That may well be. Sample buffer sizes unequal a power of 2 (48/96/192...) should be avoided anyway. Use 64 or 128 and it works correctly, right?

Hi Matthias & Dr Tone

Sadly, Dr Tone is correct.

It needs to be pushed all the way up to  1024  to ensure  crackle / pop  free use with a FFUC  -  whereas  Sonar X1 / Reaper / NTrack / Pro Tools 9  etc......  are all perfeclty happy at 48 Samples with huge VST Synths loads !?!?!?

I hadnt realized that this was a " known '  issue to Presonus.

Their excuse is  VERY  dissappointing  given that their own USB Interfaces  [ AudioBox ]  have  Buffer settings down to  32  Samples  [  I can only assume that they also use a substantial " safety "  buffer to ensure it is glitch-free at 32 Samples  ].

Ben

I have the same issue with the demo Studio One v1 : see this thread.

Hi Timur

If you have a FireFace UC you results / test  at  48 Sample Buffer with multiple VST Synths would be hugely appreciated.

Also, please see Post #5  here:-  http://forums.presonus.com/posts/list/0/16423.page

Its looking like this a  Version 2  " bug " .

Look forward to your testing.

Many thanks,
Ben

It's been a problem right from the beginning with Version 1.  I've just given up on it, I was getting no where with them.  I use direct monitoring and keep my buffers up.  It's sad because Studio 1 is one awesome DAW.

Hi Matthias

Just as a total long-shot,  I dont suppose the new  1.013 UC Drivers dated  09/11  fix the Studio One 48 Sample Buffer VST Synth crackling  issue ?

Ben

Currently I am testing a single plugin (both internal and external) in Studio One v1, will go on to v2 later. Curiously I noticed that I never got the license for v1, not even an invoice, but did the payment. So I'm running on trial versions now.

I am able to make the UC go down to 64 samples with occasional cracks on my setup, it seems susceptible to CPU power-savings though (C1E, dynamic clock-rate changes). 64 samples on the UC roughly corresponds to the same round-trip latency as a FF400 being set to 48 samples, so 48 samples on the UC is *lower* in round-trip latency than a 400 (lower output, but higher input latency to be exact).

Next thing I did was to compare CPU load behavior when running Studio One idle (nothing loaded, playback stopped) via the UC vs. 400 at different buffer sizes. I measured CPU C0 (non idle) states, which reveal much more than what Task-Manager or SO's CPU load meter tell us. For that I turned off multi-threading and forced SO's CPU core affinity to only use CPU0.

Interestingly the FF400 shows the same load behavior for buffer sizes 128-1024 samples and then increases load *slightly* with smaller buffers. The UC shows the same load as the 400 at 128 samples (around 20% C0)! Load *decreases* slightly with bigger buffer sizes (aka better than the 400), but increases *dramatically* with smaller sizes (400: 24% vs. UC: 41%). The latter should be the source of the issues you are experiencing.

Whether that increase if caused by the UC driver, Studio One, or the Windows USB driver (over which the UC driver communicates) I do not know. When I find time I will compare the behavior to Reaper and maybe some other applications. I tried a NI Kore USB audio interface for comparison, but setting that to its lowest buffer of 32 samples + 1 ms USB buffer corresponds more to setting the UC to 128 samples for output latency and 64 samples for input (64-96 on the 400).

To make things even more complex, even with Studio One being forced to a single core (and using my own customized power-profiles) there still is load on multiple cores present. One reason being that there are many more processes running in the background, another reason being how Windows balances load between Studio One, the audio driver and the USB driver. To check for that I also looked at the whole CPU package C0% state and generally found the above tendencies to be correct.

Hola,
I have also been testing the S 1 demo on my laptop
Acer travelmate 5720 Win XP Tested systems Studio one will not produce crackle free sound under 256 samples buffer.Cubase and reaper work fine with 64 samples buffer (moderate load) on the same cmp.
According to this post problem w/Studio One Pro and RME babyface Studio One works fine with other audio interfaces. HeadScratch

Cheers

I'm always curious about these things and will get in contact with Presonus. What I found out so far is that with USB based RME drivers the corresponding Studio One ASIO thread uses considerably more "context switches" (aka more taxing to the CPU even if you don't see it as classic CPU load). This would be the source for the higher C0 state percentage

I just Installed "Studio One 2 Pro x 64" - my interface is a UFX via USB.
My computer is a ADK Quad Xtreme,
Win 7-Pro x 64
i7 930 2.8GHz Quad Core - overclocked 3.4-3.6GHz
12Gig DDR3 1600 CL8 memory.
I haven't done any recording yet but playing back the Demo songs included with Studio One I could get a clear audio playback on two of the three included Demos at 512 samples but one of the songs had to be played at 1024 samples to prevent the popcorn and cracker-jack sounds.

Yes, we are already in contact. Please give us both some time to work this out.

That is great news.  I just wish they would have contacted you guys a year ago when I first pointed out the issue to them.

This was posted on the Presonus Forum by a user of the Presonus 1818 VSL. By looking at his reply concerning the test. His New Presonus USB device did no better then RME! This issue effects Fireface 800 - Firewire as well as USB! I have the same issues with my Fireface 800 in S1V2 Pro. Thanks Ben for bringing some of these issues to my attention!

Thanks,
Randy

benifin wrote:
Hi gospeltunes

1818 VSL is a very nice unit as is the FireFace UC.

Could you please perform the following 2 very quick tests with your 1818VSL - it will only take you a minute or so for each test:-

TEST 1 - 32 SAMPLE BUFFER

- Set your 1818VSL to play and record at 88.2k / 32 samples
- Open a new clean empty song in S1V2 and set it to 88.2k / 24 Bit
- Load any VST soft-synth - Ampire / Guitar Rig / Mojito / anything etc..
- engage the monitor / record on your track
- now play something with you midi controller and also record it

TEST 2 - 48 SAMPLE BUFFER

- Set your 1818VSL to play and record at 88.2k / 48 samples
- Open a new clean empty song in S1V2 and set it to 88.2k / 24 Bit
- Load any VST soft-synth - Ampire / Guitar Rig / Mojito / anything etc..
- engage the monitor / record on your track
- now play something with you midi controller and also record it

For *both* tests:-

=> do you hear any pops / crackling when you are playing the midi controller ?

=> do you hear any pops / crackling when you play back the track you have recorded ?

Your testing information would be quite crucial in determinig if this problem *ONLY* relates to RME units *OR* if it affects all USB units at ultra-low-latencies.

Many thanks in advance.

Ben

The results of Test 1:

I followed instructions, loaded a grand piano from Komplete 7, and I got nothing but strange high pitch noises during monitoring, recording, and playback.

The results of Test 2:

The 1818 won't let you set the ASIO buffer size to 48, it only list 32, 64, 128, 256, 5120, 1024, and 2048 as possible choices.

Hi Matthias !

I know its only been  7  days  but was wondering / hoping  that there may have been some progress on this issue between RME and Presonus ( ? )

Only asking because until its resolved, using my RME FireFace UC with Studio One V2 at  sub-1024 sample latencies is still crackly - plus at 1024, because I do a lot of VST synth and Guitar rig work, its just way too  " laggy " .

Either way, many thanks for begining to address this issue.

All the best,
Ben

We had sent them a Fireface UC, Timur some tons of measurements - so it all worked out!