Reaper worst Automation??

[Trancit]

Sorry for this title but I can't think of anything else to say...

Of course I cannot tell if it´s a wrong setting, if it´s a bug or just the way it is

I stumbled upon this as I noticed there is a bug in S1 regarding automation and PDC...

I ran this test using in all DAWs the same plugin (Hornet Track Utility MK2) and the same type of automation...
No latency envolved...nothing else is going on...

All DAWs show the same result more or less clean...sometimes there is more of the transient...sometimes less...

But what Reaper does here I can only call sloppy...
Watch the video and decide yourself: https://youtu.be/-kvWpLTM3Vc

[Trancit]

Btw:
I tested this behaviour with a lot of other plugins including Reaper natives too...

It was always the same result so it´s not the plugin causing this!!

[Time Waster]

Reaper's automation works per block, so there can easily be a few milliseconds variation between the occurrence of each step. Placing the automation points that close to the transients is asking for trouble.

[akademie]

You didn't say what are you automaing in the first place.
Which parameters? mute, bypass?
Are you sure you are running all the DAWs with the same latency for example?

I tried similar kick loop here and automating track's Volume, Mute, ReaEQ plugin bypass (plugin's output fader set to -inf so I get signal or not). All of them works steady - solid (only mutes were little "behind", but that's because Mute is set to have 5.0ms fades by default, again it was dull on attacks but steady as a rock for each hit, well yes for the each other hit

(32bit Reaper v6.33rc3 & Win10 x64)

EDIT:

Originally Posted by akademie

You didn't say what are you automaing in the first place.

Well, I watched the video again, carefully and I see in Studio One part it can be seen that you automate plugin's level parameter, yeah. I will test that too in a moment.

EDIT2:
So I tried the same with JS Volume... effect and also ReaDelay (Dry out fader, Wet was static on -inf) and if anything, there is some microscopic bleed of the transient from "muted" hits (really like click/pop), again steady not random. And if I tried to offset the media playback about 128samples to the right (btw 128 samples is my buffersize), then it gone.

[Trancit]

Please try with Reacomp´s dry volume (wet turned to -inf) which behaves on my side exactly like shown in the video...

Perhaps it´s just a setting on my side... I would be glad if so...

[akademie]

Originally Posted by Trancit

Please try with Reacomp´s dry volume (wet turned to -inf) which behaves on my side exactly like shown in the video...

Perhaps it´s just a setting on my side... I would be glad if so...

OK, tried with ReaComp, automating the Dry parameter.
Result is exactly the same as with ReaDelay (only steady micro pops at muted part's transients are there).

I would recommend to try "install" portable Reaper with clean config, with no extensions and scripts and create brand new project and check behavior). I would keep latency buffer size at 128 or 256 samples.

Also note that my tests were with the laptop's internal onboard audio card (Realtek) and WASAPI Exclusive interface selected in Reaper at 48kHz.

[Trancit]

I just tried again too...

I have to shift the envelope 3 ticks to the left before I get rid of most of the unwanted behaviour...

This seem to be the proof of automation working in blocks...

If this is true ...sorry again ...but then my statement from the title is valid for me...

None of the other tested DAWs needs any kind of treatment for giving better results out of the box than Reaper with treatment...
For a nowadays DAW which likes to be on a high technical level that´s quite poor...

[akademie]

I tried to switch audio interface selection in Reaper and when "WASAPI Shared" was selected then it started to act badly = randomly (kind of like in your video).
Then back to "WASAPI Exclusive" at 512 samples and also chopy results (still similar to your video = very random).
After setting buffersize to 256 it settled down and at 128 samples it was back solid as my previous tests (still bleeds are present but at least no random behavior).

[Time Waster]

Originally Posted by Trancit

I just tried again too...

I have to shift the envelope 3 ticks to the left before I get rid of most of the unwanted behaviour...

This seem to be the proof of automation working in blocks...

If this is true ...sorry again ...but then my statement from the title is valid for me...

None of the other tested DAWs needs any kind of treatment for giving better results out of the box than Reaper with treatment...
For a nowadays DAW which likes to be on a high technical level that´s quite poor...

There is a trade off between automation accuracy and CPU performance. REAPER is noted for it's CPU performance.

[DaneJH]

This is unfortunate, have you forwarded this to REAPER Support?

If there is a trade off between CPU and Accuracy, then why not make it an Automation preference?

[pepe44]

I get a nice and accurate result using the JS pink noise plugin and volume automation . no lag.
What is your audio interface and tell us what is the Buffer size and sample rate you are using for multiple automation DAW tests.

[mozart999uk]

I haven't tried your test but coming from Cubase, I was pleasantly surprised how much quicker reaper responds to automation than Cubase. :-)

[Coachz]

Originally Posted by Trancit

Sorry for this title but I can't think of anything else to say...

Of course I cannot tell if it´s a wrong setting, if it´s a bug or just the way it is

I stumbled upon this as I noticed there is a bug in S1 regarding automation and PDC...

I ran this test using in all DAWs the same plugin (Hornet Track Utility MK2) and the same type of automation...
No latency envolved...nothing else is going on...

All DAWs show the same result more or less clean...sometimes there is more of the transient...sometimes less...

But what Reaper does here I can only call sloppy...
Watch the video and decide yourself: https://youtu.be/-kvWpLTM3Vc

Why not change your title to "automation problems". The current title is insulting imo.

[cool]

This is "Envelope Display" option "Transition time for automatic created envelope edge points". My default option is 20ms. Too mutch. You can set to 0, to get same sound as other DAWs. But, you will most likely get clicks on low frequency sounds.


And. Here i see two problems.
1. I have noticed on several occasions that Reaper has some default "not convenient" settings. That is, if you compare Reaper and other DAWs out of the box, with the default settings, then Reaper will lose, although it will have potential.

2. Second. Now, if you shape the envelope from the selection, then the left part will be formed correctly, keeping the attack. But the right one is not. It will "eat" the attack of the next sound placed on the grid.
Correcting the second point will allow you to ignore the first.

[Trancit]

The hasn´t any influence at all of what I am doing here...

I am always working at a buffer size of 512 samples @44.1Khz...

It turned out that Reaper´s automation precision is indeed bound to the audio device block size...
Turning down the buffer size to 256 samples or less helps indeed with this problem...

Nevertheless is that imho a very silly behaviour...
If it´s about having a smaller CPU footprint for automation what does it serves you that you have to work with smaller buffer sizes for getting a higher precision??
Smaller buffer sizes cost tons more of CPU than having a system like all other DAWs obviously use as they don´t have these problems at all...

Sorry... either the automation in Reaper is broken and it isn´t meant to work this way or the title of this thread is spot on

[Coachz]

Titles like that wont get users like me who like Reaper excited to help you, but by all means, you do you.

[bodfum]

Options: Rendering settings: Block size to use when rendering. Set this to 4. be aware your renders will be slow

[Trancit]

Originally Posted by Coachz

Titles like that wont get users like me who like Reaper excited to help you, but by all means, you do you.

I am pretty sure there is nothing you can do to help...

[Trancit]

Originally Posted by bodfum

Options: Rendering settings: Block size to use when rendering. Set this to 4. be aware your renders will be slow

How shall this help when plying back the project in realtime??

[bodfum]

It won't change what you are hearing as VST2 does not support sample accurate automation, but rendered files will not be affected

[Cosmic Pig]

If this is important to you then by all means use another daw. I am way not a power user but I can think of several ways to stop the transient click.

Maybe Microsoft Excel might be more suited to your needs.

See now.. if that kick sound was me I'd take out all the volume humps so it was nice and flat at minus infinite. The big bonus there is it's way less nitpicky to mix.

[mschnell]

Originally Posted by Time Waster

Reaper's automation works per block, so there can easily be a few milliseconds variation between the occurrence of each step. Placing the automation points that close to the transients is asking for trouble.

This is perfectly "normal", as with VST2 plugins, the VST2 specs only allow for parameter modulation on a per block basis.

The VST3 specs allow for sample accurate parameter modulation. (AFAIK, Reaper does support this to some extent). But may Plugins don't support this up til now. E.g. the "JUCE" library, that can nicely be used as a basis for doing VST3 plugins, does not yet support sample accurate parameter modulation (I asked the makes of JUCE some two weeks ago.)

-Michael

[mschnell]

Originally Posted by Trancit

Nevertheless is that imho a very silly behaviour...

It simply is the VST2 specs. You can't use or do a VST2 with more timing accuracy.

Did you try some VST3s ? They might be better but only if as well the VST provider and Reaper implemented that part of the specs.

-Michael

[airon]

VST2 parameters are per ASIO/Coreaudio block.

VST3 parameters can(!!) be sample accurate.

JS plugin parameters can be sample accurate.

Track volume/pan/pre-fx volume/mute, send level/mute/pan run on a per-sample basis..

That's it. It's ok if this is a deal breaker. I suggest always using track volume for sample accurate stuff, if Reaper is the DAW you're using. If not, perhaps pick a DAW that suits you best. Nobody will think any less of you.

[pepe44]

Originally Posted by airon

VST2 parameters are per ASIO/Coreaudio block.

VST3 parameters can(!!) be sample accurate.

JS plugin parameters can be sample accurate.

Track volume/pan/pre-fx volume/mute, send level/mute/pan run on a per-sample basis..

That's it. It's ok if this is a deal breaker. I suggest always using track volume for sample accurate stuff, if Reaper is the DAW you're using. If not, perhaps pick a DAW that suits you best. Nobody will think any less of you.

This might explain why my test was very precise. I used a JS plugin.

[bFooz]

Originally Posted by Trancit

But what Reaper does here I can only call sloppy...
Watch the video and decide yourself: https://youtu.be/-kvWpLTM3Vc

Yes, it's like the whole Reaper (untold) philosophy = "it works but is unusable for people".

[Pink Wool]

So changing the Track mute fade under Mute/Solo in preferences to 0 ms clears this one, at least on my end...

[akademie]

Originally Posted by Pink Wool

So changing the Track mute fade under Mute/Solo in preferences to 0 ms clears this one, at least on my end...

Well, but the thing is that OP does automate Output volume/Level of the plugin, so the mute is not the solution (e.g. he/she may want to attenuate level by big amount but not to -inf and then wrong timing will still be issue).

I think that Justin's entry here would be very welcomed to step-in on this subject for clarification, @Justin, please.

(Why we cannot have the same experience like OP shows with number of other DAWs ? - "IF" set-up correctly, of course).

[bodfum]

Is this not just a case of with Reaper WYSIWYG or in this case hear. I would assume the other DAWS are compensating for VST2s lack of sample accurate automation which would explain why some do it better than others.
With Reaper you compensate for this yourself and then have the option to reduce block size at render.

[bFooz]

Originally Posted by bodfum

Is this not just a case of with Reaper WYSIWYG or in this case hear. I would assume the other DAWS are compensating for VST2s lack of sample accurate automation which would explain why some do it better than others.
With Reaper you compensate for this yourself and then have the option to reduce block size at render.

Other DAWs are user-centric meaning they use code as a mean for a usable design and if there is a need for some bending in the name of usability, they do bend it. Reaper is dev-centric and it uses code for its own sake with usability having a lower priority.

[akademie]

Originally Posted by bodfum

Is this not just a case of with Reaper WYSIWYG or in this case hear. I would assume the other DAWS are compensating for VST2s lack of sample accurate automation which would explain why some do it better than others.
With Reaper you compensate for this yourself and then have the option to reduce block size at render.

Unfortunately, there are still some tiny pops even if render blocksize is set to 4 samples (=minimum allowed), tested

[bodfum]

Originally Posted by akademie

Unfortunately, there are still some tiny pops even if render blocksize is set to 4 samples (=minimum allowed), tested

At 4 samples there is still going to be some delay, 4 samples worth.
In which real world situation would you use automation like this though? No fader moves from -inf to whatever level in 0.000000ms

[bodfum]

Originally Posted by bFooz

Other DAWs are user-centric meaning they use code as a mean for a usable design and if there is a need for some bending in the name of usability, they do bend it. Reaper is dev-centric and it uses code for its own sake with usability having a lower priority.

In this case watching the video shows me that the other DAWs workarounds do not work, in which case I would much rather Reapers philosophy, WYSIWYG

[Trancit]

Originally Posted by mschnell

It simply is the VST2 specs. You can't use or do a VST2 with more timing accuracy.

Did you try some VST3s ? They might be better but only if as well the VST provider and Reaper implemented that part of the specs.

-Michael

Sorry, I don´t understand!!

I am using the same plugin in all the DAWs shown in my video...

Every DAW handles this task very well... except Reaper...

And that´s now the fault of the VST2 specs???

[Trancit]

Originally Posted by Pink Wool

So changing the Track mute fade under Mute/Solo in preferences to 0 ms clears this one, at least on my end...

Doesn´t change anything here...

[EpicSounds]

have you seen this?

https://www.admiralbumblebee.com/mus...utomation.html

[mschnell]

Originally Posted by Trancit

And that´s now the fault of the VST2 specs???

Supposedly it's due to different configuration (e.g. block size, ...)
-Michael

[Trancit]

Originally Posted by EpicSounds

have you seen this?

https://www.admiralbumblebee.com/mus...utomation.html

This is about how smooth the automation is in a long fade...not about precise timing...

[Trancit]

Originally Posted by mschnell

Supposedly it's due to different configuration (e.g. block size, ...)
-Michael

Nevertheless there is definetely working something worse in Reaper compared to all the others...

[mschnell]

Sorry but this needs to be some kind of Hoax (if concerning VST2s).

-Michael