PlanetZ forums - scopeusers.com

I'm just wondering if anybody can get 1.5 ms?

The best usable latency that I can get is 5.8 using the latest drivers (Scope 4, Sonar 5 or Samlitude). This is on 2 different machines, one a dual Xeon 2.66 Hyperthreaded machine and the other a dual core 3.0 machine. I've tried every thing that i can to get the latency down to that of my M-Audio FW1814, but to no avail.

So I'm just wondering if anybody doing audio recording get anything better that 5.8?

<font size=-1>[ This Message was edited by: 7XL on 2005-10-15 18:57 ]</font>

at what samplerate?
at 44.1k, i always work at 4ms. it will go lower....

I often work at 48k and use the lowest setting in the ULLI, which in cubase gives me a total latency of 4ms

Messiah

44.1K.

Are you doing audio recording and if so what is your track count?

i can use well over 36 tracks. i record as many as 16 at a time. sonar's performance is poor compared to logic 5.51 and sx3, but that's not surprising as cakewalk truly dislikes using a steinberg app like asio. still, 16 tracks recorded at once works fine. i only use wav drivers with sam and rarely more than two tracks as i use it for mastering, mainly.

I have recorded 36 tracks at once with the Dual Xeon setup, but i still can't get the latency below 5.8 ms.

And the thing about Sonar having poor performance, I have no idea where you got that from. I'm not going to get into a "war" over recording apps. But I only use ASIO drivers and with all the other audio devices that I have, except my Ozone and US428, I can get 1.5 ms. I am more interested in trying to improve the performance of that I can get from these cards.

Correct me if I'm wrong, but the Tascam is 4in/2out and the Ozone is 4in/4out at 44.1Khz.

I run an older Dual Xeon here (Prestonias on i860) and have seen 4ms latency in Scope with no adverse side effects or cpu load. 4ms latency is the same buffersize that other cards refer to as 3ms, and I never saw need to go below this for what I do. Most likely the only thing holding you back there is your system configuration--its entirely possible that the dual cpu support in sonar is causing some i/o bottlenecking on the main system bus. I used to have a lot of problems with SX/Nuendo in 'SMP' mode and finally realized that logic 5.51 (with only 1 cpu supported) yields me better performance than Steinberg with all the bells & whistles enabled, and is a lot more stable to boot.

Also, 1.5ms latency is a bit excessive even with my RME unless you're tracking a VERY sensitive performer through software monitoring (and scope is SO much better than software monitoring). I typically record at 3-6ms and mix at 6-12ms (23 if needed).

On 2005-10-15 23:48, garyb wrote:
sonar's performance is poor compared to logic 5.51 and sx3, but that's not surprising as cakewalk truly dislikes using a steinberg app like asio.

Sorry to get a bit OT here, but that's an interesting point...
Anyone else experienced the ASIO support of Sonar in comparison to other apps? Do the late versions of Sonar, which DO support native ASIO still left behind?
Or can the audio engine of late Sonar now be compared to the Logic/Cubase audio performance?

i don't know about the latest version. i have sonar3 which isn't bad at all, just not as stable as logic pc or sx3. later versiions may be much improved. sonar3 is certainly sufficient, though. don't get the idea that i'm claiming otherwise...

I use ASIO with everything that I run. I do not see what the problem is with the apps that I'm running. I can set the card to 7 ms ion the ULLI settings (actual latency 5.8 ms), anything below that I occasional ticks, pops, etc.

As far as my system configuration goes, I have no problem with other audio devices that I've tried, the only one that gives me any "problems" is the Scope system. I can set the Scope to 4 ms (actual 2.9 ms), but I find that the 3 ms is completly unusable. Even at 4 ms setting I find that I have issues when I start adding native effects or the track count gets "high".

*shrugs*
i don't see a real problem. there are a lot of variables involved. i have a demo computer at work that doesn't perform as well as the machine in my studio. i need to reinstall the os properly on the machine at work, i swapped a motherboard without doing a reinstall. it works pretty well, just not as well as the machine in the studio. sonar seems to be the app that suffers on the store machine the most. 4ms will click on the sonar demo at the end of the song consistantly, but i pay it no mind as i know that the os is at fault.

on my studio computer, i do get 4ms without ANY problems(and i don't care what it sez in the sequencer), which is useful enough. i monitor in scope so it's really only an issue when playing vstis. if 4ms(8ms both ways) disturbs me or an artist, i monitor a scope or an external synth or drum machine while recording midi. i have full control over the sound afterward. usually 8ms of latency doesn't get in the way of creative flow anyway. the perceived latency in the scope environment is 0. for mixdowns i can raise the latency to 23ms and never notice.

obviously, i don't understand the problem. sorry to hear you have trouble. if you are just looking at the numbers and think that fastest is automatically the best, then by all means, use the fw device if it's "faster". i doubt the fw interface will do as much for you as a production tool, and i doubt it'll make your music any closer to getting a grammy. don't be mad, use what you like. personally, i'd take a pulsar1 at 13ms over ANY of the fw interfaces on the market, but that's me. you can tell by the number of posts that i've made that i like scope. as for sonar, i LIKE sonar, it's great. it should work great for you, i would think.


<font size=-1>[ This Message was edited by: garyb on 2005-10-16 15:41 ]</font>

valis,

You are correct, the 2 USB devices that I mentioned are not capable of doing more than 4 I/O.

I have been working with a vocalist that likes to hear what is "printed" as she records. So achieving the lowest possible latency is important. I like using the Vinco on her voice as it adds a nice warmth to the overall signal chain (GT67, A-Designs MP-2, Vinco) without getting too "mushy". Anything above what she precieves to be "tape latency" (1.5ms) is not acceptable to her. I've used hardware monitoring with her using all hardware, but she didn't like the sound of the overall signal chain (Maybe I should look for a vintage 1176).

garyb,

I'll take yor Grammy insult with a grain of salt as you have no idea of who you are typing at. I have 3 Scope cards and 1 Pulsar II that I've used for the last 3 years with no problems. I mentioned the FW1814 because I find it "odd" that I can achieve better results with my laptop that I do everything on than I can on both of my dedicated studio machines.

boy, you really ARE testy. no insult intended. i don't find what you report strange.

if you monitor off the scope mixer when tracking(not through the sequencer), then you can use vinco or whatever scope processing with no more latency than the hardware version.

Agreed. Routing through scope introduces negligable latency (+/-~1ms for the i/o buffers and only a few samples per device) and is the same as using hardware. Only by routing the signal into AND back out of sonar before sending to her headphone/monitor would the ASIO latency make any difference.

Its well established here that scope is rather finicky when it comes to hardware resource sharing. There could be any number of things you could track down to help 'tweak' your system. The answer to your question seems to be that yes people are achieving lower latencies with creamware cards, are you interested in help or...?

valis,

I am interested in help, that is why I asked if anyone was getting lower latency with acceptable results than I was achieving. I generally don't have these issues, but this is a "special" case and I am seeking solutions. Thanks for your time and any insight that you can give in helping me to solve this "problem".

I hear you loud and clear Garyb. 11/01/05 you and me, mostly you.

In a number of tests routing Nuendo 1.6 tracks out to SFP to outboard gear and back into SFP into Nuendo 1.6, I got latencies between 80 samples and 100 samples. What's that at 44.1 kHz? Ummm...

*uses digits to calculate*

1.814 ms to 2.27 ms.

That's in AND out, with ASIO.

Running on a 1.6 GHz P4 with external (Lucid) wordclock, and a minimal, non-upgraded Windoze XP (no service packs), if any of that makes any diff.

I've never done the test with VSTi's, no.

the best latency that i achieve was 7ms but its ok with me coz im still on 13ms. wish someday i'll buy a machine like grayb and rock'n roll to 4ms. 7xl your vocalist must be one hell of a musician. recognizing 2ms is nightmare. maybe you can move few samples after each vocal take to solve the problem.
all my best