Question about VST vs RE performance (uh.. hi Mattias?)

This forum is for discussing Reason. Questions, answers, ideas, and opinions... all apply.
Post Reply
chaosroyale
Posts: 730
Joined: 05 Sep 2017

01 Sep 2018

Will the upcoming VST performance optimizations also improve RE performance?

Why do I ask? I have noticed that some REs use a lot more DSP than similar native devices, even simple things like splitters and mid/side processors. In fact, it uses less DSP to swap out some REs for comparatively over-spec'ed VST devices. Now, we know the VST performance at the moment is quite heavy, so it seems that REs are heavy/un-optimized too.

So, will the under-the-hood improvements bring benefits to both? Or do RE's rely on completely different code to the VSTs? I understand if this information is still under wraps.

I like a lot of the utility extensions from people like Jiggerypokery and Lectric Panda, and it feels inelegant to be swapping them out for big expensive VSTs in order to save on CPU cycles, or having to use a chain of several spiders instead of 1 switcher/splitter RE for the same reason.

Can Props or any RE devs shed light on this?

Note: this a technical question, not a thread for complaining about stuff. Save your energy guys.

User avatar
Loque
Moderator
Posts: 11228
Joined: 28 Dec 2015

01 Sep 2018

That might be that RE benefit from that refactorings, but its very speculative, since we need to know the bottleneck of RE and VSTs. VSTs seem to have problems with sample calculation which is a fundamental design of Reason. The RE seem to have an other bottlneck, maybe its the API with too much overhead or the device sandboxing, which is stronger in RE compared to VSTs...Very speculative...i cross fingers that also the RE will be faster, because its really anoying that a few RE bring the system down on its knees on a 6 core, 12 threads and 5ghz CPU...
Reason12, Win10

chaosroyale
Posts: 730
Joined: 05 Sep 2017

14 Sep 2018

Does anyone else have any insight about this? I want to use these excellent "utility" rack extensions to build complex sounds with all kinds of audio and CV routing, but the CPU hit is just too high.

User avatar
Oquasec
Posts: 2849
Joined: 05 Mar 2017

14 Sep 2018

Your machine needs to have a bench of at least 1500 to run reason past version 5 : /
And you need good asio drivers with little to no overhead that can run chrome on the side when watchin twitch or some shit
Producer/Programmer.
Reason, FLS and Cubase NFR user.

User avatar
EnochLight
Moderator
Posts: 8418
Joined: 17 Jan 2015
Location: Imladris

14 Sep 2018

chaosroyale wrote:
14 Sep 2018
Does anyone else have any insight about this? I want to use these excellent "utility" rack extensions to build complex sounds with all kinds of audio and CV routing, but the CPU hit is just too high.
I would imagine that if you've reached out to Mattias via proper channels (the usual social media suspects), and yet he still hasn't answered, my guess would be that because it's still too early to tell.

That said, if I was a betting man, I'd wager that RE performance will not be addressed in the forthcoming free-VST-performance update later this year. It's been highly publicized that VST-performance is what's going to be addressed. No mention of RE-performance - ever - was made to my knowledge (and I feel that would be an awesome "bonus selling feature" of the update, so.... there's that).
Win 10 | Ableton Live 11 Suite |  Reason 12 | i7 3770k @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro | Akai MPC Live 2 & Akai Force | Roland System 8, MX1, TB3 | Dreadbox Typhon | Korg Minilogue XD

User avatar
Emian
Posts: 712
Joined: 16 Jan 2015

14 Sep 2018

OP is right that it's not only VST's that have performance issues, also RE's.

And i too am hoping that in the upcomming "performance" update, not only VST's will be adressed but also RE's.
it seems only logical that Props would benefit more from that, & for me personally it will be an important factor for me for future purchases...

But i still have high hopes for props ;)


"i might be established, but i'll never be establishement "
- Dave Clarke -www.soundcloud.com/emian

User avatar
Timmy Crowne
Competition Winner
Posts: 357
Joined: 06 Apr 2017
Location: California, United States

14 Sep 2018

I’d like to know too. I recently built a big guitar Combinator that I wanted to share with the community. It allows variable-speed strumming, variable strum-range, manual plucking of each string, patterns, mutes, hammer-ons/pull-offs, bending, and sliding. Each string has its own chain of CV and Midi devices. It really sounds good and uses the stock Reason samples and a couple free REs, but causes a CPU crisis because of the routing. It’s unplayable :/

Really hoping general performance sees an improvement. Otherwise some of the more “mad scientist” stuff may not be possible anymore.

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

14 Sep 2018

There is no „performance issues“ with REs, I tested the TSAR for example and while yes, you can get more instances in Logic overall you can NOT get more instances if you run them at 64 samples buffer size like Reason does.

So as always it‘s horses for courses. If you want to create crazy synth and sound design patches with a lot of intermodulation use Reason. If you want a bunch of Kontakt instruments with UAD plugins use Cubase or Logic.


User avatar
joeyluck
Moderator
Posts: 11088
Joined: 15 Jan 2015

14 Sep 2018

EnochLight wrote:
14 Sep 2018
chaosroyale wrote:
14 Sep 2018
Does anyone else have any insight about this? I want to use these excellent "utility" rack extensions to build complex sounds with all kinds of audio and CV routing, but the CPU hit is just too high.
I would imagine that if you've reached out to Mattias via proper channels (the usual social media suspects), and yet he still hasn't answered, my guess would be that because it's still too early to tell.

That said, if I was a betting man, I'd wager that RE performance will not be addressed in the forthcoming free-VST-performance update later this year. It's been highly publicized that VST-performance is what's going to be addressed. No mention of RE-performance - ever - was made to my knowledge (and I feel that would be an awesome "bonus selling feature" of the update, so.... there's that).
While that is the focus, who knows if it might not have positive results across the board?

From the blog:
Meanwhile, work on VST performance is ongoing. The result of this work will be released as a separate free update later this year. The reason it’s a separate release is because the performance work is an extensive rewrite of the inner workings of the program and requires an expert task force.
Maybe they'll find other ways to improve performance across the board? Or perhaps a culprit in the low VST performance is what is also causing performance issues elsewhere in the software?

User avatar
EnochLight
Moderator
Posts: 8418
Joined: 17 Jan 2015
Location: Imladris

14 Sep 2018

joeyluck wrote:
14 Sep 2018
While that is the focus, who knows if it might not have positive results across the board?
How dare you be a glass-half-full kind of person - that's normally my job! :D :lol: :puf_wink:
Win 10 | Ableton Live 11 Suite |  Reason 12 | i7 3770k @ 3.5 Ghz | 16 GB RAM | RME Babyface Pro | Akai MPC Live 2 & Akai Force | Roland System 8, MX1, TB3 | Dreadbox Typhon | Korg Minilogue XD

User avatar
normen
Posts: 3431
Joined: 16 Jan 2015

14 Sep 2018

joeyluck wrote:
14 Sep 2018
While that is the focus, who knows if it might not have positive results across the board?

From the blog:
Meanwhile, work on VST performance is ongoing. The result of this work will be released as a separate free update later this year. The reason it’s a separate release is because the performance work is an extensive rewrite of the inner workings of the program and requires an expert task force.
Maybe they'll find other ways to improve performance across the board? Or perhaps a culprit in the low VST performance is what is also causing performance issues elsewhere in the software?
Yeah, if they change the whole signal flow then why wouldn’t they be able to run single REs with a larger buffer as well. As I said elsewhere my suspicion is that the „extensive rewrite“ is concerning exactly this „issue“. I think they‘ll create a dynamic buffering system with separate buffer sizes for each step in the signal flow - which might also bring us _FULL_ PDC along the way.

The questions are these:
a) Do REs even need larger buffers as they‘ve been created with 64 samples in mind (most probably only the „ported“ ones)
b) Will existing REs need to be adapted if they want to use larger buffers (most probably they will)

User avatar
jappe
Moderator
Posts: 2441
Joined: 19 Jan 2015

15 Sep 2018

Even if improving RE performance/audio engine significantly would be unachievable, having knowledge about precisely what options there are for a user to optimize specific projects would be very helpful.

A project specific performance Wizard, enabling 1 click automatic performance fixes as well as manual mode where culprits are pointed out to the user.

I think it's possible to make a light weight performance profiler, but it could also be possible to switch it off.

Currently I'm keeping ballast devices in projects to deal with the situation that a project I was working with on my stationary PC turns out to be too heavy when trying to run the project on my laptop when on the go. This works since devices consumes DSP even when idle.

madmacman
Posts: 804
Joined: 18 Jan 2015

15 Sep 2018

I put high hopes into the improvements planned by Propellerheads!

Recently I watched two Walkthru videos of productions by German Synthpop veterans "De/Vision".

Both were done with Apple Logic. From what I count: 70-80 tracks (grouped), audio, virtual instruments, and lots of processing with FX. And of course this videos were captured on a MacBook in realtime during presentation/playback. The MacBook struggled a little bit here and there, but holy sh*t: after watching I was jealous! :shock:




Post Reply
  • Information
  • Who is online

    Users browsing this forum: No registered users and 5 guests