If kick-ass low latency performance is your #1 priority and all other things being equal, then I think you got it about right (Reaper > Studio One > Cubase). But there are a lot more considerations into why to chose a DAW.reggie1979 wrote: ↑Mon Apr 22, 2019 5:00 pm I'm really confused by all of this. The detail and minutia throw me. All I want to know is how/can I get my work done.
I used reaper for years and never had any issues. Studio one seems to be quite nice these days. Reason is so-so. It works, I can get the job done. Cubase seemed to be the hungriest. Always had to be careful of the CPU meter.
Studio One 4.1.4 - V - Reason 10.3 - |Performance Compared|
-
Funkybot's Evil Twin Funkybot's Evil Twin https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=116627
- KVRAF
- 11483 posts since 16 Aug, 2006
-
- KVRian
- 694 posts since 9 May, 2005
I'm not speaking to the OP's test.
With my workflow, I would never work at a 512-sample ASIO buffer size.
When pushing the limits of ultra low latency, there's a significant performance difference between various DAW applications.
That difference in performance means being able to effectively work... vs. having to compromise (raise the ASIO buffer size).
-
Funkybot's Evil Twin Funkybot's Evil Twin https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=116627
- KVRAF
- 11483 posts since 16 Aug, 2006
Jim, out of curiosity, what's your mixing process? Do you mix with a lot of instrument/effects running in real-time? Are you freezing virtual instruments (or not running any at all)? I keep my buffer at 128 (RME Fireface 800 - looking to upgrade soon) as I find it's the best compromise between latency and CPU usage, but come mix time, I usually need to increase it. I should get into the habit of creating stems, but I have a bad tendency of just leaving my instrument tracks non-rendered in case I need to start tweaking the arrangement in the mix.Jim Roseberry wrote: ↑Mon Apr 22, 2019 5:46 pm When pushing the limits of ultra low latency, there's a significant performance difference between various DAW applications.
That difference in performance means being able to effectively work... vs. having to compromise (raise the ASIO buffer size).
-
- KVRian
- 694 posts since 9 May, 2005
I'm running a pretty hefty machine (as that's what I do for a living).
I'm using a fair amount of realtime processing (mainly EQ/Dynamics and some Reverb and Delay)
I'll usually freeze any virtual instruments (prior to mixing).
I'm typically doing a rock project (not huge orchestral mock-ups like many of our clients)... so the projects generally aren't super dense.
I've got both Presonus Quantum and RME Fireface UFX+ audio interfaces.
Quantum feeds the UFX+ via lightpipe... so I can use the Quantum if I want/need crazy low round-trip latency.
All monitoring is done via the UFX+.
If I'm using hardware instruments, I'll use the UFX+ as audio interface.
If I'm using software instruments, I'll use the Quantum as audio interface.
I may pickup a SSL Six and use that for routing/monitoring (instead of the UFX+).
Quantum isn't the end-all of audio interfaces, but when it comes to round-trip latency, it's exceptional.
As good as the UFX+ is... I just don't want to take that small step backward.
For tracking, I'm running at a 32-sample or 64-sample ASIO buffer size.
For mixing, I rarely go above a 128-sample ASIO buffer size.
I'm using a fair amount of realtime processing (mainly EQ/Dynamics and some Reverb and Delay)
I'll usually freeze any virtual instruments (prior to mixing).
I'm typically doing a rock project (not huge orchestral mock-ups like many of our clients)... so the projects generally aren't super dense.
I've got both Presonus Quantum and RME Fireface UFX+ audio interfaces.
Quantum feeds the UFX+ via lightpipe... so I can use the Quantum if I want/need crazy low round-trip latency.
All monitoring is done via the UFX+.
If I'm using hardware instruments, I'll use the UFX+ as audio interface.
If I'm using software instruments, I'll use the Quantum as audio interface.
I may pickup a SSL Six and use that for routing/monitoring (instead of the UFX+).
Quantum isn't the end-all of audio interfaces, but when it comes to round-trip latency, it's exceptional.
As good as the UFX+ is... I just don't want to take that small step backward.
For tracking, I'm running at a 32-sample or 64-sample ASIO buffer size.
For mixing, I rarely go above a 128-sample ASIO buffer size.
-
- KVRAF
- Topic Starter
- 3230 posts since 30 Dec, 2014
I've added the test files to my first opening post so you can compare the results for yourself, obviously you'll need these DAWs installed with UHE's Hive, but I did add the midi file which I had initially exported from Reason and imported into Studio One. With that, you can use the test file with other DAWs.
Here's the additional settings you didn't see in the video.
Click to view larger
Here's the additional settings you didn't see in the video.
Click to view larger
KVR S1-Thread | The Intrancersonic-Design Source > Program Resource | Studio One Resource | Music Gallery | 2D / 3D Sci-fi Art | GUI Projects | Animations | Photography | Film Docs | 80's Cartoons | Games | Music Hardware |
- KVRAF
- 2338 posts since 28 Feb, 2015
Does it matter though? Even if Studio One is "cheating", you obviously get more power out of it than Reason? A Koenigsegg is faster than a Volvo, or should the Koenigsegg only be allowed to drive on the first gear for the test to be valid?
i9-10900K | 128GB DDR4 | RTX 3090 | Arturia AudioFuse/KeyLab mkII/SparkLE | PreSonus ATOM/ATOM SQ | Studio One | Reason | Bitwig Studio | Reaper | Renoise | FL Studio | ~900 VSTs | 300+ REs
-
- KVRian
- 992 posts since 27 Apr, 2005
You can get power out of reason by setting it to whatever large number of samples the dropout protection is setting it to in the dual buffer in S1. So, unless you know what the buffer size is on all the tracks, the test is pointless.
It’s all a vanity bs test anyway. “My toy is bigger than your toy”. To allude to your koenigsegg/volvo comparison, the koenigsegg won’t get to the liquor store any faster than the Volvo in normal traffic, and you’ll be more uncomfortable
It’s all a vanity bs test anyway. “My toy is bigger than your toy”. To allude to your koenigsegg/volvo comparison, the koenigsegg won’t get to the liquor store any faster than the Volvo in normal traffic, and you’ll be more uncomfortable
-
- KVRAF
- 1858 posts since 26 Nov, 2018
(plays twilight zone theme)
- KVRAF
- 2338 posts since 28 Feb, 2015
Yeah, I see what you mean. But a fair game would be if both DAWs are showing the same latency, right? And from the count how many plugins they can run without audio dropouts.
i9-10900K | 128GB DDR4 | RTX 3090 | Arturia AudioFuse/KeyLab mkII/SparkLE | PreSonus ATOM/ATOM SQ | Studio One | Reason | Bitwig Studio | Reaper | Renoise | FL Studio | ~900 VSTs | 300+ REs
-
- KVRAF
- 1858 posts since 26 Nov, 2018
No, it must be a carefully played out game where there is much intrigue and guile.
-
- KVRian
- 992 posts since 27 Apr, 2005
S1 has a dual buffer, and reason doesn’t. If Props implemented the same thing in reason, it would severely screw with the cv patching system. But if you are measuring a plugin run count, all S1 is doing is automatically increasing the buffer size until it can play without dropouts. You can set the same buffers in Reason Manually, and run the same number of plugs (or at least similar, I have no doubt S1 is a bit more efficient) but we have no idea what S1 is setting the buffer to, to make a comparison.starflakeprj wrote: ↑Mon Apr 22, 2019 11:55 pm Yeah, I see what you mean. But a fair game would be if both DAWs are showing the same latency, right? And from the count how many plugins they can run without audio dropouts.
-
- KVRAF
- 1858 posts since 26 Nov, 2018
And then there you go, makin' sense and all that. For shame.
-
- KVRAF
- Topic Starter
- 3230 posts since 30 Dec, 2014
I re-ran Reason 10.3 and increased the buffer to maximum (2048) and managed to get a little more instances, up to 53, but of course that comes at a cost of higher latency whilst playing. Increasing the buffer in the same way in Studio One 4.1.4 whilst leaving everything else as it was before, provides no benefit as clicks start to occur as more instances are added. But it shows that S1 4.1.4 performance is still well over double that of Reason on the same hardware.
On a practical level, if someone walks into a music store to buy DAW, and they say they want one that will give them the best performance for a new rig they have built...it's important to tell them them what one's do, because you want them to come back and buy more stuff... ok, that's less common as we have internet shopping these days...but still, for anyone starting out, it's useful information to have... of course there could be still people plodding along on a DAW for over 10 years...that just haven't looked over the fence to know about it...until now.
On a practical level, if someone walks into a music store to buy DAW, and they say they want one that will give them the best performance for a new rig they have built...it's important to tell them them what one's do, because you want them to come back and buy more stuff... ok, that's less common as we have internet shopping these days...but still, for anyone starting out, it's useful information to have... of course there could be still people plodding along on a DAW for over 10 years...that just haven't looked over the fence to know about it...until now.
KVR S1-Thread | The Intrancersonic-Design Source > Program Resource | Studio One Resource | Music Gallery | 2D / 3D Sci-fi Art | GUI Projects | Animations | Photography | Film Docs | 80's Cartoons | Games | Music Hardware |
-
- KVRian
- 992 posts since 27 Apr, 2005
To be clear then, you still have not run the test with the dynamic second buffer of undetermined size disabled, The Block size in S1's prefs is only for the "live" buffer for tracking.
As well, you have Reason's Hyperthreading turned off, and S1's Dropout protection also makes extensive use of processor threading to do it's Voodoo... (Now to be fair the Reason Hyperthreading support is probably crap, and I have it turned off as well. haven't run it since the update)
On a practical level, knowing how many concurrent instances of a plugin can be run in a sterile environment is not what making art is about. Penis size contests about DAW specs are irrelevant. Reaper would probably win all those anyway even though I profoundly dislike that app. There are quite a few scenarios where I would heartily recommend S1 (Mixing, Tracking Bands, etc...) I like it, and have it installed on my machine. My mouse still passes right by the S1 icon on the way to the Reason icon on a daily basis tho, even before the 10.3 update when performance was abysmal.
As well, you have Reason's Hyperthreading turned off, and S1's Dropout protection also makes extensive use of processor threading to do it's Voodoo... (Now to be fair the Reason Hyperthreading support is probably crap, and I have it turned off as well. haven't run it since the update)
On a practical level, knowing how many concurrent instances of a plugin can be run in a sterile environment is not what making art is about. Penis size contests about DAW specs are irrelevant. Reaper would probably win all those anyway even though I profoundly dislike that app. There are quite a few scenarios where I would heartily recommend S1 (Mixing, Tracking Bands, etc...) I like it, and have it installed on my machine. My mouse still passes right by the S1 icon on the way to the Reason icon on a daily basis tho, even before the 10.3 update when performance was abysmal.