Page 1 of 2

Clip changing causes stuttering

Posted: Sun May 14, 2006 1:25 pm
by borg nagar
I have live 5.2 running quite large liveset with 7 audio channels, 2 midi channels (with impulses), 4 return channels with delays and one compressor. I have tried to keep the effects to absolute minimum with max 4 effects per channels (usually only EQ-4 or EQ-4 and a compressor). I use 4 bar quantisation, so that live should have time to prepare for the changing clips and I only launch max 2 clips at a time. The CPU usage is average 42-43% with this set playing all tracks and fx.

I'm experiencing stuttering when changing clips. The clips seem to throw the CPU usage to 200-400% and Disk light flashes. Particularly if I've worked with the set for a while. After one hour it usually starts to occur very regularly.

I re-installed Windows 5 days ago. I have updated my drivers for everything I can think of. I have defragmented my harddrives very regularly and defragment the paging file and directories on every reboot. I have followed every tip on http://www.musicxp.net/ website. And I have converted my filesystem to NTFS... to increase performance. It is still not enough. I have only the utmost important software installed.

The machine is MP4 3,2ghz Asus laptop with M-Audio audiophile FW sound interface, 512MB RAM. The sound card latency settings are set quite high: Buffer size 1024 samples, which gives Output&Input Latency: 25.2ms. Sample rate 44100, I even added 15ms of latency to get total of 65.4ms. I have disabled all inputs and outputs (except for one, which I play through)

Do you think my expectations from the machine are too high?

I have a gig very soon... Any tips?

Posted: Sun May 14, 2006 3:18 pm
by longjohns
i could be wrong, but i don't think that setting the global quantize to a higher value (4 bars, 8 bars) actually has any effect on how the audio engine is working. it doesn't give it more time to look ahead at its processing tasks. only your latency and buffer settings are controlling that.

what is the speed of your HD? do you have a second HD going? i'd recommend getting up to at least 1G ram. watch the heat on your computer. i read quite often about laptops with p4's being blazing hot

Posted: Mon May 15, 2006 11:01 am
by borg nagar
I think you are right about the quantization.

The HD is standard asus laptop HD. Slow, I know.

Yes, more RAM would do the trick (more space to leak the memory into).

But... right now I can make Live stutter by loading an empty liveset, then opening preferences and switching the Test Tone switch on and off.

There's something very wrong with something and I have 2 days to figure it out.

Posted: Mon May 15, 2006 3:08 pm
by longjohns
what kind of performance do you get with the on-board sound w/ asio4all?

Posted: Mon May 15, 2006 9:23 pm
by borg nagar
longjohns wrote:what kind of performance do you get with the on-board sound w/ asio4all?
I don't have the integrated soundcard driver installed, and I even removed it from device manager.

I think I got the problem sorted with assistance from some friends. They told me it's most likely an IRQ conflict or something similar (I even started having problems with the ASIO drivers at that time). I disabled parallel and IR ports from bios and removed all useless hardware from the device manager (card reader, modem, lan, wireless lan, integrated sound etc). And everything seems to run a lot smoother now, no audio dropouts have occurred after that.

I also deleted all the .asd files and let live reconstruct them (someone said here on the forum that it could help if I had been working on the same project with older versions of live). After those I reverted to earlier version of the M-Audio audiophile FW driver (5.10.0.5036), because the newer driver jammed the card's control panel software from time to time. I stopped all unnecessary windows services, only leaving 10 important ones running. I also disabled System file cache from Live's preferences. The overall latency is at 397ms now (HUGE), but I think I can live with it as long as I don't get any kind of dropouts or noisebursts.

Everything looks to be ok now.. but I'm wary. I'm sure some problems will pop up, if not earlier, atleast while playing the gig ;) But I hope for the best.

Posted: Mon May 15, 2006 11:36 pm
by longjohns
cool

the reason i asked is that for troubleshooting purposes it would be nice to switch out the M-audio device

Posted: Tue May 16, 2006 12:03 pm
by borg nagar
Yea. Thanks for your help! I really appreciate it.

The laptop really seems to work now, and the gig is tomorrow... close call! Now I have today to fine tune everything and perhaps compose some quick suprise for myself, hehe.

Posted: Thu May 18, 2006 12:43 pm
by borg nagar
It failed me :(

After about 35 minutes I got a noiseburst during the live performance. I rebooted, and got a bluescreen. I rebooted again, sound played fine for about 2 minutes, and the stuttering started during clip change... I had to stop playing.

Apalling.

I will be posting a clip from the sounds later.

Posted: Thu May 18, 2006 1:45 pm
by borg nagar
Here it is: noiseburst, recording from last night's live.

Beautiful! This is what I call ASIO noise. Even though I'm not sure what it actually is and where it comes from.

Posted: Thu May 18, 2006 8:07 pm
by borg nagar
longjohns wrote:what kind of performance do you get with the on-board sound w/ asio4all?
I tried it and stuttering occurs with the integrated soundcard and asio4all v2 driver as well.

I also installed a heat probe software, and tried to stress the HD with "CPU Burn-in" software. I couldn't get the heat higher than 71C even with the ventilation holes blocked. It stays on 61C when playing the liveset.

Posted: Fri May 19, 2006 12:59 am
by longjohns
shoot that really sucks that you had problems at your show.

i wouldn't exactly call those temps "cool", even if they are within spec. but i have no idea if that's the issue. just throwing stuff out there

my p4 2.8 has always given a temp of between 25-35 degrees, which i have always suspected is total BS, because my other comp is a p4 1.8 which idles at like 45deg. but anyways

sorry man

Posted: Fri May 19, 2006 1:39 am
by Attack mode
It is most likely a virtual IRQ issue . The only way to resolve it would be to reinstall Windows and diable ACPI on the install so you can manually change the IRQ assignments.


goto run>msinfo32>IRQ If you see that your IEEE is above 15, that would be a strong case for your firewire interface not playing along. My lovely laptop has my ATI graphics card virtual, while having a dedicatd IRQ for my touch pad and SM media bay...good thinking!

The only way to chage IRQ it is in the BIOS ( might be able to go it on an Asus), or you will have to reformat and diable ACPI on the install so you can manually change IRQ assigments if need be.

Posted: Fri May 19, 2006 2:44 am
by longjohns
you can switch to standard PC mode without reinstalling, but i do not think that you can switch back to ACPI. AFAIK, even if you manually set IRQ info in the bios, as long as XP is in ACPI mode, it will not pay attention to any of it, and share irq's as it sees fit

Posted: Fri May 19, 2006 7:14 am
by Amaury
Hi Borg,

What does the Task manager reads when Live is playing your set ? In the RAM usage section.
That could sound like a memory swaping issue too, like Windows has to swap memory to the hard drive and the hard drive can't cope with that.

Best,

Amaury

Posted: Fri May 19, 2006 10:40 pm
by Attack mode
longjohns wrote:you can switch to standard PC mode without reinstalling, but i do not think that you can switch back to ACPI. AFAIK, even if you manually set IRQ info in the bios, as long as XP is in ACPI mode, it will not pay attention to any of it, and share irq's as it sees fit

I would like to know how to do this... as I haven't been able to find this function you speak of, so that I can reasign my IRQ's

Cheers,

-N