Rack stutter audio-noise Loading a Patch

[unlessgames]

If the existence of the noise depends on whether or not the mute module is present between the mixer and the audio-8, wouldn’t that mean, that the source of the problem is module/patch/rack related?

@gtilde is there anything else that you’ve changed besides adding the mute? If you try to remove it, do you consistently get the same noise problem and if you add it, does it solve the issue for good?

My only idea at probing is what I’ve suggested before, trying to replace the Mixer with another cpu hungry module that supposed to output silence. I don’t mean this as “use another mixer to solve the issue” but rather to see if cpu heaviness at the end of the chain has anything to do with it. This could be a completely useless endeavour though.

[gtilde]

@marc_boule I stand corrected-it IS possible. I placed the scope on the mixer outputs and can show the noise at load. I have a screenshot software I can load and, with fingers fast and configured correctly, I captured the scope. I will share all of this in the morning.

[gtilde]

@unlessgames That is all I changed was adding the big mute button. The noise stopped during load engaged, not engaged, it came back. Sorry if i confused - regarding my 1st screen shot, abridged representation, I moved a couple modules out of the way for that screenshot - vs the larger full shot above. I will do your test in the morning - adding it, removing it, adding it and report back.

As for the your second point. I understand. It is worth exploring and I can also test this in the morning too…

@unlessgames your question to @Squinky, ---- good question.

[unlessgames]

Might be overthinking this but an alternative to the scope + fast screenshot method could be to use Simpliciter with PRE-RECORD enabled, and a constant high voltage patched into the REC input

[gtilde]

@unlessgames even if its overthinking,which imo it was not, I learned something new which is always cool-thank you for that. It took me a bit to get Simpliciter figure out, but I captured an event.

the audio file does not represent what I heard, but the anomaly is clearly there. I am attaching the audio file, the screenshots of fastest fingers on the east coast scope results too. Good think I play keys and the fingers still work.

I have six different shots showing: using the ASIO driver and then the WASAPI driver - both had the noise. The ASIO was a bit more dramatic aurally, but the scope looked the same. The WASAPI sounded like a strong ugly tic. Also worth noting, I had to repeatedly load to finally get a fail - with supports what I said before it didn’t happen all the time. I loaded sometimes 3 to 4 times before a fail. I tested on three different patches, in various configurations of that patch.

scope screenshots:

• pic1 patch w/asio drivers (why does “impromptu clocked” click on the load?)
• pic2 new patch w/asio drivers
• pic3 w/asio drivers / hey cool, both channels
• pic4 w/wasapi drivers
• pic4a w/asapi drivers (same patch)
• pic5 captured before the video finished rendering thus the vacant graphics. notice the glowing of the right ch, kinda cool - not! notice the meters??

I also attached the audio file created by the simpliciter module, result from the load. For fun, I took a screenshot of it in RX which shows the entire recording from simpliciter

pic1

pic2

pic3

pic4

pic4a

pic5

oops cannot upload wav file, but here is the rx picture of the audio file, what I hear is much more dramatic . . . i can assure you, its not a “tic” i am hearing. this pic mimics what the Scope shows in VCV

for complete disclosure, here is how the audio file was created:

ideas?

my conclusions: mindmeld is being used in all the testing. mute obvs does not equal mute for what is being passed through - the big mute button - yes, it mutes and stops it - what is different? why does “clocked” click on the load.

[ mindmeld: DC blocker enabled, Clipping default, not running in Eco mode, (doesn’t matter if engaged) ]

The testing has disclosed more and explained a bit more what has been happening during the load - a lot. Interesting for sure … seeing the meters on the mixer light up - when nothing is to be running - but they most certainly are running - so using the big mute button keeps it from passing on through which is great. We still have no idea what is actually causing such a disturbance ?? drivers seem to be not the issue since i displayed switching was no help. I am not thinking either, honestly, its a big oversized load at startup. To me, there are way to many things starting up - when they shouldn’t be starting.

I am going to use the big mute button as a work around. I am going to test today - i think i can - swap mixers and see what this brings.

[marc_boule (Marc Boulé)]

Thanks for the detailed testing! I will have to look at this later, but if ever you can try with the DC blocker in MixMaster turned off, that would be very interesting for me. Incidentally, we placed the DC blocker after the mute in the signal flow, so even when muted, if ever it’s misbehaving, it will be audible.

[marc_boule (Marc Boulé)]

I think I should go hide in a corner , I’m looking at the source code, and it looks like I forgot to reset the DC-Blocker filters upon initialization/startup. If ever you would be willing to try a new version, please let me know what OS you use, and I’ll put up a build for you. I would suggest anyone following this to not spend any more time on it until @gtilde can try a new version of MixMaster with this bugfix.

[unlessgames]

It’s still the best mixer around even if it hails the start of Rack with loud enthusiasm! Thanks for looking into it, and hats off to @gtilde for the report!

[gtilde]

@marc_boule I will be able to check this out this afternoon; I’ll report back soon there after. I concur with @unlessgames, and also, thank you.

I’ll report back results.

[gtilde]

@marc_boule I spent about . . . wait, let me turn the room light back on and allow you to come out from the corner …

I tested about an hour loading random patches, first with just the ones with the scoop attached which showed much desired flat-lined. Then, against my better judgement, others with headphones on, and main levels up. . . happily, with no issues. With the Mute enabled, dc blocker I always turn on, hey - it did in fact mute! As far as I can tell, the mute and DC blocker is working!!

Awesome. @marc_boule thank you for addressing so quickly. Quick question though, why would DC blocker ever be an option? I cannot see this never not enabled - imho. I think it is defaulted to off too btw…I may be wrong.

From my testing, I say push this version out to one and all. This is terribly exciting. I, I am not kidding, the relief knowing that I don’t have to muck about with my mains level with every load… or if I forget I get a burst of audio-noise that scares the every living crap out of me . . . . yay! My Genelecs thank you @marc_boule

I am very curious though, is it “normal” standard behavior for during an initial load process crap blurts out?? Is there a plethora of reasons why this does happen? Load/system/weather?

I still like to find out why the impromtu clock starts/activates on load and sends a clk, when obviously it is not enabled. With this Bang, it initiates what every is connected - seems odd to me.

TBH, I wish, the audio globally, had an on/off option, the “run” engine is in a pause/off state. All modules / deactivated at load. Similarly the way MAXmsp sorts it out having to ‘start audio’ to get things moving. Having Rack at load ALL be inactive, until “rack-engine” becomes “turned on…” manually … just jabbering now

[marc_boule (Marc Boulé)]

Great news! Glad this was sorted out. I’m sorry it caused you so much lost time and thank you so much for the detailed debugging

For Clocked, if you turn off the option in the right click menu called “Outputs reset high when not running”, it should start up with a low clock signal, which may help for you case. There are many use cases and possible implementations for clocks, so this may seem weird, but in some cases it’s better like this. In any case, there are many options in Clocked’s menus, so it’s good to take a look in there if there’s a behavior you want changed.

For DC blocker, agreed that it could probably be on by default since it’s not a big thing to compute anyways.

Cheers and thanks again!

[gtilde]

you are very welcome @marc_boule I was happy to provide all the detailed debugging and testing efforts. The important thing here is we able to figured out the problem and you were able to address it!

Thank you too regarding the Clocked hint and explanation. I took last evening and started to review and study all the documentation you provide for the massive impromtu library - I have so much to learn but appreciate the challenge. I will make that adjustment you suggested that will help in this case and be mindful of it moving forward.

Cheers

@unlessgames I still am curious why this happens in the first place. Is it a systemic issue that can crop up time-to-time of Rack, is it my misuse of modules, my system performance . . . ? should I be just glad it now won’t be "“heard” and roll - with - it ?

[gtilde]

@marc_boule almost forgot - could be no big deal but - your release preview build link text is v1.1.12, but the zip file is 1.1.11. old guys like me get confused fast seeing this . . .

[unlessgames]

@gtilde

Is it a systemic issue that can crop up time-to-time of Rack

Not sure what you mean. My assumptions about the source of the noise turned out to be false, it was a bug in MixMaster.

Are you saying the noise is still happening despite the fix?

[marc_boule (Marc Boulé)]

@gtilde

your release preview build link text is v1.1.12, but the zip file is 1.1.11.

That is intentional, so you don’t get the red update dot which would overwrite the temporary version you manually installed. But as soon as the 1.1.12 hits the library, it will update it automatically and you’ll be automatically back in sync with the library

[gtilde]

@unlessgames sorry for the delay I actually had to work today and I wanted to squeeze in a test or two before I answered fully. The short of it, the answer is yes absolutely and also no, not quiet. It’s late for me and I need to gather my thoughts so not to speak out my arse. I’ll get back to this tomorrow morning. I felt I needed to at least clarify a little now, but tomorrow it will become clearer-we all freakin hope right? cheers

[unlessgames]

No need to apologize for spending your time on other things! I don’t work here either.

Your answer kinda confused me tho, maybe because you answered both yes and no, or because you’ve used the phrase “not quiet” when talking about stuff related to something being audible or not

I assume you still have some sort of noise burst on startup.

I also experience similar stuff sometimes, when loading up more complex patches that are supposed to be making “full sound” right from the start. I don’t know if this is something related to my setup, the modules I use, or is a general issue that most people just don’t care much about.

Reflecting on your first post, this behaviour might be a crucial difference between Rack and other music software. Most other software is not trying to make sound automatically as soon as possible (while still loading up graphics etc.), usually they have a master play or something that you have to toggle to start hearing things. Rack has engine off, or you can mute your mixer or whatever every time you quit a patch, but I don’t like this solution, when I want to quit a program, I want to do it asap.

Maybe the protection of speakers and sanity might be a good justification to put that “space bar toggles everything” feature into Rack that I’ve seen many people ask for…

Or have a work around module made specifically for this, that is basically a mute that waits a bit before enabling sound… this feels wrong too

But I’m rushing ahead here, maybe it has to do with something else…

[gtilde]

Your answer kinda confused me tho, maybe because you answered both yes and no, or because you’ve used the phrase “not quiet” when talking about stuff related to something being audible or not

…arrrrrg true, that is confusing and the wrong spelling - I was tired and banging the answer out on my phone which is never a grand idea. Isn’t it amazing how one letter out of place changes its entire meaning. sorry yes meant to say [not quite] / not yet

Your assumption is correct I still have a noise burst on startup - granted it is a lot more “pleasant.” This is still at the root of it, which ultimately revealed mindmeld’s DC-block issue that since had been corrected. Ty @marc_boule!!! It’s like, I thought, I had one problem somewhere, but it turned out there are multiple things going on here.

You are absolutely right, regarding your comments reflecting on my first post. That is at the crux of it isn’t it? I am with you, if I wish to quit (aha, I spelled it correctly too) a program I want to quit it and not need to do things manually for next time start up - there needs better safe-guards overall imo.

I don’t believe that you are rushing ahead, from my analysis and testing yesterday, it is disclosing another issue. In my case, in laymen’s terms, one thing is banging another thing once, thus causing audio burst for no good known reason. It’s the “premature downbeat” of the patch

I would need to deactivate, assuming it really does this to the module, each module that I might suspect initiates a LoadBang (sorry maxmsp term) at load which initiates the sound source module to produce a sound. Further what is nuts, imho, it’s just one pulse/gate/cv amount …a module, on load - pulses a cv value out.

I’m digging the idea of a special module to just stop this on load until all is loaded including video graphics, or the “press space bar I’m all loaded let’s get started” action! I totally agree…I vote yes!

…so this is what I meant above by saying, so I should then expect it’s just what vcv rack does.

Using the previous really cool let’s pre-record at load module simplicitier - I received these results. It’s screenshots of the audio burst. The 2nd one, has the sound source going thru a delay … as it shows … nuts!

Fun fact: I loaded/unloaded this 10 times and got the exact same results for both patches. Obviously mindmeld mains-out were not muted. ( heyho - when muted, the result is as expected - mutes…yayyy )

…as you can see, it’s a much more pleasant but extremely bad for the heart if not expecting at loud levels - pro tip: I always expect it now

[unlessgames]

lol, love the delay approach, you could even make music from this

I was rushing ahead because this might not be something that everyone experiences, it could still be an issue stemming from audio/video drivers, general audio setup etc.

Until (if ever) this problem is solved properly, I’ve made a module that mutes the signals on launch for x seconds then applies a one second fade-in to avoid popping as well, so you only need to put it between the Audio-8 and your mixer, no other action is needed.

Keep in mind that this is nothing more than a work-around. In exchange for it, you could make an issue on Rack’s github page reporting this problem! If you do so, link it here so I and others can stand behind it.

For now, you can find the module inside my dev-builds (it is called pre-muter). Tell me if the 5 seconds max is not enough…

[unlessgames]

in case someone installed my dev-build before I’ve made this comment, I’ve changed the versioning from now on to use the technique explained by Marc

thanks @marc_boule!