ReBuzz

[mcbpete]

If it would be possible to save a bmx with used machines into a zip file an reload the bmx into ReBuzz by using the same zip file as a termporary gear folder?

That's a great idea, though for compatibility maybe have it as a new file type. So you could have

.bmw - Song without samples
.bmx - Song with samples
.bmz - Zipped song with samples AND machines

[magmavander]

Very good idea HF

[mcbpete]

Odd, never had problems with the vst3shells in old Buzz and also ReBuzz and I had hundreds of vst3's installed. I even had some tweakings to speed up the scanning of the vst3shells. In older loaderversions it could take some minutes on first scan though.

For some reason it seems to be a doing a full scan every time reBuzz starts - viewtopic.php?p=17370&sid=2be9bc6c73a04 ... e0b#p17370

[HerrFornit]

If it would be possible to save a bmx with used machines into a zip file an reload the bmx into ReBuzz by using the same zip file as a termporary gear folder?

That's a great idea, though for compatibility maybe have it as a new file type. So you could have

.bmw - Song without samples
.bmx - Song with samples
.bmz - Zipped song with samples AND machines

That would be elegant. And a requester after loading the .bmz: "Do you want to import the loaded machines or discard after closing the song?"

[polac]

Does the debug console only show what ReBuzz does until the Machine View has loaded? The minute or two boot time seems to be a result of 'UIGetInstrumentPath: C:\Program Files\ReBuzz\Gear\Vst\vsti3shell.x64.dll' getting called nearly 2000 times every time ReBuzz gets launched(1868 times to be precise ), but I'm still trying to work out why it takes around a minute to add any VST3 as nothing seems to get logged in the console

Hmm, it should call the vst3shells only once on buzz start to check if there are some new plugins added to the shell. It shouldn't load the vst3 separately to check if it's instrument or effect. If it is another shell vst (like waveshell) it would do the check wether it's a instrument or a effect, but it shouldn't scan previously scanned vsts again, only the new ones.

However, if you load a song with lots of pvst instances, it would load the shells again and again for each instance(each time only one time load per instance), perhaps this is the problem, but i doubt you have 2000 instances open . I think I could circumvent the unnecessary loads somehow, I talked about it with WDE, we thought it is fast enough as it is.

[mcbpete]

Thanks for looking into this, I'd love to get back to Buzzin' again - for the past few years I've leapt to the dark side (Ableton)

To confirm the vst3 loader is being called this many times on just loading up reBuzz to a blank project, unfortunately as reBuzz doesn't appear to write anything to the debug console after the application is fully launched I can't see whether it's being called that many times again when loading a single PVST instance thus causing this >minute delay in adding it to the machine view (perhaps I can run Process Monitor to check...)

Additional info if it's of use: I ditched all my Waves plugins back in early '23 after their back-and-forth with perpetual licenses vs subscription, also their WUP plan made keeping plugins up to date cost more than buying them in the first place - so all the waveshells (think there were about 10 of them as I refused to update just to keep them on the save version number) have been removed from my machines.

Also my laptop is a relatively (when it was purchased about 4 years ago anyway ) well spec'd machine: i9 10900k, nvidia rtx 2080, 32gb ram, ssd drive hosting the OS, Buzz and the VSTs

EDIT: OK so yeah reBuzz is definitely going wild with it's file reading. Just launched Process Monitor and added 1 (one!) VST3 (u-he's Bazille) to an empty project -

This is called 6155 times:
ReBuzzEngine64.exe 11552 ReadFile C:\Users\mcbpete\AppData\Roaming\Polac\Vst Loader\1.1.10\X64\plugins.dat

and then this is iterated on for *every* vst in my VST3 folder (obviously listing a different VST in each instance!):
ReBuzzEngine64.exe 18240 QueryOpen C:\Program Files\Common Files\VST3\Airwindows Consolidated.vst3\Contents\x86_64-win\Airwindows Consolidated.vst3
ReBuzzEngine64.exe 18240 CreateFile C:\Program Files\Common Files\VST3\Airwindows Consolidated.vst3\Contents\x86_64-win\Airwindows Consolidated.vst3
ReBuzzEngine64.exe 18240 QueryNetworkOpenInformationFile C:\Program Files\Common Files\VST3\Airwindows Consolidated.vst3\Contents\x86_64-win\Airwindows Consolidated.vst3
ReBuzzEngine64.exe 18240 CloseFile C:\Program Files\Common Files\VST3\Airwindows Consolidated.vst3\Contents\x86_64-win\Airwindows Consolidated.vst3
ReBuzzEngine64.exe 18240 IRP_MJ_CLOSE C:\Program Files\Common Files\VST3\Airwindows Consolidated.vst3\Contents\x86_64-win\Airwindows Consolidated.vst3

and this on every subfolder (usually 3 for each VST3 - Contents, Contents\Resources and Contents\x86_64-win):
ReBuzzEngine64.exe 18240 CreateFile C:\Program Files\Common Files\VST3\OP-X PRO-3.vst3\Contents\x86_64-win
ReBuzzEngine64.exe 18240 QueryDirectory C:\Program Files\Common Files\VST3\OP-X PRO-3.vst3\Contents\x86_64-win\*.vst3
ReBuzzEngine64.exe 18240 QueryDirectory C:\Program Files\Common Files\VST3\OP-X PRO-3.vst3\Contents\x86_64-win
ReBuzzEngine64.exe 18240 CloseFile C:\Program Files\Common Files\VST3\OP-X PRO-3.vst3\Contents\x86_64-win
ReBuzzEngine64.exe 18240 IRP_MJ_CLOSE C:\Program Files\Common Files\VST3\OP-X PRO-3.vst3\Contents\x86_64-win

In total there's over 100k events created in ReBuzz.exe or ReBuzzEngine64.exe when I launch a plugin. I guess that's why it take a minute or so

@polac and/or @wde - let me know if you want the whole log sent to you if it's of any use ....

[HerrFornit]

...
However, if you load a song with lots of pvst instances, it would load the shells again and again for each instance(each time only one time load per instance), perhaps this is the problem, but i doubt you have 2000 instances open . I think I could circumvent the unnecessary loads somehow, I talked about it with WDE, we thought it is fast enough as it is.

Hello polac,

do you mean with "instances" also the same VST appearing serveral times in a song? Loading the VST only once per song would speed up the song loading, too? (Not sure how it works internally, but I describe it as it looks alike)

Thanx a lot for your support!

[Buzztler]

Ehm, I still have problems to get ReB goin'.
I copied my 32bit Buzz Gearfolder to Rebuzz and then installed version 1782 as Admin over it again.
When I start Rebuzz.exe for the first time, I suppress all Errormessages with ok, after this I get a final error, saying that an xml cannot be written to harddisk. After pressing "ok" I can work with ReB almost normal, old bmx-files load etc., Vst(i) still take several seconds to load ...
When I close ReB and start it again I get the first Errormessage I uploaded here. When I want to open an old bmx/bmw file, I get the second Errormessage I uploaded.

The "Splashscreen" stays on top, nothing happens, I have to use the TM to close ReB.
So, are these errors Win10 related, or what happens here, hope the jpg's help in any way?!

[HerrFornit]

Ehm, I still have problems to get ReB goin'.
I copied my 32bit Buzz Gearfolder to Rebuzz and then installed version 1782 as Admin over it again.
When I start Rebuzz.exe for the first time, I suppress all Errormessages with ok, after this I get a final error, saying that an xml cannot be written to harddisk. After pressing "ok" I can work with ReB almost normal, old bmx-files load etc., Vst(i) still take several seconds to load ...
When I close ReB and start it again I get the first Errormessage I uploaded here. When I want to open an old bmx/bmw file, I get the second Errormessage I uploaded.

The "Splashscreen" stays on top, nothing happens, I have to use the TM to close ReB.
So, are these errors Win10 related, or what happens here, hope the jpg's help in any way?!

Here 1782 started withoud errors. As you know I use W11, but I suppose your copied gearfolder is a candidate for the errors.

[mcbpete]

I get a final error, saying that an xml cannot be written to harddisk.

This might be a Windows Defender thing - ReBuzz is probably trying to write a config file to your user documents folder which is protected by default to prevent ransomware encrypting the folder

Try turning off controlled folder access and see if that suppresses the error - https://learn.microsoft.com/en-us/defen ... ed-folders (for Step 4, replace the step with: If controlled folder access is turned on, you'll need to turn it off )

[Buzztler]

Here 1782 started withoud errors. As you know I use W11, but I suppose your copied gearfolder is a candidate for the errors.

Well, I don't think the copied gearfolder is the problem, 'cause ReB should overwrite the data therein with it's new files, so that it can work otherwise it would be strange ... (imo), but thank you very much for your fast answer ...

This might be a Windows Defender thing - ReBuzz is probably trying to write a config file to your user documents folder which is protected by default to prevent ransomware encrypting the folder

Try turning off controlled folder access and see if that suppresses the error - https://learn.microsoft.com/en-us/defen ... ed-folders (for Step 4, replace the step with: If controlled folder access is turned on, you'll need to turn it off )

Thank you for your hint about the controlled folder access ... but
I use a seperate Virus-Scanner, and installed ReB with Admin-rights, so I think it might be another cause for the problem, I post the Error Message right here, hopefully someone knows why "96" is not valid ...

[HerrFornit]

Here 1782 started withoud errors. As you know I use W11, but I suppose your copied gearfolder is a candidate for the errors.

Well, I don't think the copied gearfolder is the problem, 'cause ReB should overwrite the data therein with it's new files, so that it can work otherwise it would be strange ... (imo), but thank you very much for your fast answer ...

Yes you are right. Just thought that RB scans the machines at start and, if there are some (perhaps new) incompatibilities, that might lead to (sometimes strange) errors.
But for sure, a dev has to have a look at it.

[Buzztler]

Yes you are right. Just thought that RB scans the machines at start and, if there are some (perhaps new) incompatibilities, that might lead to (sometimes strange) errors.
But for sure, a dev has to have a look at it.

O.k., the errormessage is caused by something in the gen's folder, that's for sure, as a result of testing so far, but I could not find out which gen(s) or combination of gen's exactly cause the problem ... this will take a lot of time and perhaps a dev has a faster solution, it would be fine ...

[thepedal]

Feature request:

I was talking with an old buzz-user friend the other night and we thought how cool would it be if you could 'clone with patterns' a set of machines but then switch them to be linked to the original set of machines. The idea being you could take a set of machines and use them as an isolated group for purely controlling stuff (peer env or anything like that) but then param changes to the original set would just be applied to the clones as well. Can't be tricky to implement I'd think but would be a nice feature. That's it

[HerrFornit]

Yes you are right. Just thought that RB scans the machines at start and, if there are some (perhaps new) incompatibilities, that might lead to (sometimes strange) errors.
But for sure, a dev has to have a look at it.

O.k., the errormessage is caused by something in the gen's folder, that's for sure, as a result of testing so far, but I could not find out which gen(s) or combination of gen's exactly cause the problem ... this will take a lot of time and perhaps a dev has a faster solution, it would be fine ...

Hi Buzztler,

I asked and got response at discord by dev lolokimono to your error messages. He had a look at it and said that the first error looks like modern patern editor was not at it's place, or registry config is corrupted.
The "96" error is realy odd.
For now, it would help if you could reproduce the error(s) and reduce the song to one machine (producing the error), at least if it is related to loading/save a song. Another way is to post your list of gens and effects (if they are free to download)

[thepedal]

First off, massive thanks for the latest update! It's mind-blowing to see this project getting some traction and the Buzz legacy living on.

I use basically the same setup for all my tunes with various different instrument machines around the edges. No VSTS, always 6 synths, 1 bass and several drum kits. It all hangs off some mixIO machines and crucially some Peer machines. Almost always PeerChord and most importantly a PeerCTL which I use to map lots of bits to a Behringer controller. This latest update lets me load this setup so that's great. Audio-wise the tunes don't really play without shredding the cpu but there's a couple of more interesting things I've noticed.

1.) Where I have several instances of the same synth, say MX7 all renamed to different things the PeerChord only loads with the first instance of that type of machine assigned. I can go into the settings and assign the others ok and all the note data etc is intact in the patterns. It's as if it doesn't recognise the renamed machine instances as separate from the first one, treating them all as 'MX7' and when it assigns the first one it thinks it's done.

2.) The PeerCTL is loading without almost all of the assignments in place but the tracks are loading the different param/value mapping curves I've set. I say 'almost all' because strangely it is loading two assignments. One for an instance of LarsHa Funkyverb and one for a Jeskola NiNjA dElaY. Nothing special about either of them, in the general layout. There's plenty of machines doing similar stuff in similar positions and other instances of those exact machines for which the assignments don't load. It seems, to use a very over-used word quite random that those assignments are in place but no others. Also, the PeerCTL mixer does seem to let me control params on the NiNjA but not anything on the Funkyverb. Even with de-assigning and re-assigning them.

3.) This may be a well known thing, even without pressing play all the loaded machines seem to be doing a fair bit of CPU and Buzz overall is saying it's at 100% CPU. Windows sees this as 30-40% CPU usage even though I've told Buzz to use all 8 cores. That's using WASAPI and or Asio4all on my laptop with a standard Realtek onboard soundcard. Nothing special but classic Buzz lets me do a whole lot on that.

Happy to share more info and BMX etc if this is useful..

[DJ Saint-Hubert]

First off, massive thanks for the latest update! It's mind-blowing to see this project getting some traction and the Buzz legacy living on.

I use basically the same setup for all my tunes with various different instrument machines around the edges. No VSTS, always 6 synths, 1 bass and several drum kits. It all hangs off some mixIO machines and crucially some Peer machines. Almost always PeerChord and most importantly a PeerCTL which I use to map lots of bits to a Behringer controller. This latest update lets me load this setup so that's great. Audio-wise the tunes don't really play without shredding the cpu but there's a couple of more interesting things I've noticed.

1.) Where I have several instances of the same synth, say MX7 all renamed to different things the PeerChord only loads with the first instance of that type of machine assigned. I can go into the settings and assign the others ok and all the note data etc is intact in the patterns. It's as if it doesn't recognise the renamed machine instances as separate from the first one, treating them all as 'MX7' and when it assigns the first one it thinks it's done.

2.) The PeerCTL is loading without almost all of the assignments in place but the tracks are loading the different param/value mapping curves I've set. I say 'almost all' because strangely it is loading two assignments. One for an instance of LarsHa Funkyverb and one for a Jeskola NiNjA dElaY. Nothing special about either of them, in the general layout. There's plenty of machines doing similar stuff in similar positions and other instances of those exact machines for which the assignments don't load. It seems, to use a very over-used word quite random that those assignments are in place but no others. Also, the PeerCTL mixer does seem to let me control params on the NiNjA but not anything on the Funkyverb. Even with de-assigning and re-assigning them.

3.) This may be a well known thing, even without pressing play all the loaded machines seem to be doing a fair bit of CPU and Buzz overall is saying it's at 100% CPU. Windows sees this as 30-40% CPU usage even though I've told Buzz to use all 8 cores. That's using WASAPI and or Asio4all on my laptop with a standard Realtek onboard soundcard. Nothing special but classic Buzz lets me do a whole lot on that.

Happy to share more info and BMX etc if this is useful..

Re: 3) Make sure to set View ---> Settings... Multithreading to 'True'. But if you have very interconnected machines/effects/spaghetti it might not help

[polac]

Thanks for looking into this, I'd love to get back to Buzzin' again - for the past few years I've leapt to the dark side (Ableton)

To confirm the vst3 loader is being called this many times on just loading up reBuzz to a blank project, unfortunately as reBuzz doesn't appear to write anything to the debug console after the application is fully launched I can't see whether it's being called that many times again when loading a single PVST instance thus causing this >minute delay in adding it to the machine view (perhaps I can run Process Monitor to check...)

Additional info if it's of use: I ditched all my Waves plugins back in early '23 after their back-and-forth with perpetual licenses vs subscription, also their WUP plan made keeping plugins up to date cost more than buying them in the first place - so all the waveshells (think there were about 10 of them as I refused to update just to keep them on the save version number) have been removed from my machines.

Also my laptop is a relatively (when it was purchased about 4 years ago anyway ) well spec'd machine: i9 10900k, nvidia rtx 2080, 32gb ram, ssd drive hosting the OS, Buzz and the VSTs

EDIT: OK so yeah reBuzz is definitely going wild with it's file reading. Just launched Process Monitor and added 1 (one!) VST3 (u-he's Bazille) to an empty project -

Hmm cannot reproduce here. I think I'll add the "old" shell startup scanning as an option, which means no scanning at all on Buzz startup, only on demand shell scanning via preferences. This won't be the default scanning method though.

[polac]

Hello polac,

do you mean with "instances" also the same VST appearing serveral times in a song? Loading the VST only once per song would speed up the song loading, too? (Not sure how it works internally, but I describe it as it looks alike)

Thanx a lot for your support!

Yes it does, because ReBuzz is sandboxing every vst in a song in separate processes. In standard Buzz it does scan only once. The only problem are the shell plugins like vst3shell/waveshell etc which i have to instanciate to check if there are new vsts in the shell.