|
|
|
05-20-2009, 10:00 PM
|
#1
|
Mortal
Join Date: Jan 2006
Location: Wickenburg, Arizona
Posts: 14,051
|
Soundcards and Latency
Since this comes up so much I thought I'd make a sticky and hopefully that will help for buying cards and such
If you want to help, download the Latency Test Utility from CEntrance
http://www.centrance.com/downloads/ltu
Remember to turn off any monitoring so as not to skew the results. Please also check to see if the value changes every time you hit test. State Stable/Unstable according to whether it changes or not
Here's the list, Ill add to it as it grows:
LTU results(in msec) @44.1khz
32/64/128/256 sample buffer size
**NA/12.56/15.46/25.54 - Stable Echo Audiofire 12
**NA/ 7.35/13.15/24.76 - Stable Egosys Wami Rack PCI
**NA/***NA/18.44/27.14 - Unstable Line6 KB37 *
**NA/ 3.06/***NA/***NA - Stable Maudio Delta 44 PCI
**NA/ 3.20/ 6.10/11.88 - Stable Maudio Delta TDIF PCI
**NA/ 5.06/ 7.91/13.72 - stable within .05 msec M-Audio Fast Track Ultra USB
2.04/ 3.85/ 6.76/12.54 - Stable MOTU 896 (measuring digital I/O)
3.79/ 5.24/ 8.14/13.95 - Stable MOTU 896 (measuring analog I/O)
2.49ms/ 3.95ms/ 8.46ms/ 14.26 ms - Stable MOTU traveler driver rev 3.6.7.4 (measuring analog I/O)
3.61/ 5.06/ 7.96/13.76 - Stable RME Multiface
N/A / 6.10/ 9.00/14.81 - Stable RME FF800
* The line6 drivers are acting very screwy and reporting thruput without anything plugged in. In 16 bit mode the measurment appears accurate. In 24 bit mode, there is some internal passthru fudging the numbers
|
|
|
05-20-2009, 10:42 PM
|
#2
|
Human being with feelings
Join Date: Jul 2006
Location: Cowtown
Posts: 1,562
|
M-Audio Fast Track Ultra: NA / 5.06 / 7.91 / 13.72
Scott
|
|
|
05-20-2009, 10:58 PM
|
#3
|
Human being with feelings
Join Date: Jul 2006
Location: Hungary
Posts: 3,129
|
RME FF800, analog
(Gigabyte EP45-DQ6 MoBo FW)
@ 32 : N/A
@ 64 : 6.10 (269 samples)
@128 : 9.00 (397 samples)
@256 : 14.81 (653 samples)
exactly the same results everytime i press "Measure!"
__________________
panda in the desert
Last edited by beatbybit; 05-21-2009 at 04:07 PM.
|
|
|
05-20-2009, 11:15 PM
|
#4
|
Administrator
Join Date: Jan 2005
Location: NYC
Posts: 15,746
|
MOTU traveler, driver rev 3.6.7.4
32: 110 samples / 2.49ms
64: 174 samples / 3.95ms
128: 373 samples / 8.46ms
256: 629 samples / 14.26 ms
|
|
|
05-21-2009, 06:03 AM
|
#5
|
Human being with feelings
Join Date: Mar 2007
Location: Mediterrenean Sea
Posts: 979
|
RME HDSP9632 PCI
@44.1KHz
32 (156 samples): 3.54 ms
64 (220 samples): 4.99 ms
128 (348 samples): 7.89 ms
256 (604 samples): 13.70 ms
[EDIT]
Always exact same results.
use 64 for recording and 512 for mixing. Usually audio only projects larger than 50 chanels, 24 tracks playin at once. CPU raises to 80-90% when full of fx. Very reliable.
P4 4GigRam
winxp sp3 nlited by dux
Reaper v2.58
[/EDIT]
Last edited by antiClick; 06-11-2009 at 01:24 AM.
|
|
|
05-21-2009, 03:29 PM
|
#6
|
Human being with feelings
Join Date: Jan 2008
Location: UK
Posts: 139
|
Line6 POD Studio UX1
32 : NA
64 : NA
128: 159 samples / 3.61 ms
256: 281 samples / 6.37 ms
512: 543 samples / 12.31 ms
512: 1061 samples / 24.06 ms
Rather unstable Results keep changing eveytime i click on Measure.
|
|
|
05-21-2009, 03:39 PM
|
#7
|
Human being with feelings
Join Date: Nov 2007
Location: Denver, CO
Posts: 153
|
Focusrite Saffire ASIO Control Panel Says:
ASIO Buffer Size: 2.0ms
Firewire Buffer Size: 1.0ms
Input Latency: 3.6ms
Output Latency: 5.5ms (this is what is reported by the soundcard)
Centrance says:
88 samples @ 44.1k
Measurement results: 475 samples / 10.77 ms
192 samples @96k:
Measurement results: 924 samples / 9.63 ms
All values are stable.
I realize that this isn't done in 32/64/128/254 samples, but I don't see an option to change the buffer based on size, only on time... (and I tried to make that work to no avail.)
Hardware: Dell 630 laptop w/4G RAM, SIIG Firewire card w/TI Chipset running winxpsp3
Like Syd says below, I also run at 2.0ms/1.0ms settings for very responsive recording. I've never noticed latency recording to it, nor has anyone I've recorded with. It's not as fancy as my old motu 24IO, but since I don't do this for a living anymore, this is completely acceptable.
Last edited by themensch; 02-01-2010 at 11:23 AM.
Reason: include hardware/os specs
|
|
|
05-25-2009, 09:51 PM
|
#8
|
Banned
Join Date: Sep 2008
Location: gently scraping your delicate sensibilities with a potato peeler
Posts: 6,791
|
Saffire Pro 26
Driver version 2.3.8467
Dux WinXP32 SP3
Abit IP35 Pro on-board TI Firewire
Q6600 4gb
@44.1k
32 samples = 5.87ms
66 samples = 8.16ms
132 samples = 11.66ms
264 samples = 17.64ms
all measurements were stable under Centrance LTU
The first two however have not been tested under practical applications, ie- jamming.
I've currently been using ASIO 88 sample (2.0ms) buffer with fw 1.0ms buffer:
88 samples = 9.66ms
This setting is very responsive with guitar/bass/vsti so 9.66 seems a little misleading especially when compared to some of the measurements posted here.
I'd be curious to know more details from everyone regarding OS (XP32/64/ vs Vista32/64) and firewire PCI vs on-board. As always with these things, the more data the better.
Last edited by Syd Masters; 05-28-2009 at 10:35 PM.
Reason: driver shit
|
|
|
05-26-2009, 10:37 PM
|
#9
|
Human being with feelings
Join Date: Jun 2006
Location: Finland, Kuopio
Posts: 911
|
M-Audio FW1814
32 samples N/A
Every result was unstable. What does that mean? Result did not change every time, sometimes 5-6 times the same.
Last edited by Finnish; 05-26-2009 at 10:39 PM.
|
|
|
05-27-2009, 03:27 PM
|
#10
|
Human being with feelings
Join Date: Apr 2008
Posts: 2,036
|
M-Audio Delta 44
M-Audio Delta 44
44.1K
buffers: 64 195 samples / 4.42 ms
buffers: 128 323 samples / 7.32 ms
buffers: 256 579 samples / 13.13 ms
buffers: 384 835 samples / 18.93 ms
buffers: 512 1091 samples / 24.74 ms
(all stable readings)
Last edited by LCipher; 05-27-2009 at 03:33 PM.
|
|
|
05-29-2009, 10:25 AM
|
#11
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
on my echo audiofire12, the results change every time as well. what should i be looking for to correct this?
|
|
|
05-29-2009, 02:58 PM
|
#12
|
Mortal
Join Date: Jan 2006
Location: Wickenburg, Arizona
Posts: 14,051
|
Quote:
Originally Posted by d. gauss
on my echo audiofire12, the results change every time as well. what should i be looking for to correct this?
|
I'd like to know as well
|
|
|
05-29-2009, 09:41 PM
|
#13
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
pipe, are you having the same issues?
Quote:
Originally Posted by pipelineaudio
I'd like to know as well
|
|
|
|
05-29-2009, 10:01 PM
|
#14
|
Mortal
Join Date: Jan 2006
Location: Wickenburg, Arizona
Posts: 14,051
|
On certain setups I am. What Im not sure about yet is whether the same card goes from stable to unstable on different pc's
|
|
|
05-30-2009, 06:14 AM
|
#15
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
i can't figure it out. at buffer of 256 it varies from 23 to 25 ms with every test --1018 to 1160 samples. i've gone through and shut down every device and service until the computer was totally crippled and still it has no effect.
even tried on 3 different computers all with similar results.
Last edited by d. gauss; 05-30-2009 at 04:36 PM.
|
|
|
06-02-2009, 12:44 AM
|
#16
|
Human being with feelings
Join Date: Jun 2006
Location: Finland, Kuopio
Posts: 911
|
Quote:
Originally Posted by d. gauss
i can't figure it out. at buffer of 256 it varies from 23 to 25 ms with every test --1018 to 1160 samples. i've gone through and shut down every device and service until the computer was totally crippled and still it has no effect.
even tried on 3 different computers all with similar results.
|
Is this somehow interface-related? Some are giving "stable" results and others don't, but they still work "as expected"?
|
|
|
06-02-2009, 05:29 AM
|
#17
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
exactly that's what i'm asking, as i have no idea what causes this.
|
|
|
06-02-2009, 08:54 AM
|
#18
|
Human being with feelings
Join Date: Apr 2007
Location: Germany
Posts: 100
|
ST-Audio DSP2000 C-Port
ok, here the values from my ST-Audio DSP2000 C-Port (bought in 2003 !!)
http://www.st-audio.de/products/dsp2000.html
Driver: dsp24ewdm 8.4.1021
LTU results(in msec) @44.1khz all stable
48: 2.20 ms
64: 2:95 ms
128: 5.85 ms
256: 11.63 ms
Win XP SP3 Dux's version
Intel(R) Core(TM)2 Duo CPU E4400 @ 2.00GHz
Gigabyte GA-P35-DS3
4GB Kingston 2G-UDIMM
cheers, hb
__________________
Intel Core 2 Duo E4400 @ 2 GHz, 4 GB Ram
ST Audio DSP2000 C-Port w/V 8.4
|
|
|
06-02-2009, 10:12 AM
|
#19
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
Quote:
Originally Posted by hannebambel
LTU results(in msec) @44.1khz all stable
48: 2.20 ms
64: 2:95 ms
128: 5.85 ms
256: 11.63 ms
|
just curious, by stable, do you mean you get EXACTLY 11.63ms @256 EVERY single time you do the test?
|
|
|
06-02-2009, 10:53 AM
|
#20
|
Human being with feelings
Join Date: Sep 2008
Location: Sweden
Posts: 7,432
|
Almost ashamed of the following, given the results of others, but what the heck...
Lexicon Lambda with LambdaASIO driver
@32: not possible, couldn't set it that low
@64: not applicable, CEntrance couldn't detect its own pulse (I guess)
@128: 17.5 ms (rather unstable but mainly around that)
@256: 35 ms (rather unstable but mainly around that)
Lexicon Lambda with ASIO4ALL 2.9
@32: not possible
@64: not applicable
@128: 17.53 ms (much more stable than LamdaASIO)
@256: 23.33 ms (much more stable than LambdaASIO)
Dell Vostro 1510, Intel Core 2 DUO T5870 2GHz, 3GB 667MHz DDR2, 250GB WD Scorpio Black 7200SATA
DuX' audio optimised XPsp3
__________________
// MVHMF
I never always did the right thing, but all I did wasn't wrong...
|
|
|
06-02-2009, 11:14 AM
|
#21
|
Human being with feelings
Join Date: Apr 2007
Location: Germany
Posts: 100
|
Quote:
Originally Posted by d. gauss
just curious, by stable, do you mean you get EXACTLY 11.63ms @256 EVERY single time you do the test?
|
yeah, i measured it about 30 times in a row, with random breaks in between and allways get 11,63 ms @ 256 samples. wanna tell me why you are curious ?
cheers,
hb.
__________________
Intel Core 2 Duo E4400 @ 2 GHz, 4 GB Ram
ST Audio DSP2000 C-Port w/V 8.4
|
|
|
06-02-2009, 12:58 PM
|
#22
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
Quote:
Originally Posted by hannebambel
yeah, i measured it about 30 times in a row, with random breaks in between and allways get 11,63 ms @ 256 samples. wanna tell me why you are curious ?
|
thanks. just trying to get to the bottom of why my results waiver a ms or so every time. i.e. close, not always the same.
still can't figure it out.
|
|
|
06-02-2009, 03:09 PM
|
#23
|
Human being with feelings
Join Date: Dec 2008
Location: US
Posts: 255
|
Not to be a bother but i don't see how this is helpful. Doesn't latency change on everyone's system and even at different sample rates?
Heck I've even gotten varying latencies at the same sample rate. Is this just to give people a rough idea of the latency that you might get with that soundcard?
|
|
|
06-02-2009, 03:14 PM
|
#24
|
Mortal
Join Date: Jan 2006
Location: Wickenburg, Arizona
Posts: 14,051
|
It should be pretty consistent between pc's
It changes per sample rate but again, should be pretty consistent when measuring the same rate
|
|
|
06-02-2009, 03:15 PM
|
#25
|
Human being with feelings
Join Date: Jul 2006
Posts: 20,859
|
Quote:
Originally Posted by GuitarKeyes
Not to be a bother but i don't see how this is helpful. Doesn't latency change on everyone's system and even at different sample rates?
Heck I've even gotten varying latencies at the same sample rate. Is this just to give people a rough idea of the latency that you might get with that soundcard?
|
not only the latency but the stability of the latency and consistency of the same interface in different systems. i'm wondering how daw's (reaper included) handle latency compensation for unstable interfaces.
|
|
|
06-02-2009, 04:16 PM
|
#26
|
Human being with feelings
Join Date: Dec 2008
Location: US
Posts: 255
|
Quote:
Originally Posted by brainwreck
not only the latency but the stability of the latency and consistency of the same interface in different systems. i'm wondering how daw's (reaper included) handle latency compensation for unstable interfaces.
|
I'm guessing that the asio drivers would have to keep reporting the new latency to the host. Not sure if the asio reported latency remains fixed or changes. To me if this is not the way asio drivers work currently then I think adding this would be a good way to make sure that the correct latency is properly compensated for even in unstable interfaces. Anybody in the know wanna chime in?
|
|
|
06-02-2009, 04:23 PM
|
#27
|
Human being with feelings
Join Date: Jul 2007
Posts: 1,891
|
laptops versus desktops, latency will differ.
Quote:
Originally Posted by pipelineaudio
Since this comes up so much I thought I'd make a sticky and hopefully that will help for buying cards and such
|
it is perhaps also good to know that latency and other issues will differ depending on whether you are using a laptop or not. (esp. with pc's) >>>laptops use acpi to check the battery and that service causes dropouts as well as added latency. also, having wifi enabled adds latency. i have found that users' desktops almost always coast in comparison to what their laptops do.
if you are using a laptop, run your latency checking program (such as dcplat.exe) and go into your devices and disable acpi battery method and study the difference.
this may not work for everyone, but give it a try if you haven't already.
peace.
|
|
|
06-02-2009, 04:43 PM
|
#28
|
Super Moderator (no feelings)
Join Date: Dec 2007
Location: On or near a dike
Posts: 9,836
|
You are maybe confusing DPC latency with audio latency. There is only a faint connection between these two as in "raising ASIO buffers may or may not help avoiding dropouts due to DPC related usurpation of the CPU".
Quote:
Originally Posted by reapercurious
>>>laptops use acpi to check the battery and that service causes dropouts as well as added latency.
|
Not laptops in general. Some models suffer from this, others don't. It's also not always the ACPI battery services causing DPC problems in laptops.
Quote:
Originally Posted by reapercurious
also, having wifi enabled adds latency.
|
These problems are not confined to Laptops, some desktop PCs have problems with DPC latency as well. This is a rather complex issue and generalizations are not possible since the culprit can be a lot of things from faulty hardware to bad BIOS implementations and of course drivers.
|
|
|
06-02-2009, 05:17 PM
|
#29
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
this is what i get:
|
|
|
06-02-2009, 10:32 PM
|
#30
|
Human being with feelings
Join Date: Apr 2007
Location: Germany
Posts: 100
|
d.gaus,
on my tests the sample size is allways the same. maybe there is a driver issue with your device since your sample size seems to be oddly high varying from 1066 to 1114 @ 256 samples where on my system it reports constantly 513 samples. In all the tests at any sample size and sample rate the measured sample size is allways close to twice the tested sample size where yours is allmost four times the sample size ... how are your dcp latency results ?
cheers,
hb
__________________
Intel Core 2 Duo E4400 @ 2 GHz, 4 GB Ram
ST Audio DSP2000 C-Port w/V 8.4
|
|
|
06-03-2009, 07:37 AM
|
#31
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
Quote:
Originally Posted by hannebambel
on my tests the sample size is allways the same. maybe there is a driver issue with your device since your sample size seems to be oddly high varying from 1066 to 1114 @ 256 samples where on my system it reports constantly 513 samples.
|
yeah, this isn't good. i hooked up a mackie onyx firewire mixer and the results were rock solid (see below). echo hasn't been much help yet. and all this time, i thought the audiofire was stable/reliable?
anybody else have an audiofire 8 or 12 that they've tested?
|
|
|
06-03-2009, 09:51 AM
|
#32
|
Human being with feelings
Join Date: May 2006
Posts: 1,631
|
update -- just got this from echo:
<<Hi - Our Q.A department has confirmed the problem and looking for a solution...>>
fingers crossed.
|
|
|
06-04-2009, 12:56 AM
|
#33
|
Human being with feelings
Join Date: Nov 2008
Posts: 18
|
Asio Roland Sonic Cell
Buffer size/Latency: 96 sample (2.18ms)
Sample Rate: 44100Hz
Measure:
640 sample / 14.51 ms
641 sample / 14.54 ms
638 sample / 14.47 ms
639 sample / 14.49 ms
Buffer size/Latency: 192 sample (4.35ms)
Sample Rate: 44100Hz
Measure:
832 sample / 18.87 ms
833 sample / 18.89 ms
834 sample / 18.91 ms
831 sample / 18.84 ms
Result: Unstable
|
|
|
06-04-2009, 01:21 AM
|
#34
|
Human being with feelings
Join Date: Sep 2008
Location: Sweden
Posts: 7,432
|
Quote:
Originally Posted by pro77
Result: Unstable
|
Well, it varies between 0.02267 and 0.02268 ms/sample.
Is that not stable enough?
__________________
// MVHMF
I never always did the right thing, but all I did wasn't wrong...
|
|
|
06-04-2009, 01:51 AM
|
#35
|
Human being with feelings
Join Date: Nov 2008
Posts: 18
|
Quote:
Originally Posted by Fabian
Well, it varies between 0.02267 and 0.02268 ms/sample.
Is that not stable enough?
|
Yes, it is.
How did you get these numbers 0.02267/8?
I'm sorry, I'm not so technical
|
|
|
06-04-2009, 03:48 AM
|
#36
|
Human being with feelings
Join Date: Sep 2008
Location: Sweden
Posts: 7,432
|
Quote:
Originally Posted by pro77
Yes, it is.
How did you get these numbers 0.02267/8?
I'm sorry, I'm not so technical
|
That's ms/sample.
I just divided 14.51/640 = 0.02267 ms/sample, etc
EDIT:
BTW, those numbers (ms/sample) multiplied by the sample rate (sample/sec) should give 1000 (ms/sec), and they come close enough, 0.02267 * 44100 = 999,75, and 0.02268 * 44100 = 1000.19
__________________
// MVHMF
I never always did the right thing, but all I did wasn't wrong...
Last edited by Fabian; 06-04-2009 at 04:45 AM.
|
|
|
06-04-2009, 03:57 AM
|
#37
|
Human being with feelings
Join Date: Nov 2008
Posts: 18
|
Quote:
Originally Posted by Fabian
That's ms/sample.
I just divided 14.51/640 = 0.02267 ms/sample, etc
|
Ok, thank you Fabian
|
|
|
06-04-2009, 04:52 AM
|
#38
|
Human being with feelings
Join Date: Nov 2008
Posts: 35
|
Results on Line 6 Toneport Ux 2 with 3.4.3.6 Driver:
@44.100 hz,32 bit
stable 1024/23.22 ms
stable 512/11.61 ms
stable 256/5.8 ms
stable 128/2.90 ms
But in Reaper I get different latencies, for example at 256 samples I Have 17ms of output latency... What are the right values?
Last edited by Highlander; 06-04-2009 at 04:55 AM.
|
|
|
06-04-2009, 08:24 AM
|
#39
|
Human being with feelings
Join Date: Jun 2006
Location: Finland, Kuopio
Posts: 911
|
Quote:
Originally Posted by Highlander
Results on Line 6 Toneport Ux 2 with 3.4.3.6 Driver:
@44.100 hz,32 bit
stable 1024/23.22 ms
stable 512/11.61 ms
stable 256/5.8 ms
stable 128/2.90 ms
|
Are you absolute sure you got those values as final values? Seems quite low...
|
|
|
06-04-2009, 01:40 PM
|
#40
|
Human being with feelings
Join Date: Nov 2008
Posts: 35
|
Quote:
Originally Posted by Finnish
Are you absolute sure you got those values as final values? Seems quite low...
|
yes, these are the values given by the program. In reaper the latencies reported are much larger, maybe these values are much realistic.
|
|
|
Thread Tools |
|
Display Modes |
Linear Mode
|
Posting Rules
|
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
HTML code is Off
|
|
|
All times are GMT -7. The time now is 05:14 AM.
|