KVR Audio

jumper24 · Post by **jumper24** » Sun May 31, 2015 8:33 pm

Hi

Just an unscientific benchmark test - Semi real world scenario
Doublet checked results.

On each track there is:
U-he Hive
Preset: ARK Blue Star (First preset ind the Pads and Strings cat)
Playing 5 vocies

Fabfilter Pro Q2 with 5 active band
Fabfilter Pro C default

Computer Dell xps 15 / Win8.1 / RME Babyface @256samples
Every DAW must play at least 30 sec without any dropout or crackle/pop sound.

Studio one 3.0.2
Nr of tracks: 24

Reaper 5.01
Nr of tracks: 31

Reaper 4.78
Nr of tracks: 31
Nr of tracks: 5 !!! (woot ?? Anticipative FX Processing off)

Cubase 8.0.20
Nr of tracks: 32
Nr of tracks: 24 (with ASIO Guard off)

Bitwig 1.1.7
Nr of tracks: 20
Nr of tracks: 17 (sandboxing)

Machine 2.3
Nr of tracks: 14

Ableton Live 8
Nr of tracke 25

Ableton Live 9
Nr of tracke 26

ericj23 · Post by **ericj23** » Sun May 31, 2015 10:22 pm

Try Cubase and Reaper without the pre rendering (ASIO Guard) - I suspect the answer will then be about 18. The maschine answer is the odd one for me.

5Lives · Post by **5Lives** » Mon Jun 01, 2015 12:05 am

I gave this a go as well on OS X on a top of the line Macbook Pro. I don't have Hive, so I used Serum's PD Defiant pad patch along with FB Pro-Q2 and Pro-C. 5 voices - had to loop at least once without dying and no pops / clicks.

Logic 10.1 - 24 tracks
Cubase 8 Pro - 13 tracks
Pro Tools 11 - 10 tracks
Studio One v3 - 7 tracks (it popped a little on 7, but I just gave it to it)

I had a suspicion that Serum's VST and AAX versions are for whatever reason not as CPU efficient as AU (it also crashed PT and Cubase), so I decided to try again with just Massive (custom preset, Ultra setting, 5 voices):
Pro Tools 11 - 48 tracks
Logic 10.1 - 45 tracks
Cubase 8 Pro - 44 tracks
Studio One v3 - 29 tracks (this was surprisingly high given my general experience with S1, S1v3 is a weird one though - sometimes it can play through once without pops, but then you loop it, and it goes to hell)

Anyway, Logic, PT 11, and Cubase 8 are all powerhouses, which was expected. S1 still has a ways to go.

Sparky77 · Post by **Sparky77** » Mon Jun 01, 2015 1:03 am

Perhaps things like features, sound and layout is a better real world test for a DAW.

foosnark · Post by **foosnark** » Mon Jun 01, 2015 1:04 am

ericj23 wrote:Try Cubase and Reaper without the pre rendering (ASIO Guard) - I suspect the answer will then be about 18. The maschine answer is the odd one for me.

I like Maschine, but it does have significantly higher CPU overhead than FL Studio (I haven't compared anything else). Particularly the 32-bit version. That said, in real-world usage I don't run into problems.

5Lives · Post by **5Lives** » Mon Jun 01, 2015 1:25 am

Sparky77 wrote:Perhaps things like features, sound and layout is a better real world test for a DAW.

They all have pretty similar features for the most part. CPU performance is as much a part of workflow as layout.

zvenx · Post by **zvenx** » Mon Jun 01, 2015 1:34 am

@ericj, what would be the point of purposely handicapping Cubase when they designed asio guard 2 for this specific purpose?

rsp

jumper24 · Post by **jumper24** » Mon Jun 01, 2015 3:56 am

The reason for the test:
I was mastering a project in Studio one 3, and with only 2 tracks and about 10 plugins (mostly fabfilter) there was click and pop´s all the time.
Also it was going crazy on its Cache-meter (not sure why.. will investigate.)
And normally this would not be a problem in cubase, are doing this all the time.

- I really like studio one 3 workflow wise (but like 5Lives said, CPU performance is as much a part of workflow as layout)
I stay in cubase for now - also Cubase is not bad - The biggest dislike for me is the dongle (as i´m on the move with laptop).

beatmangler443 · Post by **beatmangler443** » Mon Jun 01, 2015 4:08 am

jumper24 wrote:The reason for the test:
I was mastering a project in Studio one 3, and with only 2 tracks and about 10 plugins (mostly fabfilter) there was click and pop´s all the time.
Also it was going crazy on its Cache-meter (not sure why.. will investigate.)
And normally this would not be a problem in cubase, are doing this all the time.

- I really like studio one 3 workflow wise (but like 5Lives said, CPU performance is as much a part of workflow as layout)
I stay in cubase for now - also Cubase is not bad - The biggest dislike for me is the dongle (as i´m on the move with laptop).

This is studio one's first version. Steinberg had to do two updates to get it right. I"m sure in months the S1 performance will be better.

qtheerearranger · Post by **qtheerearranger** » Mon Jun 01, 2015 4:21 am

hmm...

my test on s1 3. 79 Tracks for me.

Screen Shot 2015-05-31 at 10.18.38 PM.png

Got a worse performance in bitwig 1.1.8 with 61 tracks compared to s1s 79

95 in logic pro, however, logic was doing some core bouncing when I got to 60+ tracks, as in, I had to stop and resume playback several times when adding tracks because logic would overload a core, then when id hit play again it would balance out the load.

I would rather have 79 tracks in s1 than 95 in logic pro any day tbh.

ericj23 · Post by **ericj23** » Mon Jun 01, 2015 7:02 am

zvenx wrote:@ericj, what would be the point of purposely handicapping Cubase when they designed asio guard 2 for this specific purpose?

rsp

Because you aren't really comparing the efficiency of the programs - remember ASIO guard, the process in reaper and logic add large latency buffers to pre render the tracks. It is not more efficient, it is just different. If for example you keep changing the selected track the CPU use is much higher as the program keeps having to change what is in real time.

The test does give an idea of how efficient hosts are, but to keep it fair you need to be comparing equal set ups. ASIO guard and the like tell you how efficiently Cubase etc render with massive latencies. I will bet the numbers from most other hosts are about the same at 1024 plus latencies.

But really all hosts have freeze functions (apart from Maschine) so this extra 'efficiency' is hardly the most important factor in choosing a DAW

Ogopogo · Post by **Ogopogo** » Mon Jun 01, 2015 11:35 am

Sparky77 wrote:Perhaps things like features, sound and layout is a better real world test for a DAW.

You can't test features and layout, those are matters of preference. Sound, well that's been gone over a million times, there is no evidence of any discernible difference except for things like pan laws or ableton leaving time-stretching on as default.

Performance isn't the end of be all of daw qualities but it does matter at times.

jumper24 · Post by **jumper24** » Mon Jun 01, 2015 5:27 pm

Updated with Anticipative FX Processing off (reaper), ASIO Guard off (cubase) numbers

zvenx · Post by **zvenx** » Mon Jun 01, 2015 5:29 pm

Interesting..... Over to the reaper crowd...

can we try the cubase with the computer off?

rsp, very happy Cubendo user...

I use 64 sample size buffers on both my PC Nuendo and Mac Cubase...Its the only settings for me that I don't feel the latency and I can play keyboards without having to quantize anything....
Yes everybody may have freeze but for me it is a workflow killer so I avoid it . (of course this depends on your projects..I am doing ads/jingles/scoring/music for infomercials etc and I am constantly with very short deadlines having to readjust stuff to satisfy the client, new changes in video etc)

Passing Bye · Post by **Passing Bye** » Mon Jun 01, 2015 5:33 pm

Language barrier I guess ooopsss

Daw benchmark test: Cubase / Bitwig / Studio one 3 / Reaper / Maschine