xp daw, periodic stuttering
zikester
annoying half-second freeze/stutter. Interestingly, I took the time
down of each of these stutters, and of the 7 I measured, they all fall
on the 36 second mark, and generally either 2 or 3 minutes apart. So
something periodic must be going on.
Details:
- dell 700m, 1 Gb RAM, 5400 RPM 120Gb hard drive
- interface: presonus firebox via firewire
- softsynth: NI B4 II ( b3 emulator, CPU load under 20% whole time ),
Electrik Piano 1.5 ( stutters on both, only 5-7% load )
- did the a bunch of common tweaks listed on internet forums, like
emphasizing background processes over applications, no virtual memory,
disabled unneeded services, turned off wireless and unneeded devices.
- controller: oxygen 8 / novation 49 SL MkII ( stutters on both )
I did a fresh install of Windows XP home SP2, installed only the NI
Komplete suite and basic MIDI tools, and feel like I've stripped
everything down as much as I can. What could be going on? Any known
problems with the dell 700m firewire or something?
Thanks
Soundhaspriority
"zikester" <isaa...@gmail.com> wrote in message
news:de96f69e-7146-4964...@z35g2000prh.googlegroups.com...
Zikester,
It sound as if your softsynth has a data cache, and the cache isn't big
enough to hold everything it needs. The problem is the 1GB of RAM; not big
enough for a complex soft-synth. When it hits the hard disk for missing
data, you get the stutter. You may be able to test this hypothesis if the
softsynth has an adjustable quality setting. Setting the quality level to
minimum reduces the size of the samples, increasing the period of the
stutter, or possibly eliminating it.
In my experience with Steinberg's "The Grand", 2.5 GB of RAM is
required with Windows XP to have a cache large enough that hitting the disk
is not necessary. This is an unusually complex plugin; yours should not
require as much, but 2GB is advisable.
Bob Morein
(310) 237-6511
zikester
3 different soft synths. The other clue that says this isn't the
problem:
the stutter happens at the 36second mark, meaning I took the literal
wall-clock time down at each stutter, and the second hand showed "36"
on each of the stutters. This could not happen if it were a caching
issue and I was freely improvising. I'd think that the mark wouldn't
be so consistently on the 36second mark. More likely to be some
periodic service or something in the background...but what?
On Dec 13, 3:44 pm, "Soundhaspriority" <nowh...@nowhere.com> wrote:
> "zikester" <isaac...@gmail.com> wrote in message
Soundhaspriority
and is hitting the disk almost continuously, then a service that preempts
disk access could cause this.
Another test might be possible, if you can control where the synth pulls its
data from. If it doesn't provide that kind of control, plug in a usb disk,
and try installing it to that.
DPC Latency Checker may be a useful diagnostic:
http://www.thesycon.de/deu/latency_check.shtml
Bob Morein
(310) 237-6511
"zikester" <isaa...@gmail.com> wrote in message
news:2ee0a75e-b139-490e...@z35g2000prh.googlegroups.com...
geoff
> I don't think this is the problem--I experienced the same problem with
> 3 different soft synths. The other clue that says this isn't the
> problem:
>
> the stutter happens at the 36second mark, meaning I took the literal
> wall-clock time down at each stutter, and the second hand showed "36"
> on each of the stutters. This could not happen if it were a caching
> issue and I was freely improvising. I'd think that the mark wouldn't
> be so consistently on the 36second mark. More likely to be some
> periodic service or something in the background...but what?
>
Soundcard buffer size ?
geoff
zikester
> It could still be a caching problem. If your synth has a very small cache,
> and is hitting the disk almost continuously, then a service that preempts
> disk access could cause this.
>
> Another test might be possible, if you can control where the synth pulls its
> data from. If it doesn't provide that kind of control, plug in a usb disk,
> and try installing it to that.
>
> DPC Latency Checker may be a useful diagnostic:http://www.thesycon.de/deu/latency_check.shtml
>
> Bob Morein
> (310) 237-6511
>
Thanks--more info:
- DPC latency checker shows the 12-13ms spikes of latency in the DPC
queue that coincide with the stuttering.
- I tried disabling every conceivable device that is not critical /
the presonus driver itself.
- the spikes show up *even when I'm not playing anything*, so we know
it's not softsynth related in all probability
Now I'm at a loss as to what to try next!
Thanks for your help so far...
zikester
And, btw, my machine has 1.3 Gb RAM. TaskMgr shows > 925Mb available
at all times.
Soundhaspriority
"zikester" <isaa...@gmail.com> wrote in message
news:13a17e92-9470-44ff...@y32g2000prd.googlegroups.com...
If this is a laptop, go into system-->devices, and look for the APIC
compliant battery device. This device can be freely enabled and disabled --
the only loss of functionality is the battery meter. On my machine, it's not
even necessary to reboot. The problem disabling it sometimes solves is that
the driver for the battery measuring device takes control, in what is known
as a "spinlock", halting the machine for a brief interval. But I am
surprised that the Presonus would be so susceptible to a rather small
latency. I have heard bad things about their drivers.
Bob Morein
(310) 237-6511
zikester
It appears to be solved! I disabled the display driver and even
removed the battery, and one of those two seems to have fixed it....10
minutes straight with no dpc spikes, even while playing softsynths.
So, those of you with a dell 700m for DAW use, might try disabling the
dispaly driver ( and using standard vga ), and/or not using the
battery.
Thanks for all of you guys's help...
Soundhaspriority
--------------------------------------------------
You don't have to remove the battery - just disable the battery driver. The
battery will remain functiona, you just can't monitor the charge state.
Bob Morein
(310) 237-6511
Arny Krueger
Not enough RAM for a current version of XP.
Jay Ts
> "zikester" <isaa...@gmail.com> wrote
>> - DPC latency checker shows the 12-13ms spikes of latency in the DPC
>> queue that coincide with the stuttering. - I tried disabling every
>> conceivable device that is not critical / the presonus driver itself.
>> - the spikes show up *even when I'm not playing anything*, so we know
>> it's not softsynth related in all probability
>>
>> Now I'm at a loss as to what to try next!
>>
>> Thanks for your help so far...
>
> If this is a laptop, go into system-->devices,
That's Control Panel -> System -> Hardware -> Device Manager ->
Batteries -> Microsoft ACPI-Compliant Control Method Battery
(on my Toshiba laptop, running Windows XP, at least.)
Then right-click and select "Disable". A red X appears on the icon there.
> But I am surprised that the Presonus would be so
> susceptible to a rather small latency. I have heard bad things about
> their drivers.
It's not Presonus's fault. 12-13 ms of system latency here is HUGE.
It's more than enough to cause a problem with the audio driver's
buffers, which nowadays we expect to be well under 5 ms.
The delay of 12+ ms. in the device driver causes the application
to be put off for at least that much time. Even if the audio device
driver is able to fill the buffers, the application cannot process
the data, resulting in an overflow of the buffer, and clicks/pops
as the audio system (incl. application) resynchronizes.
Jay Ts
--
To contact me, use this web page:
http://www.jayts.com/contact.php
geoff
> Thanks--more info:
> - DPC latency checker shows the 12-13ms spikes of latency in the DPC
> queue that coincide with the stuttering.
> - I tried disabling every conceivable device that is not critical /
> the presonus driver itself.
> - the spikes show up *even when I'm not playing anything*, so we know
> it's not softsynth related in all probability
>
> Now I'm at a loss as to what to try next!
>
> Thanks for your help so far...
MIDI or soundcard on PCI ? Try moving the lot (shared interupt ?).
geoff
geoff
Video adaptor that uses shared system RAM ?
geoff
Badmuts
The 'current' version of XP (i assume you mean XP SP3) does not use much
more memory than the first release of XP.
Extra things they added like windows defender or the firewall may take up
some memory, but that's still not spectacular.
Any XP system used for light work should run fine with 512MB RAM and 1024 is
luxury.
For Vista and Windows 7 it's a different story, those like at least 2048 MB
Arny Krueger
message news:4b26c8a6$0$16661$2e0e...@news.tweakdsl.nl
>> Not enough RAM for a current version of XP.
> The 'current' version of XP (i assume you mean XP SP3)
> does not use much more memory than the first release of
> XP.
I wish. XP memory usage for common functions has increased significantly
with every service pack.
Every once in a while I build up a bootable XP hard drive starting with a XP
SP0 install disk, and take it through the service packs.
Back in the days of SP0, a 512 K machine was useful. As of SP3 2 GB is about
the point of diminishing returns unless you're running some really
RAM-intensive software such as video editing. Then there is never too much
RAM, and 64 bit addressing and > 4 GB is your friend, even if you are
running 32 bit applications.
> Extra things they added like windows defender or the
> firewall may take up some memory, but that's still not
> spectacular.
The extra weight seems to have gone in at a lower level, probably in the
name of better security and integrity.
> Any XP system used for light work should run fine with
> 512MB RAM and 1024 is luxury.
I wish. With XP luxury starts at 2 GB.
I'm adding RAM to XP systems all the time, both desktops and laptops. Right
now my target is 2 GB, and 4 GB despite the obvious waste of RAM in the 4th
GB, is not a bad idea for most power users.
> For Vista and Windows 7 it's a different story, those
> like at least 2048 MB
I'm sure there is enough RAM in the world to make Vista run adequately, but
Windows 7 was a big improvement.
Peter Larsen
> While playing a softsynth using a midi controller, I repeatedly get an
> annoying half-second freeze/stutter. Interestingly, I took the time
> down of each of these stutters, and of the 7 I measured, they all fall
> on the 36 second mark, and generally either 2 or 3 minutes apart. So
> something periodic must be going on.
I'm not really certain about this, but your laptop appears to be a "small
un'" and those generally utilize shared ram. Do you really really need
portability? - imo it is too costly if you need a reasonably but not too
capable graphics subsystem and a multispindle disk subsystem, and those you
DO need for a high performance box.
Bob Moreins suggestion:
> DPC Latency Checker may be a useful diagnostic:
> http://www.thesycon.de/deu/latency_check.shtml
Is a good one.
Mine is to use a stationary computer with a separate graphics card,
preferably _no_ _more_ than 256 megabytes ram on it, minimum 2 gigabyte ram
and multiple harddisks so that you can avoid concurrent reads and writes.
The simplest ram setup for 3.2 or so gigabyte available ram in 32 bit
windows may be to put two 2 gigabyte sticks in and live with a bit of it
being unused. XP tends to become surprisingly faster with 2 or more
gigabytes ram compared to having ""only"" one.
Using rdp or radmin to operate a stationary computer from a laptop can work
very well and gives you the best of both worlds.
Kind regards
Peter Larsen