Can Someone check me please: MVibrato phasy issues even with Depth = 0.00%

[alexis1]

Hi -

Can someone check me please? Running a vocal through MVibrato sounds as if it may be adding some kind of delay, so that it sounds a bit phasy. (Cubase 7.5.40)

Repro:
1) Route some audio to the master out. I used a vocal with a fair degree of sibilance.
2) Send it at unity gain to an aux channel, with MVibrato inserted on it (Depth = 0.00%), in turn routed to the master out.
3) Listen while toggling MVibrato on/off while looping the audio.
4) Then render the master track twice, once with MVibrato bypassed, once not. Reverse phase on one of the rendered tracks and play both.

What I found on my system:
Step 3: Toggling MVibrato in/out made a definite difference in how things sounded. Especially prominent in the "Ssss" of the vocal track.
Step 4: The vocal tracks didn't null (the backing tracks did, as expected).

Can anyone else reproduce that please?

Thanks much - Alexis

(PS: The application I was trying to use MVibrato for: In the Aux Channel, inserted just before MVibrato, I had a reverb plug in. I was trying to apply a modulated pitch shift to the reverb. Things didn't sound right - with MVibrato inserted/not bypassed, there was noticeable phasing even at Depth = 0.00%; when the Depth was turned up to around 30-40%, the phasing was very obvious. So I wound up simplifying, and getting the results as above).

BTW - this was the free version of MVibrato, 10.02. Thanks!

[_al_]

i'll check after work, if no one else has by then

oops its friday, i'll check tomorrow

[alexis1]

i'll check after work, if no one else has by then

oops its friday, i'll check tomorrow

Thank you!

[DarkStar]

MVibrato introduces latency (1 sample here, with results in 256 samples latency (1 buffer size) -could that be it?

[alexis1]

MVibrato introduces latency (1 sample here, with results in 256 samples latency (1 buffer size) -could that be it?

Yes, I believe so. But I can't imagine why Cubase doesn't compensate!??


PS-what kind of testing allows you to determine 1 sample of latency?

[DarkStar]

Reaper displays the reported latency for each FX (and rounds it up to the next "buffers-worth" of samples).

[alexis1]

Reaper displays the reported latency for each FX (and rounds it up to the next "buffers-worth" of samples).

Ah, thx.

Anybody with Cubase here that could maybe take a moment please to see if it is compensating for the MVibrato's delay? I've never had a problem with other plugs where Cubase wouldn't compensate. Is there a way I can check what delay MVibrato is reporting to Cubase?

Thx -

[alexis1]

Reaper displays the reported latency for each FX (and rounds it up to the next "buffers-worth" of samples).

Ah, thx DarkStar.

Anybody with Cubase here that could maybe take a moment please to see if it is compensating for the MVibrato's delay? I've never had a problem with other plugs where Cubase wouldn't compensate. Is there a way I can check what delay MVibrato is reporting to Cubase?

Thx -

[DarkStar]

Oops, I should have said that the tracks do not null here either (quite why that vanished of my first reply, I do not know). So I can confirm your findings.

Here are some more diagnostics:
-- I put the same loop on two tracks,
-- inverted the phase of one track,
-- add MVibrato to the other AND set its Depth to 0
-- added MMultiAnalyzer to each track and the Master (where the two tracks should be nulled)

And the result? ...

>>> http://i.imgur.com/GfjCytD.png

So the tracks (red and blue) are almost the same (except at the very high frequencies) but the combined audio (yellow) is not null. If I bypass MVibrato (using its [Bypass] button or in the host) then the tracks null perfectly.

I get a similar result in Tracktion 7.

Maybe I'm missing something obvious (not for the first time). Or a Depth of 0% does not mean no effect at all.

[DarkStar]

MTremolo's Depth works OK here - set it to 0% and the tracks null. As does MLimiter with a Dry/Wet of 0%

But I cannot get MMultibandChorus with every Dry/Wet control set to 0 (100% dry, 0% wet) to null.

[MeldaProduction]

Vibrato MUST be used as insert. It works based on a short delay line (otherwise it would need a latency and it would be out of phase anyway), so there will be a variable phase shift if you mix it with the original signal.

[alexis1]

Thank you DarkStar for your amazing analysis, and Mr.Melda for clarifying there is a delay in MVibrato and it needs to be used as an insert.

When I get back I'm going to put it on an aux bus right after a reverb plug-in, so I can mess with the reverb pitch, maybe put a delay on there after that as well. Maybe that sample or two delay won't matter since it's just going on a reverb bus?

Will report back. Thank you again!

[DarkStar]

It is an Insert with my method.

[MeldaProduction]

With a reverb bus it would be just fine. The vibrato delay length is quite small, but in general it is hard to use effects in parallel - it is impossible unless there is absolutely no delay nor phase shift. And most effects has at least some of it I think.

[alexis1]

With a reverb bus it would be just fine. The vibrato delay length is quite small, but in general it is hard to use effects in parallel - it is impossible unless there is absolutely no delay nor phase shift. And most effects has at least some of it I think.

Thank you for that information.

I am confused why Cubase doesn't compensate for the delay. I thought it routinely did that unless Plug-in Delay Compensation was bypassed.