pcrecord, post: 442405, member: 46460 wrote: This topic should be on it's own thread tho ;)

What would we title it and where would we split it off?

Good question...
It could be : How does 3rd party plugins affect my computer
or : Plugin vs audio degradation, myth or reality ?
or : Why is it better to keep in simple (Software and plugins)
or : More plugins ? How bad could it be ?
.... ;)

In DAW's, Computers & Software seems the right place

Thought I'd start a thread here and give us a place to discuss plug ins...

;)

Just to add context, we discussed plugins in many other thread.
One aspect that was intriguing to many was about how plugins could make the DAW or computer unstable and how a inserted plugin, deactivated or removed could continu to affect the sound.

If any of us have scientific data to support those interogations, that would be the place to put it ;)

audiokid, post: 442402, member: 1 wrote: I would rather keep stability and phase dead locked to how my DAW was designed.

Lol I've been thinking about this non stop. Whatever the cause of the lingering is yet to be determined.

But I do belive the effect is in essence ohase related. I thought I nailed it last night trying to sleep when I thought it might simply be delay compensation that wasn't shut off when the pluggin was removed. But pt7 didn't even have delay compensation!! Gross lol.

But if what I describe is 'modulation' and Chris describes it as noise, and I've heard image collapse in DP, those can all be phase related.

Take a simple noise/hiss just for example. Assume audible noise or cranked channel/Ect, just for the example even though digital noise floor is quite low... If the noise was steady state, and when the pluggin was added it and it altered the sync/phase even a tiny bit then in theory (to my mind) that would make the noise louder. This would be becuase the intrusion caused an upset in the phase or +/- (polarity?) of the steady noises waveform. Similar to how crappy clocking throws off the image depth. If the noise floor itself is so low due to stability/balance in the daw's internal sync/clock. Then it would stand to reason that anything that effected the timing/phase could effect/modulate the track.

Further, Chris used a delay, I used and eq, both of which are time/phase related effects.

So my theory of the what is that due to an imbalance/phase shift/sync/clock type error it caused noise that was otherwise phase canceled, to become louder becuase of a partial shift. Hence less cancelization.

As to why and how the symptom lingered after the effect was removed is still a question.

But that's my best theory on what we experienced.

'Stability' or the merit of the internal clocking can also be a determining factor in why some DaW's have more solid imaging and depth or less smear. Between the clocking and the headroom I belive that's what makes up the DaW's 'sound' overall.

What's even more fascinating to me is the phenomen of this noise/modulation seems to be reguardless of pluggin format.

I would wager that even the 32 bit 2001 version of Adobe audition still sounds more solid/open/clear than some current DaW's. Also apparently the resolution/headroom is effected by the bit being "integer" math or not. I came across this while researching an adda I'm interested in. The post was reguarding a new product by mytek.

----Michal from mytek New York, (quoted from another forum)

"Both are 32 bit integer and the goal is to provide complete signal path in 32 bit integer. (not float- float is 24 bit + 8 mantissa with resolution of only 24 bit)

ADC is dual mono (2 balanced chips sampling at 12 MHz) , real 130dB DR, with 32 bit integer output.

Most importantly there is true integer 32bit USB transmission and that works with Core audio and ASIO , so if you have an app capable of 32 bit (Integer!) I/O, like Reaper for example, in a 64bit float environment you can maintain 32 bit integrity which basically means you increase working headroom by 48 dB (32x6dB=192dB, as opposed to 144 dB with 24 bits). That gives you wider path for digital processing without unnecessary rounding and throwing out detail.

The result is more natural sound with more low level detail, it's similar to going from 24 bit to 16 albeit a smaller effect. Totally audible on a good monitoring system."
-------

Haven't read this article yet but I'm gonna in a few.

http://mytekdigital.com/download_library/papers/Beyond_24_bit_Michal_Jurewicz_Resolution_2014.pdf

kmetal, post: 442424, member: 37533 wrote: Further, Chris used a delay, I used and eq, both of which are time/phase related effects.

Also Reverb in a Pro Tools HD track that was judged by many Mastering Engineers. All were dumbfounded. I was the one that heard it, proved it through a null test. The remainder was reverb. This is another topic but the bad code part is related to this.

audiokid, post: 442432, member: 1 wrote: Also Reverb in a Pro Tools HD track that was judged by many Mastering Engineers. All were dumbfounded. I was the one that heard it, proved it through a null test. The remainder was reverb. This is another topic but the bad code part is related to this.

Interesting another time based effect.

Lol- an awful terrible reverb. Yet gets used still to this day. Beyond me man. Hell I hated the thing first try. Of course that was after swithiching to pt from audition, which has truly excellent reverbs. The new Coldplay record used it on Beyoncé's vocals..... Read it in inside track in sos.

I really wish I understood code better, this really is perplexing.

Between 384k, decoupled master, decoupled amp sim computer/rack, lingering pluggin noise, I'm really re inventing or finding my workflow.

I'm literally gonna have to test each thing one by one.

I'm contemplating having a mix daw that runs no efx. No processing. Just using it for the mixer. In other words printing all the tracks, then importing on a fresh session audio only, for final balances.

Heavily considering sequioa again. Now that I have deducted video software. I may end up using my 3rd party plugins on the decoupled systems only as part of the vsti / amp sounds 'on the way in' to the tracking daw. Basically using a computer / vsti / interface like a standalone synth.

Man this is not easy!

Well dropped about 2k on mix plugins and vsti's only to find an audio interface that goes up to 384k (mytek 8x192) coming out next year. Besides hoping the plugins and vsti catch up soon, I've once again flip flopped DaW's and may gone sequioa.

I've got no science. But in dp7 if you use a certain amount of plugins the imaging collapses in the way an MP3 collapses a .wav it's subtle but once the line gets crossed it doesn't return.

What I can't figure out is if it's a cpu threshold, a stacking effect of the pluggin code, maybe a limit to the 32 bit system I was on, ram caching? Undo history? Summing?

There has got to be an explanation for that.

The whole lingering noise/modulation is a strange artifact as well.

As much as I don't ever want to deal w it, I'll be curious when my stuff is setup if I can make PT or Sam exhibit those phenomena and either mix it down or capture it on vid or something.

Figures I finally get all the vsti an plugs I really liked and wanted, now I'm paranoid they'll either do weird stuff, or won't be at the 384k sample rate very soon.

I really wonder if it's a native vs dsp card thing too.

I'll stop babbling and try and find some facts. I know the guy who coded the URS plugins, I've been meaning to drop him a line, he's usually super busy, and I'm not like a friend of his, but if I get him on the phone I'll try and mention it.

kmetal, post: 442434, member: 37533 wrote: Interesting another time based effect.

Which is why I use hardware processors, and a DAW that has awesome stock, tested essential plug-ins. A DAW that is a hybrid fully awesome beast. Now you are getting haha!! :D

kmetal, post: 442434, member: 37533 wrote: I'm literally gonna have to test each thing one by one.

2 DAW's and an independent monitoring controller = forensic ABC comparisons. This is why I use the same DAW platform for both DAW1 and DAW2.

kmetal, post: 442434, member: 37533 wrote: I'm contemplating having a mix daw that runs no efx.

The mixdown DAW rocks for all related to mastering, including uploading, studying your mix and keeping it there while you may keep mixing others into to it, as you are picking your best mix. .

kmetal, post: 442434, member: 37533 wrote: Man this is not easy!

No its not. Just keep reading and thinking.

http://recording.org/threads/vsts-the-good-the-bad-the-ugly.59726/

kmetal, post: 442435, member: 37533 wrote: I've got no science. But in dp7 if you use a certain amount of plugins the imaging collapses in the way an MP3 collapses a .wav it's subtle but once the line gets crossed it doesn't return.

What I can't figure out is if it's a cpu threshold, a stacking effect of the pluggin code, maybe a limit to the 32 bit system I was on, ram caching? Undo history? Summing?

On my side, with Sonar, if I put too many plugins, I hear pops and glitches. When it happens I just put a bigger buffer value and it comes back to normal. I can vary between 128 when tracking and 1024 when mixing. I haven't worked on more than 40 tracks but I assume that the more processing I ask, more buffer will be required.
I need to test this but I pretty sure that when I hear pops and glitches, they won't show on the final product if I use the export features... So it might just be the playback computing that is failling. I understand many won't use export feartures but record summings instead. This is problematic because it asks for a perfect playback. Even dual DAW workflow can suffer from this. If your computer is not able to playback everything perfectly, recording that signal is not a good Idea. That's a bit why, against many opinions, I prefer sonar exporting fonctions...

384k sample seems overkill for tracking and mixing. Altought I don't even know what mastering engineers will say about it yet. I'm quite content with 96khz right now...
I admire your search for perfection K ;)

audiokid, post: 442436, member: 1 wrote: 2 DAW's and an independent monitoring controller = forensic ABC comparisons. This is why I use the same DAW platform for both DAW1 and DAW2.

Ya know, that's a good point. Didn't think of things that way. But it makes sense. Your taking yet another variable out of the equation. You helped me decide whether to upgrade, or grab another full version (academic pricing) the same $$. I'm gonna grab the second license. This gives me the ability to have SAM on 6 computers. Plus i can still upgrade my current one too, so they'd all be prox3. This for the cost of one prox3 full price ($600).

I think I'm starting to get the hang of things 18 years in. Lol

What's cool about magix is you can open Sam sessions in sequioa. I'm eyeing sequoia to pair w the mytek currently....

pcrecord, post: 442445, member: 46460 wrote: On my side, with Sonar, if I put too many plugins, I hear pops and glitches. When it happens I just put a bigger buffer value and it comes back to normal. I can vary between 128 when tracking and 1024 when mixing. I haven't worked on more than 40 tracks but I assume that the more processing I ask, more buffer will be required.
I need to test this but I pretty sure that when I hear pops and glitches, they won't show on the final product if I use the export features... So it might just be the playback computing that is failling. I understand many won't use export feartures but record summings instead. This is problematic because it asks for a perfect playback. Even dual DAW workflow can suffer from this. If your computer is not able to playback everything perfectly, recording that signal is not a good Idea. That's a bit why, against many opinions, I prefer sonar exporting fonctions...

The clicks and pops from buffer sizes I'm quite familiar w my old core 2 duo PT mpowered system. I belive your correct that they are glitches in the playback engine, and not printed to the audio. In my experience they've never showed up on the wave form or mixdown. Lol still frightening.

Have you noticed a differnce on tone/texture, or imaging, in sonar? That's what I'm baffled by. Based on what you've said where talking similar sized sessions.

pcrecord, post: 442445, member: 46460 wrote: I understand many won't use export feartures but record summings instead. This is problematic because it asks for a perfect playback. Even dual DAW workflow can suffer from this. If your computer is not able to playback everything perfectly, recording that signal is not a good Idea. That's a bit why, against many opinions, I prefer sonar exporting fonctions...

Ya know the potentiall for error in either case is frightening. I'm eager to try both. I think the key to the two daw system would be to hear the error in realtime. Just guessing.

I have compared mixes from the session next to the bounced waveform both in iTunes and bias peak pro (mastering software, now discontinued) and I've experience the full multitrack as fuller, or the bounces as 'smaller'.

This hasn't been the case in every song or daw, mainly ones with a ton of mid stuff/energy and processing like rock and metal....

I've never done the capture daw thing yet so only know bouncing, which in general I found fine, with those occasional exceptions.

pcrecord, post: 442445, member: 46460 wrote: 384k sample seems overkill for tracking and mixing. Altought I don't even know what mastering engineers will say about it yet. I'm quite content with 96khz right now...

96k seems to be the sandard I general at all levels, even the mega studios. I've been at mainly 44.1, up until my hiatus last year. So I'm wondering if I'll perceive any differnce.

pcrecord, post: 442445, member: 46460 wrote: I admire your search for perfection K ;)

Lol there's a line between not being satisfied ever, and being insane. Not sure what side I'm on, lol, let's say I walk the line.

Honestly not sure if I'll hear the diff. But if my software already does 384k and I gotta drop 3k or more for my 'flagship' interface, I figure why not be up to the cutting edge. It seems it's an eventuality that sample rates will keep hiking. So between 'future proofing' the expensive items, and archival purposes it seems to make sense. I'll be super excited to try the Abc of each sample rate for comparison here.!!!!

Gotta hope the vsti players are timely w the 384k or I'll have to do a work around. Ditto for my plugins.

Worse comes to worse, they'll still get plenty of use at the current 192 and below when I'm working on other peoples recordings which likely won't be 192k as standard for a while.

I'm going to throw a further equation into this. Running on a Xenon dual CPU system (Mac Pro of course) or a single CPU i7 PC. I'm going to continue to argue that I personally have found my Macs far more robust under all levels of stress in my experience.

I'm not saying newer dedicated music pcs can't hack a load. I just haven't found that in my experience. I think the amount of worrying about how this or that works VSTi etc is a detraction from creativity.

What we all should look for at this point is, turn on, plug in, play. If you can't do that by now, time to go back to tape and all external gear, no? :).

Tony

Makzimia, post: 442522, member: 48344 wrote: I'm going to throw a further equation into this. Running on a Xenon dual CPU system (Mac Pro of course) or a single CPU i7 PC. I'm going to continue to argue that I personally have found my Macs far more robust under all levels of stress in my experience.

In my experience similarly specd computers performed about the same Mac or PC. I've experienced this pluggin thing on both.

Makzimia, post: 442522, member: 48344 wrote: I just haven't found that in my experience. I think the amount of worrying about how this or that works VSTi etc is a detraction from creativity.

I think it's part of the engineers job to be as knowledgeable and well rounded as possible. My creativity gets stifled when the unexplained pluggin lingering artifact, occurs in the middle of my mix. It harshes the vibe, like wtf just happened to my mix?

So to me knowing the cause can help make better engineering decisions to avoid such oddities. Just imho

kmetal, post: 442476, member: 37533 wrote: Have you noticed a differnce on tone/texture, or imaging, in sonar? That's what I'm baffled by. Based on what you've said where talking similar sized sessions.

I honestly didn't notice anything. But I'm wondering how we could test that.
I guess I'd need a second DAW or external recorder to grab the live playback..
I could play a simple project and record the playback once then add a bunch of plugins on a muted track and record the playback again..
If there is any difference on a null test it would be interesting.
How about that ?

pcrecord, post: 442595, member: 46460 wrote: I honestly didn't notice anything. But I'm wondering how we could test that.
I guess I'd need a second DAW or external recorder to grab the live playback..
I could play a simple project and record the playback once then add a bunch of plugins on a muted track and record the playback again..
If there is any difference on a null test it would be interesting.
How about that ?

That would be interesting. Ideally it would be a capture ala audiokid style. I'm wondering if onstamctiating the plugins would cause then not to null, just by nature, even if there was no error.??

That said instansiating a bunch of plugins and keeping them bypassed would do what I'm talking about (or not). I don't think the plugins have to be active. i know the slate bus compressor adds a db or two even with the settings zeroed out, when the pluggin is active.

kmetal, post: 442610, member: 37533 wrote: That said instansiating a bunch of plugins and keeping them bypassed would do what I'm talking about (or not). I don't think the plugins have to be active. i know the slate bus compressor adds a db or two even with the settings zeroed out, when the pluggin is active.

My Idea is to test if more CPU processing will affect the live playback rendering. I do know that in Sonar, any plugins on an active track (even muted or with the volume down) get calculated even if we don't hear the track. The reason being, if you unmute the track the DAW needs to be ready to deliver.
The only way to prevent a plugin from using ressources is to archive the track or remove it.
I honestly don't know if all DAWs are like that... But it's the behaviour I get with Sonar up to the late versions.

I would not venture in leaving a plugin on an unmuted track at default setting because, like you say, it would certainly not help the null test.
Of course, going analog and back to digital, may give files that never null but if the differences are getting wider with more plugins, it would be a sign that something is going on...

pcrecord, post: 442629, member: 46460 wrote: My Idea is to test if more CPU processing will affect the live playback rendering. I do know that in Sonar, any plugins on an active track (even muted or with the volume down) get calculated even if we don't hear the track. The reason being, if you unmute the track the DAW needs to be ready to deliver.
The only way to prevent a plugin from using ressources is to archive the track or remove it.
I honestly don't know if all DAWs are like that... But it's the behaviour I get with Sonar up to the late versions.

I would not venture in leaving a plugin on an unmuted track at default setting because, like you say, it would certainly not help the null test.
Of course, going analog and back to digital, may give files that never null but if the differences are getting wider with more plugins, it would be a sign that something is going on...

I'm interested in the results of you end up doing any testing.

As far as resources affecting live bouncing, I know my completely unstable old PT7 setup which bounce while playing back, sometimes depending on the weather it wouldn't even bounce with a certain amount of plugins on. It'd stop somwhere in the middle. Soooo annoying lol.

I'll get around to it in a couple of weeks.. I keep you posted ;)

Awsome. Keep your ears open for image collapse and graininess lol. I wonder if sonar does that like DP and PT