Audio drop out from CPU spiking in Tiger?...

[toodamnhip]

Just updated to Tiger....tested system extensively.
I am noticing audio drop out when monitoring. When recording to a stero track inside DP and checking the drop out, the drop out is not IN the stereo track, leading me to believe it is a firewire/828 issue or output issue. Has anyone else experienced drop outs in monitor play back with Tiger, 10.4?

<small>[ May 15, 2005, 10:28 PM: Message edited by: toodamnhip ]</small>

[spon-g]

hey,
i ve been having some issues too check my post ( http://www.unicornation.com/ubb/ultimatebb.php?ubb=get_topic;f=1;t=002784 ),
Mainly spikes resulting in audible artefacts/glitches..
Havent tried recording yet..

[toodamnhip]

ANy one else having these cpu spikes?///

[aaronlyon]

I was having trouble before Tiger arrived (not sure if the problem still exists). When too many of my soundbites started or ended simultaneously, I would get a CPU spike and a droupout in the playback. This also affected recorded tracks. For example, I had six vocals in a chorus, and I trimmed all the ends of the soundbites to the same point, just to be tidy. The problem went away when I trimmed the ends of the soundbites differently so they were staggered.

In another circumstance, I solved the problem by Merging Soundbites.

-aaron

[Timeline]

This may be related to too many access bites instead of continuas streamed files switching as a result of MOTU's DCPUM.

If you can, mearge the waveforms of the multipal bites together, do it. This way DCPUM won't be threatened to turn on and off.

Just a guess.

[Studio615]

I have solved this issue on my machine. I found it to be a firewire issue. I have a brand new 1.67ghz G4 powerbook with 1.5gb of ram, and an 828mkII. I am also using an external firewire 60gb 7200rpm harddrive for my audio. At first I was using both devices on the same firewire bus. I had steady dropouts at an even rhythm, basically the bus couldn't handle the amount of data for the drive and 828. I looked at the cpu meter in DP and it was running at about 1%, with overload spikes all over the place. So I then put in a firewire pcmcia card in the powerbook. I hooked up the 828 with this card and hooked up the drive to the built in one. It got better, but i was still getting dropouts and the cpu meter still spiked when I had the dropouts. I then loaded the firewire driver update from Motu's site. The problem has gone away. I don't know what this means for people with towers, but if you are using a firewire device the update does fix this problem. It is specifically for tiger, tiger must use the firewire bus differently than panther. I hope this helps.

[toodamnhip]

Doesn't seem to apply to my situation, I have a serial ata drive. Fire wire is going out my 828, but the firewire update did nothing to alleviate the spiking.

[Studio615]

Toodamnhip, have you tried to go back to 10.3.9? If it is tiger that is causing the problem, then this might be a workaround until they figure it out. I can't believe that you would get cpu spikes with that much power, something is seriously wrong with either tiger or DP 4.52. That is quite sad. You could run a space flight to the moon and back with the amount of raw power your machine has, and still should be able to record a few tracks at the same time. If I find out any more info I will post it. Good luck, I really hoped that this was a firewire driver issue, but apparently it is not.

<small>[ May 23, 2005, 10:01 AM: Message edited by: Studio615 ]</small>

[toodamnhip]

Thanks for your concern and time studio 615....

Going to 10.39 would be an incredible hassle.

Only my largest projects spike enough to cause audio drop outs.

So at least i can survive this stupid problem til a fix arrives..hopefully soon.

[Studio615]

Bump....... anyone? Bueller? Bueller? Bueller? MOTU WHERE ARE YOU? This problem is unacceptable and needs fixed now! Come on guys, help us out!

[Kaspar]

I think it might be a disk reading problem. Somehow the disk gets busy getting all the files ready. I can use 2 Altiverbs or my Powercore 'verbs; it doesn't matter. So I don't feel it's a processor problem. But the processor stresses because it don't get the needed files on time. This also explains why it is a problem on very fast computers. Of course, with the little knowledge we have on the details in OSX and DP we can only hope for fixes from Apple or MOTU. They've done it before. After a lot of OSX and DP updates it worked, but now it's broken again.

G4 533 DP, 1.25GB, OSX.4.1, DP 4.52, Powercore PCI, Waves 5, AltiVerb 5

[peteruno]

Don't know if Studio 610's PCMCIA firewire card fix really could have affected the spikes with 828mk2 by dispersing the load on the firewire bus. I tried the same thing (by putting a firewire card into my magma chasis) to try to unburden the firewire load from the 828 and it didn't work (tried 2 different cards, firewire only and firewire usb). I know from minimal success to get Mlan working with 828mk2 that even using my firewire 800 drives seems to tax the firewire 400 bus but that may be an MLAN hogging issue and the inability for the 828mk2 to live happily with low work priority.

[Studio615]

I hear what you are saying, but when I put the pcmcia card in, I only hooked the 828 to it. I left the hard disk hooked up to the built in firewire port. Correct me if i'm wrong, but I thought that by adding the card, I added a new firewire bus, free from the internal. Even if I didn't, and they are the same, maybe my problem was related to available bandwidth of the firewire cable, and splitting them up solved that problem. Regardless of the answer, I can now record without a dropout every 4-5 seconds. If you know more about the technical aspects of the fire bus please post on it, i would like to understand more. I am a wintel It guy who got fed up with wintel crap and fell in love with the ease of use from mac and motu. It is so nice to not have to yell at my recording machine all night, except for minor troubles like this. I can tell many horror stories of how trying to record on wintel with crapwalk just does not work. period. Thanks for the info. I would love to hear more. Thanks to everyone here for all of the help and knowledge. I have learned so much from everyone here!

[toodamnhip]

So Peter...

I take it that you too are having processor spiking?...

Originally posted by Peter Wetzler:
Don't know if Studio 610's PCMCIA firewire card fix really could have affected the spikes with 828mk2 by dispersing the load on the firewire bus. I tried the same thing (by putting a firewire card into my magma chasis) to try to unburden the firewire load from the 828 and it didn't work (tried 2 different cards, firewire only and firewire usb). I know from minimal success to get Mlan working with 828mk2 that even using my firewire 800 drives seems to tax the firewire 400 bus but that may be an MLAN hogging issue and the inability for the 828mk2 to live happily with low work priority.

[Studio615]

Guys, I just talked to a Motu tech today. He told me that they could not duplicate this problem in house. But, they told me a couple of things to try to resolve the problem. Try to turn off auto beat detection. It is a processor hog and could be the culprit, at least in my situation. Make sure that Tiger is a fresh, clean install. Don't upgrade. That one is a biggie. Something doesn't translate right when you upgrade. I happened to buy a new Powerbook with a panther disc in the box, it was shipped right before before Tiger came out. Apple sent me an upgrade disc. A friend is going to give me his full install disc to try. He also said that it could help if one of us sent them a project that this problem is happening in, it might help them address the problem. They seemed very eager to help, and I'm going to send them the project I am currently working on. I hope it helps, as the problem came back and I lost a couple of good takes last night because of it. I'll keep you guys posted.