I use reaper for some heavy vsts. Does anyone have a link or ss of the most efficient settings that dont sacrifice quality ? Thankyou. Also, oversampling? can you turn that on by channel? what is its intended purpose ? Does it hurt or help (majority)?

 

oversampling = usually not useful. Work in 44-48k. And set the resampling settings to 64pt for render, 16pt for realtime.

For pitchshifting, soundtouch, soloist, efficient, use less CPU than Pro, but sound worse. So test on your material first.

 

Use the freeze option (and undo for edits) as you work through your project.

 

Last edited: Aug 13, 2022

I use the Reaper oversampling option, but always listen first for any improvements.

 

You will almost always see the greatest improvement caused by allowing your cpu/plugin the additional time to "do more work", by bouncing with oversampling enabled. And then you disable it after you bounce so you do not get unnecessary lag out of the deal. The benefit is not very obvious for realtime playback processing.

 

I would only use 64pt for rendering if I know that there would be no resampling happening at all.

For mastering I always use the resampling algo in iZotope RX, but if I use Reaper then r8brain free is absolutely the best, and it's also very CPU efficient!

Here you can see some tests between different resampling algos https://src.infinitewave.ca/

 

- You can activate this for the complete FX chain of a track or for specific pluggies only.
- Its purpose is to reduce aliasing which can be caused by any plugin that adds harmonics.
- It hurts for sure the CPU but can help to sound more analogue/clean.

 

ATME, there's no difference. Where should it come from anyways?

 

Sometimes even the best math is best checked at 16X ?

 

Sorry, if you're adressing me, do you mind to rephrase, I didn't understand it.

 

default, there isn't much you can improve,
maybe question should be "best windows settings for audio" but such topics have been discussed before in various threads

 

Make sure anticipative processing is turned on. It is turned on by default, but if you accidentally turned it off, turn it back on and you will see RT CPU plummet. This is especially important for plugins, but if you're recording a lot, which I assume you are not, it could be beneficial to turn it off for lower latency. Media buffering is used for playing WAVs and when you play many WAVs you might consider making it bigger than 1200ms.

Aside from Anticipative processing and media buffering, there isn't many settings that can help CPU usage except for rendering or freezing tracks [use r8brain OverSampling for whole FX chains, 2x or 4x should be enough]. Reaper is incredibly efficient by default. I am obsessed with performance [COD definitely ] so you can be sure I've done my homework. I've been using Reaper since 2006. Cheers!

 

For us geeks that's not obsessive, it just means "Call Of Duty"
Anyways, there's many options for performance in Reaper deep settings like the ones SineWave said. There's even options for assigning cores, disable fonts or waveform view antialiasing... it's just incredible.
Be sure to backup all your settings first, there's a nice feature to export them in one file.

I seem to recall there's a video-tutorial in Reaper Mania or similar YT channel for this.

 

Keep in mind any artefacts <120dB will be lost in 16b, and anything <60dB will likely be lost in the music mix, so you can get away with "poor" SRC.

 

This video is probably worth watching too..

 

Everytime someone's linking that vid, I need to mention that Dan made a mistake by comparing HQ in Saturn, which is 8x OS, to less than 8x higher SR.

 

What sample rate do you use?

 

Less than 352,8kHz but that's not the point. Any higher OS or SR will cause less aliasing in such a case. The comparison is flawed, that's my point.

 

hoping someone can confirm I'm not completely missing the boat here, as I just switched to reaper mostly due to getting into nebula which is CPU heavy and has a bunch of 96khz samples. basically Ive been running reaper in 96k (to run the 96khz nebula libraries without down sampling it, I have an amd 5800 and between that and reaper being cpu friendly it's not inhibiting on the CPU. I used to just do 44.1k cause that seems to be fine, but uh now I'm not. probably just because I can I've been using a ton of heavy processing on tracks and rendering them down (thanks to reapers track management options. great daw) in segments so if I decide to change something later I can unfreeze it just to the vst I want to tweak. necessary? no but that's what I've been doing, and at the same time I've been picking like 2/4x oversampling as I render the files.

I don't think the oversampling is like enhancing anything but I guess my thought process is that it's just an extra step to try to avoid any unwanted aliasing. I guess I'm a bit paranoid that by doing so much processing in steps I might be adding what individually maybe inaudible artifacts but by the end have stacked up and I hear some weird artifact type stuff. although I'm generally confused and some nebula libraries seem to be better than others, or at least you def gotta pay attention to input levels and such.

anyway, in the end I just do this to keep busy so it doesn't really matter. and hey a placebo high is still a high right?

 

Last edited by a moderator: Aug 21, 2022

In principle, yes but if this happens

there's probably a plugin (or a few) which doesn't support a higher SR than 96kHz and with additional oversampling this can lead to artifacts.