[SOLVED] Sound corruption after upgrade on 3g ipod

[alleyoopster]

Just upgraded to 17131 tonight and immediately noticied a sound problem when playing back any format (ogg, flac or mp3). The problem starts when I navigate away from WPS. It is a light clicking noise and seems to switch from left to right channels. It is very obvious and does not make for good listening. The only way I can see of resovling is to reboot. but it will return after a while or when I navigate. This is on a 3G 20GB ipod.

thanks for a great bit of software

[soap]

1 - What version were you using before the upgrade? (roughly if you can't recall exact.)
2 - If you upgrade now to the latest, can you reproduce this problem?

[alleyoopster]

1. I upgraded from 17091, I tried the latest (17138) and this didn't work at all for some reason, it would not boot, just an apple. I have been upgrading a lot in hope that another problem gets fixed (idle timeout causing reboot when charging)
2. I could always reproduce the problem by navigating away from the WPS. I have downgraded back to 17091 and all is working fine again. (Didn't want to be without music for the day) I am happy to try and reproduce again if needed or I can try the next build.

[soap]

Keep trying the current build, once you have successfully tested a current build (and if the problem persists) we will move on from there.

[yapper]

I'm getting a similar issue with r17148 on a 2nd Gen iPod. The 'clicking' stops if a track is paused, but returns within 10 or 20 seconds of playback being resumed.

Daily build 2008-04-15 r17111 wouldn't boot (froze after Apple logo).
Daily build 2008-04-14 r17103 also has the 'clicking' problem.
Daily build 2008-04-13 r17091 seems to be okay.

[Llorean]

Could you try to narrow it down to the specific revision, then, and file a proper bug report?

[Buschel]

Hi,

I'm very interested in checking whether r17097 introduced this bug. As owners of PP5002-based targets are quite rare: Is someone of you able to build rockbox by his own? If so, checking r17096 vs. r17097 would really help tracking this issue down.

[alleyoopster]

I'll compile them both tonight and if there are no problems I'll give them a try tomorrow.

[yapper]

I built with the current version of pcm-pp.c (revision 17097) and the problem was present.

I then rolled back to the previous version of pcm-pp.c (revision 16981) and the problem disappeared.

[soap]

I built with the current version of pcm-pp.c (revision 17097) and the problem was present.

I then rolled back to the previous version of pcm-pp.c (revision 16981) and the problem disappeared.

Do you mind hooking up alleyoopster with the same two test builds and see if we can't get a conformation on this?

[yapper]

He's got a 3G?  I'll try and get them built (mine is a 2G).

[amiconn]

I was able to reproduce the problem on my 2nd Gen. Unfortunately, the problem tends to go away during intensive testing, and so I can't be 100% sure it was caused by r17097.

I'm quite sure it is though, as I found a bug in the SVN assembler code for PP5002: It also enters the loop and writes a pair of samples if there's no or just one free FIFO slot. I prepared a fix, put a test build on my 2nd Gen, and prepared a test build for alleyoopster (3rd Gen). Attached the patch if someone else wants to test.

Will use the test build tomorrow during the day, and if it really fixes the problem, commit tomorrow night.

[yapper]

Unfortunately, the problem tends to go away during intensive testing

I noticed that too - despite using the same track repeatedly.

I'm testing using r17153 patched with amiconn's patch, and it seems to be good.

NOTE: r17153 unpatched doesn't boot (screen goes blank after the apple logo)

[alleyoopster]

I will continue with the r17152 amiconn patched last night. It seems good so far. The ipod is working for a good part of the day, so I'll give it a good soak today.

thanks guys

[yapper]

amiconn committed the fix with r17158