Ok, first of all, this is with v3alpha3, and ogg files. So I'm not looking for a fix, but maybe someone has some ideas about what's happening.

Before I turn off the ignition I pause (v3a3 doesn't seem to remember where in an Ogg file payback is at) playback on the empeg. Turn off the ignition, empeg goes into standby for a few seconds. If I leave the empeg in the car and turn the ignition on later the playback is stuttering. None of the controls seem to have any effect.

I have to pull the empeg out and reseat it and then playback is fine. I've downloaded the file and verified it with ogginfo, so it's not a file or container format issue, as the FAQ suggests for mp3 files. This hasn't happened on an mp3 yet, but I've got over 5000 ogg files and maybe a hundred or two mp3s.

The problem has never happened when I've seated the player after the car has started.

Any ideas?

--Nathan

In theory, pulling the player out and reseating it should be functionally the same as turning off the ignition, letting the player go into full-power-off mode (instead of standby) and then restarting the ignition. So I'm surprised to hear that you get two different behaviors in that situation.

I also get it, I assumed it was something to do with the serial port but haven't actually tried to figure it out yet, because the easily accessible serial port was in the car, and we now only have the van.

Any chance it's related to the IDE cable in some way and the reseating jars the loose connection in there enough to establish a solid connection?

Stu

I'll run some more experiments and let you guys know, specifically seeing if there's a difference between inserting the player before and after starting the car vs. leaving it seated.

It's a second hand unit, but was someone's spare, so I have no idea how much aging the IDE header could have taken. This is the only weirdness I've seen, and the the unit comes into the office every day and sits in a mlord dock until I leave. This is the only weirdness I've noticed, so if it were a loose connection type of thing, I'd expect to have other stuff happen, but who knows?

--Nathan

I've found (well, in fact - no, I didn't, my 7-year old Bug Hunter William did) that this is not just the Ogg Decoder, it is sometimes also the case with the WMA and MP3, indicating it might actually be something other than the decoders themselves, but something else in the player software structure.

Quick convenient fix to stop this: before you turn your ignition off (assuming your ignition is properly wired, of course!) press PAUSE, then SLEEP, then IGN off, then wait until your power off timeout expires before removing the player unit from the sled. I have set my power off timeout to 5 seconds as a result.

Oh, and by the way - I didn't even work out this fix MYSELF - William did: if ever there is a software engineer in the making, this 7-year-old is IT. I am appalled, and at the same time surprised and gratified that I seem to have some influence on the gene pool. What's worrying is that a 7 seven year old with zero training is better than I am at bug observation/fixing than I am as a professional software engineer with some 20 years experience behind me!! Jeez.....

I would tend to agree, this is beginning to sound like a dodgy cable. However, I have seen similar behaviour on a player with a known good cable that worked fine on V2 final, and then started to do this after upgrade to V3A3. I suspect that for now, the only way to test, is to FIRST attempt a downgrade to V2.00 for a few days of continuous use (at least) and see if it still happens. If it does, then it's cable. If it doesn't re-occur, then it's the "A3 Stutter" problem.

Don't forget there was a similar problem in ( I think) V2B10/11, which was fixed in B13. Rob's major warning at Amersfoort 2002 was that the release could have problems in it that had been fixed in 2B13 but had not been ported across to the V3 tree for the player - hence problems you thought had already been fixed, may in fact re-occur. Be wary of this - maybe this is the case with this problem.

Nope. Let's just say I had occasion recently to determine this for sure and leave it at that.

playback is stuttering. None of the controls seem to have any effect.

This can be a symptom of the player running out of memory. v3alpha3 uses a lot more
memory than its predecessors. On my 12MB mk2 player, v3alpha3 runs out of
memory simply by having more than 7000 tunes in the database. The stuttering
you describe are accompanied by lots of VM messages on the serial port.

I had to reduce the tune cache to make v3alpha3 work for me.

IIRC, Ogg tunes can demand extrordaniry resources from the decoder,
making your player run out of memory.

Pim

Re: memory limitations.... Interesting. I hadn't noticed that, although I had seen that ogg playback does seem to require more CPU power. I'll have to plug in the serial cable at some point and see what happens. I think I've got just over 6k songs in the player right now.

On the day I first posted about this I had the stuttering problem twice, and haven't seen it since. Go figure.

The stuttering just happened for the first time on AC power, so I was able to get a serial cable attached.

VM: do_try_to_free_pages failed for player...
VM: do_try_to_free_pages failed for player...
VM: do_try_to_free_pages failed for player...
VM: do_try_to_free_pages failed for player...
VM: do_try_to_free_pages failed for kswapd...

Guess it's time to go digging into the FAQ about tunings for freeing up physical memory.

--Nathan

You're probably interested in ReserveCache.

I did consider this, and I suppose I should try it as well, but I'm curious why 2 subsequent boots would result in stutter once and no problem the next time: the playlist didn't change.