Does Bounce in Place free CPU?

This forum is for discussing Reason. Questions, answers, ideas, and opinions... all apply.
Post Reply
dustmoses
Posts: 197
Joined: 04 Oct 2015

11 Jun 2016

I'm wondering if bounce in place actually frees CPU.

In reason 8 it seems you have to either delete or completely unwire a device to do this.

User avatar
Benedict
Competition Winner
Posts: 2747
Joined: 16 Jan 2015
Location: Gold Coast, Australia
Contact:

12 Jun 2016

So long as a device isn't processing audio it should drop most of its consumption. Usually this is achievable by Muting the Track so no Midi triggers any notes. Un-linking/removing should really only be a total last resort.

I don't know about the new R9 Bounce In Place but the existing Bouncing a Track will auto Mute the original - if you check the box.

:)
Benedict Roff-Marsh
Completely burned and gone

User avatar
Gaja
Posts: 1001
Joined: 16 Jan 2015
Location: Germany
Contact:

12 Jun 2016

Well bounce in place is no freeze function alright. Your synt won't be triggered and therefore use less CPU, unless of course it has LFOs running and controlling some filter in another synth ir whatever. It still goes on and uses CPU accordingly, which is of course why freeze is a tad difficult to implement.
Cheers!
Fredhoven

User avatar
Marco Raaphorst
Posts: 2504
Joined: 22 Jan 2015
Location: The Hague, The Netherlands
Contact:

12 Jun 2016

I did a quick test using a Thor which I ran through many effect devices, including a few Rack Extensions.

I did the test on my MacBook Pro running without the power adapter plugged it. ( I don't think this will make much of a difference when running Reason btw, but I just wanted to make this clear).

Using the OSX Monitor, here are the result.

Thor + effects:
thor plus effects.png
thor plus effects.png (1.81 MiB) Viewed 2562 times
Bounce in Place:
bounce in place.png
bounce in place.png (1.83 MiB) Viewed 2562 times
Bounce without Thor + effects (the original track and rack devices):
bounce without original track.png
bounce without original track.png (1.21 MiB) Viewed 2562 times
In my opinion Bounce in Place is doing its job perfectly. There's just a tiny bit of overhead compared to the version without the original track. Thor has free running LFO's so it will always take a tiny bit of processing power. And think about it: the track is not muted, Thor is plugged in and able to produce sounds any time I want it.

Bounce in Place is a fantastic feature. It's better than the Freeze functionality a program like Ableton uses, because it creates a fully functional audio file NEXT to the original/source chain of devices.

WongoTheSane
Moderator
Posts: 1851
Joined: 14 Sep 2015
Location: Paris, France

12 Jun 2016

Marco Raaphorst wrote:Using the OSX Monitor, here are the result.
Very interesting, thanks! How come the monitor reports a 171% CPU usage in the first picture? (on Windows, the total always comes up to 100%, so no individual process can go above that).

User avatar
Marco Raaphorst
Posts: 2504
Joined: 22 Jan 2015
Location: The Hague, The Netherlands
Contact:

12 Jun 2016

WongoTheSane wrote:
Marco Raaphorst wrote:Using the OSX Monitor, here are the result.
Very interesting, thanks! How come the monitor reports a 171% CPU usage in the first picture? (on Windows, the total always comes up to 100%, so no individual process can go above that).
Because my MacBook Pro uses multiple CPUs.

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

12 Jun 2016

Benedict wrote:So long as a device isn't processing audio it should drop most of its consumption. Usually this is achievable by Muting the Track so no Midi triggers any notes. Un-linking/removing should really only be a total last resort.

I don't know about the new R9 Bounce In Place but the existing Bouncing a Track will auto Mute the original - if you check the box.

:)
Bounce Mixer Channels mutes the channel but not the track. To save CPU you need to mute the track. Bounce in place will mute the track, but has fewer options. I still prefer Bounce Mixer Channels, you just have to remember to mute each track after bouncing.


Sent from my iPhone using Tapatalk
Selig Audio, LLC

User avatar
Marco Raaphorst
Posts: 2504
Joined: 22 Jan 2015
Location: The Hague, The Netherlands
Contact:

12 Jun 2016

selig wrote:
Benedict wrote:So long as a device isn't processing audio it should drop most of its consumption. Usually this is achievable by Muting the Track so no Midi triggers any notes. Un-linking/removing should really only be a total last resort.

I don't know about the new R9 Bounce In Place but the existing Bouncing a Track will auto Mute the original - if you check the box.

:)
Bounce Mixer Channels mutes the channel but not the track. To save CPU you need to mute the track. Bounce in place will mute the track, but has fewer options. I still prefer Bounce Mixer Channels, you just have to remember to mute each track after bouncing.


Sent from my iPhone using Tapatalk
And you need to imported all Bounced Mixer Channels if you're using that proces, which is a bit of extra work right? I use Bounce Mixer Channels for archiving, but never thought about using it for keeping my CPUs happy. Will experiment with it probably soon. Thanks! :)

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

12 Jun 2016

Marco Raaphorst wrote:
selig wrote:
Benedict wrote:So long as a device isn't processing audio it should drop most of its consumption. Usually this is achievable by Muting the Track so no Midi triggers any notes. Un-linking/removing should really only be a total last resort.

I don't know about the new R9 Bounce In Place but the existing Bouncing a Track will auto Mute the original - if you check the box.

:)
Bounce Mixer Channels mutes the channel but not the track. To save CPU you need to mute the track. Bounce in place will mute the track, but has fewer options. I still prefer Bounce Mixer Channels, you just have to remember to mute each track after bouncing.


Sent from my iPhone using Tapatalk
And you need to imported all Bounced Mixer Channels if you're using that proces, which is a bit of extra work right? I use Bounce Mixer Channels for archiving, but never thought about using it for keeping my CPUs happy. Will experiment with it probably soon. Thanks! :)
There is an option to bounce to new mixer channels. So it's all one step.



Sent from my iPhone using Tapatalk
Selig Audio, LLC

User avatar
Marco Raaphorst
Posts: 2504
Joined: 22 Jan 2015
Location: The Hague, The Netherlands
Contact:

12 Jun 2016

selig wrote:
Marco Raaphorst wrote:
selig wrote:
Benedict wrote:So long as a device isn't processing audio it should drop most of its consumption. Usually this is achievable by Muting the Track so no Midi triggers any notes. Un-linking/removing should really only be a total last resort.

I don't know about the new R9 Bounce In Place but the existing Bouncing a Track will auto Mute the original - if you check the box.

:)
Bounce Mixer Channels mutes the channel but not the track. To save CPU you need to mute the track. Bounce in place will mute the track, but has fewer options. I still prefer Bounce Mixer Channels, you just have to remember to mute each track after bouncing.


Sent from my iPhone using Tapatalk
And you need to imported all Bounced Mixer Channels if you're using that proces, which is a bit of extra work right? I use Bounce Mixer Channels for archiving, but never thought about using it for keeping my CPUs happy. Will experiment with it probably soon. Thanks! :)
There is an option to bounce to new mixer channels. So it's all one step.



Sent from my iPhone using Tapatalk
Ok thanks!

WongoTheSane
Moderator
Posts: 1851
Joined: 14 Sep 2015
Location: Paris, France

12 Jun 2016

So I did the same test on Windows 7. Not the same setup as Marco (he uses REs I don't have); I used an Antidote with full voices instead of a Thor (because I couldn't even see the footprint of a Thor, it's too small), and a chain of effects, some of them quite CPU intensive (Echo, Alligator, Pulveriser, RE_2A, Ozone Maximizer, Chenille, Fet Compressor, DR-1 reverb, Scream4, RV7000, MClass Stereo, RV-7, DDL-1, Polar, Rotor, 4dyne compressor, DCAM env shaper, Disperser, Faturator and RE202 Exciter).

This is the CPU usage when Reason is not playing, and the track is un-muted.
3-idle.PNG
3-idle.PNG (72.39 KiB) Viewed 2528 times
The value we're looking for is the last one on the right (10.21% average CPU use in the last minute).

Here it is when playing the loop (again, averaged on 1 minute):
1-live.PNG
1-live.PNG (70.89 KiB) Viewed 2528 times
And here it is when playing the bounced track, the live track being muted:
2-bounced.PNG
2-bounced.PNG (78.28 KiB) Viewed 2528 times
For the sake of completeness, I also measured without the track (empty song), just to check Reason's overhead: there's almost none, it reports 0% use on average.

On my setup, bouncing in place reduces the CPU usage to what it is when the instruments are loaded but the song isn't playing.

User avatar
selig
RE Developer
Posts: 11685
Joined: 15 Jan 2015
Location: The NorthWoods, CT, USA

12 Jun 2016

WongoTheSane wrote:On my setup, bouncing in place reduces the CPU usage to what it is when the instruments are loaded but the song isn't playing.
Just to be clear, have you deleted the instruments that WERE loaded, because otherwise of course it's the same CPU because the instruments would still be loaded. Also, loaded instruments on muted tracks will not show any more CPU when the song is playing than when not, right?

Meaning…
Audio tracks don't take any CPU when the song isn't playing, but loaded instruments DO. Wanting to understand what you're seeing/reporting here.
:)


Sent from my iPad using Tapatalk
Selig Audio, LLC

electrofux
Posts: 863
Joined: 21 Jan 2015

13 Jun 2016

Thor is not the best device to test this as it is extremely efficient. Other devices consume a lot more CPU just by sitting in the Rack.

Also some devices have effects. I assume Antidote with active reverb still consumes cpu even when the track is muted- not sure though.

WongoTheSane
Moderator
Posts: 1851
Joined: 14 Sep 2015
Location: Paris, France

13 Jun 2016

selig wrote:
WongoTheSane wrote:On my setup, bouncing in place reduces the CPU usage to what it is when the instruments are loaded but the song isn't playing.
Just to be clear, have you deleted the instruments that WERE loaded, because otherwise of course it's the same CPU because the instruments would still be loaded. Also, loaded instruments on muted tracks will not show any more CPU when the song is playing than when not, right?

Meaning…
Audio tracks don't take any CPU when the song isn't playing, but loaded instruments DO. Wanting to understand what you're seeing/reporting here.
:)


Sent from my iPad using Tapatalk
I believe we agree, the idea was only to answer the question "by how much". There were two things I wanted to measure: 1) whether muting a track and not playing have the same overhead (because I could think of several optimizations that would save some further CPU when the track is muted) and 2) on a scale (CPU usage-wise) that goes from "no instruments loaded" to "instruments loaded and playing", where does "instruments loaded and not playing" stand.

I was quite certain the answer to the first question would be: yes, they have the same overhead (they do, more or less), but just wanted to make sure. The second question gives me a 60% saving on CPU usage when instruments are not playing, and merging the two answers gets me "bouncing saves 60% CPU usage on heavy tracks".

Bear in mind that this is only a back-of-the-envelope calculation (and not meant otherwise!), as my selection of devices was completely random, and I didn't repeat the test several times under different circumstances. It is only meant to give a very rough average of what muting the original track achieves in terms of CPU saving (as opposed to the hypothetical ideal "Deep Freeze" function that's been discussed here, one which would completely unload the devices from memory - I don't have another DAW so I can't compare).

Hope I'm not too far out on this one :?

Post Reply
  • Information
  • Who is online

    Users browsing this forum: No registered users and 18 guests