I am currently unable to use Scaler 2 the way I would like. When using Midi out in FL Studio, the plugin loses the exact midi timing after 10 bars, which causes audible phase problems with the kick/bass.
I made the problem visible in a video with a multichannel oscilloscope. Here you can see from bars 10-11 how suddenly there is no longer a perfect sync and the phase problems begin as a result. It took me ages to figure that out. I have a kick in G and a sub in G (triggerd by scaler) and the phases are perfectly aligned and optimal for my taste. … up to bar 10 (at 126 BPM) as you can see in the video.
In the second half I let the same Serum play the MIDI directly. You still have to do the nice alignment there, but you can see that it stays the same and there are no wild fluctuations after 10 bars.
As a point of reference for the devs: I would look in the direction of multithreaded processing, because if I disable the GLOBAL option in FL Studios audio settings, the problem does not occur. (?) But this is not a real solution as all plugins are affected now and it becomes very painful for large projects.
Here’s my video - no audio but the analyzer is pretty clear:
What is it Scaler is doing here? There is nothing in Pattern 1 even though you have DAW sync turned on there is nothing for Scaler to play in sync in the pattern. You don’t have Bind turned on in Scaler either so you can’t trigger the pads in section B. You have C3 on MIDI notes but the way Scaler is set those would just pass through. So as far as I can see Scaler isn’t set to do anything here.
Scaler is used as a Midi effect. It’s Midi out feeds Serums MIDI in. Scaler is used all over the project just to translate “only white keys” into the selected scale. Scaler is getting “C” translates to “G minor harmonic scale” - what means G in this case and passes “G” and all velocity information to Serum. This happens when no module (A, B or C) from Scaler but a scale and Key Lock is selected). This way I can sync the selected scale inside scaler over a very huge project with two clicks, without changing or transposing any midi. Some Scalers play chords, some just single notes.
— Edit from the future: the following passage is not right —
But since I spend the half day with this problem I now have to admit: The slight de-sync is not Scalers fault. It’s a bug in Serum. If the plugin Serum (VST2) is allowed to multithread the shown effect happes each 10 bars. So I have to tell Steve Duda and not the Scaler team. Sorry for that, I simply didn’t expect Serum to be the cause.
I keep the video up for some time. But no bug fixing and support needed. This can be closed, I guess.
Yes. Of course… I’m a fanboy of scaler and install each version right away. You can’t imagine what this tool did for me…! Oh, well… possibly YOU can
But as I wrote in my last post, my problem has to do with Serums handling of Midi-in (with FL Studio, and multithreaded processing enabled). I didn’t realize at first and opend this thread.
Scaler is pushing out it’s midi pretty fast and acurate. I now tested it with other synths, my oscilloscope and sinewaves. Like I did in my video - result: perfect timing, only a few samples later than a midi directly feeding the synth.
Only Serum had the shown problem - and since I use Serum a lot I didn’t noticed. But now im happy, my beloved scaler plugin is working fine and I contaced Xfer about my issue - and I even found a workaround for now.
Yes, me too - and Steve is much more structured and efficient than me, it seems …: He yesterday just asked me to print the notes from Scalers output… and he was right. There is something off, too. Then I tested more and used NI-Massive X as Scalers Midi “target”. And Massives audio shows the same shifting as in my video if I use Scaler2 as Midi in. I did the same experiment as in my video. It’s not shifting if it plays the same notes as “normal” Midi on the grid. This seems to be not over… Im totaly confused now
Sorry for the back and forth…
My system:
Windows 11 pro 64
9900k 8 Core
32GB RAM
FL Studio 20.9.2
Scaler 2.5.0 VST3
One of my findings is right and I can reproduce it: The entire problem is gone if I disable “Multithreaded generator processing” in FL Studios Audio settings. And there is no problem if I don’t use scaler to trigger the same instrument. Not waveform shifting.
I wouldn’t be that picky about it, but this stuff is pretty audible because the (Bass)waveform shifts that much it cancels out with other waveforms (Kick in my video). That way I started investigating, because the sound changed and it shouldn’t.
I don’t know what to do anymore. I could just create more Videos about the issue if needed. And to permanently disable multithreading in FL is not an option since the CPU usage skyrockets that way.
(Now the thread title is not right anymore, but I can’t edit it.)
This issue seems to be restricted to FL Studio only. I have run the same test with FL, Ableton and Logic Pro X and only noticed any timing discrepancies with FL Studio. Though my findings did not change regardless of having Multithreading on or off.
It could have something to do with the third party drivers that FL uses, ASIO4ALL.
I’m not using the third party ASIO(4all) Drivers in FL, I use the Audient own “ID” ASIO driver for my interface. I tried them with 128 and 512 Samples latency.
Maybe I wasn’t patient enough but for about 20 Bars I had no such timing discrepancies with “Multithreaded generator processing” disabled in FL > Options → Audio Settings.
Awww, I was hoping this is not on Imageline/FLs site now, but if required I’ll try to contact them about this, too. I really wish I could solve this problem.
I don’t have any other “midi VST” plugins like Scaler. Maybe I can test FL with Scaler 1 and/or the midi-out of one of my VST-Synths ARP, to see if there is this “shifting in time”, too.
Thanks for the support and again for reproducing the issue.
I contacted Image-Line and showed some video examples to them. They now told me they will improve this in the future. Very friendly and clear.
I’ll try to update this thread after the fix/improvement is rolled out.
(I can only speculate that this could take a while, but I don’t know.)
Who requires a better midi-out timing with Scaler 2 in FL now, for scheduled productions etc.:
-Backup your project file first
-Go to Options → Project general settings → and set “Timebase (PPQ)” to 960.
This will result in much less time-shifting for now. (5x - 10x less shifting in ms/samples). It needs to be set for each project individually. The downside: Your CPU must be able to handle this setting in large projects. But anything above the default PPQ of 96 will give an improvement, just try another high PPQ above 96, if your CPU is struggling with the max. value.
Thanks for the update, we’ve been extremely pedantic about timing issues in scaler as more and more people become reliant upon it and we are always needing to open old projects with scaler in it. So thanks for the reporting and we hope to see some improvements from imageline soon.
You can’t imagine what this tool did for me…! 
Oh, well… possibly YOU can 
…: He yesterday just asked me to print the notes from Scalers output… and he was right. There is something off, too. Then I tested more and used NI-Massive X as Scalers Midi “target”. And Massives audio shows the same shifting as in my video if I use Scaler2 as Midi in. I did the same experiment as in my video. It’s not shifting if it plays the same notes as “normal” Midi on the grid. This seems to be not over… Im totaly confused now 
