BUG: plugin delay compensation is broken

Nuendo
61

Bingo. Broke PDC.

I have only been testing for 3 minutes, and am up against a deadline - so better to get a first quick report in now, and I will write more details later on (and test far more thoroughly.

My Steps - very similar to the repro offered.

System. Mac Pro 3.1, 16GB Ram, OSX 10.6.7. 1 x Metric Halo 2882 2d expanded interface, clocked internally.
Nuendo 5.1
Sample rate : 48K only. (none others tested)
Opened a new project in nuendo.
Buffers set to 1024.
Imported a mono sine tone. It created a mono track. (Dragged the tone from desktop)
Duplicated the track.
Hit phase flip - everything cancelled.
Here I differed a little on the repro.
Inserted Altiverb (latest) on track one, set it to 100% WET.
As expected, I hear the altiverb (altiverb track still phase reversed!)
Hit Bypass on the pluggin (in the mixer) - and sound cancels as expected.
Buffers set to 768.
Sound still cancelled. Tried turning bypass on/off.
Buffers set to 512.
Sound still cancelled. Tried turning bypass on/off.
Buffers set to 384.
At this stage, altiverb is bypassed, and the track is phase reversed. I get sound. At around -50dBFS.
The sine waves are at -10dBFS, 440Htz.
Tried turning bypass on/off. PDC still broken (sound passing when in bypass.)
Set Buffers to 256.
Sound once again cancels.
Set Buffers BACK to 384.
It seems (conjecture) to be the same signal passing through (when it should be nulled.) It comes up at the same level (-50.1dBFS) on my peak meters inside spectrafoo
Set Buffers to 512.
Sound once again cancels.

I must admit to being quite amazed I have not come across this issue, as I DO quite often change buffers in the middle of projects. However, from habit, and habit only, I usually start a project at 256buffers, and as I start getting problems, I increase them - to 512, then 1024 and finally 2048 for mixing. I do not think I have ever turned on 384 or 768 before now, except perhaps when doing some benchmarking tests.

Thanks for sticking with this guys (Tuckermix - your patience is appreciated!)

I will report more results as I get them - positive or negative.

Cheers, Brendan.

(edit : just tried this on protools 9 on the same machine. Everything exactly the same as nuendo - and could not repro.)

62

We pretty much get the same results on the betaforum; Mac specific and only for buffers which are not half/double value of the “usual ones”. Many soundcards only offer 32,64,128,256,512,1024,2048 … buffers.

And only with some plugins (for the moment only UAD and Altiverb)

Someone seems to remember that this issue with “odd” buffers (384, 768) is confirmed by RME which advises not to use those buffer settings on OSX. But I haven’t found any trace of that.
We investigate further.

Fredo

63

okay, here’s an update:

other plugins that are broken with this repro are:

steinberg compressor, envelope shaper, expander, gate, limiter, reverence
audio ease altiverb 6
sonnox transmod

also, for the record, it’s not broken at the following sample rates:

32k, 64k, 176.4k, 352.8k, 384k

(ie it is broken at 44.1, 48, 88, 96, 11.025, 22.050, 192k)

it’s also worth mentioning that the following waves plugins won’t null (even in bypass) at any buffer setting:
maxxbass, renaissance bass, debreath, h-comp, l1, l2, maxxvolume, mv2, renaissance compressor, renaissance deesser, renaissance axx, renaissance vox, c4, l3-ll

64

Hey Max - interesting.

I did notice that I could break it with LOADS of plugs - but only after I had first broken it with altiverb.

On my tests, I could not get it to break with other plugs after a system / nuendo restart.

I also managed to break it at 44.1, 48 and 96k. I did not try any others. Thanks for such a big test!

Cheers, Brendan.

65

Is this now an official bug?
Perhaps fixed in next release?

Thanks much.

66

Hi,

Yeah, this has been succesfully repro-ed by beta & QA, so it’s an official bug.
I can’t comment on when it’s going to be fixed, but the issue now has an official DevTrack number.

Fredo

67

Glad to hear, thanks for hanging in on this one.

Cheers!

68

I found this conversation 11 years later. I noticed that a latency compensation error while using UAD plugins.

In my case it is the precision desser. I’m working on a movie and using boom and lavalier together is normal. When the dessers are active, phase subtractions are created and the sum of the microphones changes radically in sound. For me the only solution is not to play the desser on the channel but on the master.

Have you solved your problems? There is a solution?

Nuendo 12.0.3 - Monterey 12.4 - mac pro 2019 (3.2 GHz 16-Core Intel Xeon W) - Apogge Ensemble Thunderbolt.