Saving DSP by switching off devices when not in use?

This forum is for discussing Reason. Questions, answers, ideas, and opinions... all apply.
Post Reply
User avatar
MrFigg
Competition Winner
Posts: 9137
Joined: 20 Apr 2018

19 Oct 2020

Can’t remember if I dreamt this but I seem to remember reading somewhere that by having devices automated on/ off until they are actually in use in a song is a good way to reduce DSP. Is that a known method? Or what was it instead? Thanks.
🗲 2ॐ ᛉ

chaosroyale
Posts: 728
Joined: 05 Sep 2017

19 Oct 2020

Yes, I wrote a thread on this a couple of years ago. EDIT> this is about VST instruments and FX.

Simple version:

1.You can automate VST devices OFF to save DSP

2. VST instruments will need to be in a combinator, and you switch the combinator off. VST FX have a switch on each device.

3. This CAN change latency compensation reporting. In this case, some people have reported problems like clicks. So you may want to check a test render. In cases where there is no latency difference, there will be no problem.

4. I came up with this workaround in Reason 9.5. The performance boost in 10.3~ has made this workaround less important. Your mileage may vary.
Last edited by chaosroyale on 19 Oct 2020, edited 2 times in total.

User avatar
MrFigg
Competition Winner
Posts: 9137
Joined: 20 Apr 2018

19 Oct 2020

chaosroyale wrote:
19 Oct 2020
Yes, I wrote a thread on this a couple of years ago.

Simple version:

1.You can automate the devices OFF to save DSP

2. VST instruments will need to be in a combinator, and you switch the combinator off. VST FX have a switch on each device.

3. This CAN change latency compensation reporting. In this case, some people have reported problems like clicks. So you may want to check a test render. In cases where there is no latency difference, there will be no problem.

4. I came up with this workaround in Reason 9.5. The performance boost in 10.3~ has made this workaround less important. Your mileage may vary.
That’s great. Thanks for the clarification:).
🗲 2ॐ ᛉ

User avatar
mcatalao
Competition Winner
Posts: 1827
Joined: 17 Jan 2015

19 Oct 2020

chaosroyale wrote:
19 Oct 2020

2. VST instruments will need to be in a combinator, and you switch the combinator off. VST FX have a switch on each device.
You mean Reason instruments because VST's can be disabled in the vst device.

However, in reason and RE instruments, which don't have an off switch, this does not work because they don't stop cycling to be ready for a midi input. What I mean is that the fact they are inside or outside a combination is irrelevant for performance in the case of instruments even if you disable the "superseeding" combinator .

But you can test it. Just put an instrument inside a combinator, and feed midi to the combinator and you'll see it still plays even if the combinator is disabled (thouth the combinator doesn't output audio the device is still on and if you put a mixer in the disabled combi you'll see there's signal going into the mixer).
The off button will only work over devices that have that functionality.

So, in the case of instruments, as I've been telling for some years, if they are heavy in your mix, your best option is to bounce them to audio and delete them. For the case of effects processors, you can disable them, but for the sake of not having pops and so on, you have to be sure you activate and deactivate them from absolute silence.

IMHO, all this issue would not be a problem if we had a more mature freeze function. Bounce in place is a way to do it but the bypass of devices or only the track is not enough as we know a lot of re's still eat CPU over silence (not a big of an issue with reason core devices and the stuff in the main mixer).

If you want a good workflow for less powerful machines, what i usually do is separate production steps into 2 or 3 projects (I also have been talking about this in the forums, here and the old PUF). The idea of separating processes is not only important for performance, at it also helps a lot with "objectivity". A lot of times we are a Jack of all trades, and not only compose, sequence, record, mix, and master our songs. When my projects are too big, i separate them into 2 or 3 projects where:

1 - Composition/Sequencing/Recording
This project is more related to the song building process, and it's focused on music making tasks, like harmony development, melody creation, so sequencing and recording live instruments happens here. I usually try not to do any mixing tasks, aside from maybe panning and primary leveling on the devices.

2 - Mixing
So at this stage if i have a very busy project eating more than 50% CPU, i know i will not have a lot of juice for mixing. So at this time, i will create a consolidated export of my project. I prefer to start with a new project so i export all tracks to a folder and then re-import. For this i use the Export function under File -> Bounce mixer Channels. In there you can specify how you want to export stuff, but you really want these to be as raw as possible so you start with a completely clean mix. Usually my settings here will be (with the exception of the project sample rate and depth which should be the same as the projects):
BounceMixerChann.JPG
BounceMixerChann.JPG (78.85 KiB) Viewed 646 times

Then I create the new project and import the files for the mixing stage. As audio loads a lot less DSP than the synths, you get a nice clean project with almost 0% cpu/dsp to start your mix.

3 - After the mix is done, there can be a lot going on too. If the song is single, I usually try to master it in the master bus, with ozone and other devices. But if again, the mixing project is above 50/60 % chances are I'll have clicks and pops on the mastering project. Also, if the song is part of a bigger project it is good to master it in context, so I'll create a mastering project where my mastering chains will be loaded on channel inserts and each stereo song is on a stereo track.
Again, as the songs are audio at this point, I start with a clean slate for this process and I have virtually 0 for my mastering chains. It's important to say, Ozone and TAPE witch are widely used in my mastering chains usually don't load CPU when they are not fed with audio which allows me to have 12 instances of ozone on the same project for 12 songs, as they never play at the same time in this kind of project.

I hope this helps. At some point with Reason 6 i was using a Q6600 4 CPU processor that needed this more often. Then i started working with a I4790k and though I don't need to work like this for performance purposes, I incorporated this in my workflow and my CPU almost never hits more than 50% in 40 to 60 track projects, specially after RS improved VST and overall performance in R10.4

Cheers,
MC

chaosroyale
Posts: 728
Joined: 05 Sep 2017

19 Oct 2020

I was only talking about VST, not RE instruments and FX, I did not make that clear enough, and have edited the post to make it obvious.

How would you automate turning off VST instruments without putting them in a combinator? I do not believe it can be done. VST FX have a switch on the device, but VST instruments do not.

mcatalao wrote:
19 Oct 2020

You mean Reason instruments because VST's can be disabled in the vst device.

However, in reason and RE instruments, which don't have an off switch, this does not work because they don't stop cycling to be ready for a midi input.

User avatar
mcatalao
Competition Winner
Posts: 1827
Joined: 17 Jan 2015

19 Oct 2020

chaosroyale wrote:
19 Oct 2020
How would you automate turning off VST instruments without putting them in a combinator? I do not believe it can be done. VST FX have a switch on the device, but VST instruments do not.
I don't have reason in this machine, but If the vst device does not have a button, then it's similar to the RE and Reason device case, that i am sure. I assumed it would work because i don't remember the device being different from vst instrument to vst effect.

So, if the vst device does not have the off button, it's useless to put them inside a combi. But again, no matter as much reason i know, it's not easy to remember everything.~

Anyhow the OP's question was not specific to VST's so this information is useful to him (i hope).

Again, i can't stress enough in case RS is lurking around, we need a good freezing feature ASAP.

User avatar
ctreitzell
Posts: 12
Joined: 20 Oct 2020

22 Oct 2020

MrFigg wrote:
19 Oct 2020
Can’t remember if I dreamt this but I seem to remember reading somewhere that by having devices automated on/ off until they are actually in use in a song is a good way to reduce DSP. Is that a known method? Or what was it instead? Thanks.
I wrote about that a few years ago; version was 8.3
I was talking specifically about Reason/PHeads devices and REs; VST wasn't in Reason a few years ago
There is absolutely an on/off button on the VST device panel

my old account got deleted....thus so did the insights from my reporting many use cases to this forum.
turning off devices absolutely saved CPU cycles and I was able to finish the massive song I was working on


my machine is plenty powerful now, I can roll freely
I'm a rack guy guitar player
Reason 1, 2.5, 4, 8, 10
SawStudio
i7 8700K 64GB RAM
RME UFX+
Adam A7X

Post Reply
  • Information
  • Who is online

    Users browsing this forum: Faastwalker and 33 guests