepod 10 ftw, 10 channels of ins and outs, works plain and simple out of the box, except for .. Pulseaudio) but I will say this: if there is one place Pulseaudio should work extremely, beautifully, elegantly well, its the Pandora.

So lets get ALSA sorted first and then glom Pulseaudio on top. I really want this to be superlative, personally.

One thing we need is to find out what all the audio-capable apps for the .PND, right now, are using for output. IF they're doing ALSA callbacks, or if they're writing to /dev/dsp, we need to know that and use it to balance the debug results.

Does anyone know an app that *Absolutely* works, always, audio-wise? What is the best audio app so far, as far as you know .. ?

 

[HackModford]

If I was to make a wild guess in my situation it would be notaz's picodrive.