Different maximum number of plugins depending on the DAW

[Hochaufloesend]

First this: my point is not to start DAW bashing, but to find out if this is a problem specific to my setup or if other people have observed this as well.
I'm posting on the U-He forum because I use mostly U-He plugins and some of them are not known to be frugal with the available processing power. But I can reproduce the problem also with Omnisphere.

I recently upgraded to Ableton Live 11 and started a small project with it (12 tracks, 2 x Zebra, 1 x Diva, 1 x Ableton Drumrack with Simplers, the rest are audio tracks or external synths).

The project was already kicking out hard on the Ableton CPU load meter at a buffer size of 128 samples, but I didn't think anything of it. Only when I started to add UAD plugins (which are calculated on a dedicated DSP card!) and then audio crackles appeared, I became suspicious.

I migrated the project to Reaper 6 and lo and behold: everything runs without a crackle and the CPU load hardly ever rises above 10% - and this with a buffer size of 48 instead of 128 samples!

I then put together a few tests and got the following results:

Plugins_in_Reaper_vs_Live.png

For this experiment, I fed a polyphonic MIDI loop to a single track/instance in the project and duplicated the track until audio artifacts became audible. I then noted the number of tracks/instances where it still ran cleanly.

Below a few details about my studio computer on which I performed the above tests:
- OS: Windows 10 Pro / Version 10.0.19042 Build 19042
- CPU: Intel(R) Core(TM) i7-7700K CPU @ 4.20GHz, 4200 Mhz, 4 Core(s), 8 Logical Processor(s)
- RAM: 16Gb RAM
- Graphics: NVIDIA GeForce GTX 1050 Ti / 4Gb
- Audio IF: RME Fireface 400 & Fireface 800 (combined 46 channels In & Out)
- All Synths used were VST2 / 64bit
- Windows power plan: High performance


Can anyone reproduce this behavior with a similar setup? Is this also reproducible on Apple machines? Why does Reaper handle so many more Divas/Repros than Live? Am I doing something wrong? Should I rather use VST3?

[Funkybot's Evil Twin]

Got a third DAW you can test? I mean, if DAW #3 operated more like Reaper that might point to Live 11 being the issue. Heck, do you still even have Live 10 to compare to? Add another DAW in to see what happens. Could just be Live 11 is new and under-cooked and maybe a little buggy from a performance perspective with some plugins. Also, what about other manufacturers plugins?

Reaper is probably the most efficient DAW (at least that I've used), but it's not THAT BIG of a difference.

[pdxindy]

IIRC, Reaper is automatically using a higher buffer for tracks that are not recording live. That is why there is such a big difference at 48 and not so much at 1024.

[MitchK1989]

IIRC, Reaper is automatically using a higher buffer for tracks that are not recording live. That is why there is such a big difference at 48 and not so much at 1024.

Wow that’s actually really clever of them. I assume there are reasons that wouldn’t work in a more real time performance oriented DAW like live but I’m not computer savvy enough to say what they would be

[Hochaufloesend]

Got a third DAW you can test? I mean, if DAW #3 operated more like Reaper that might point to Live 11 being the issue. Heck, do you still even have Live 10 to compare to? Add another DAW in to see what happens. Could just be Live 11 is new and under-cooked and maybe a little buggy from a performance perspective with some plugins.

Good point. Tanks.
Did a quick and dirty test today with Mixbus 6 (DIVA, HS Giant Sloth): Seems even worse here than with Live 11 (barely 1 instance (MC On) with 48 samples and 4 instances with 1024 samples).

Unfortunately, I have already thrown away Live 10. But it's possible that version 11 still needs some fine tuning. Let's see what the coming updates bring.

Also, what about other manufacturers plugins?

Yes, as mentioned above, I also tested this with other Plugins.
With Omnisphere, I got about twice the number of instances out with Reaper.

Reaper is probably the most efficient DAW (at least that I've used), but it's not THAT BIG of a difference.

Yes, it seems the Reaper people squeeze the most out of the hardware.
It is also the only DAW know to me that manages to drive the CPU load up to 100% according to Windows Task Manager without the audio output suffering (unlike the GUI response time).

For me, it makes a big difference, because I don't want to buy a new PC.

[Hochaufloesend]

IIRC, Reaper is automatically using a higher buffer for tracks that are not recording live. That is why there is such a big difference at 48 and not so much at 1024.

That would explain a lot.
Have already wondered why the different buffer sizes have hardly affected the results.

Reaper probably also calculates the tracks in advance, which do not require live monitoring.