Digital noise after level reduction during playback

[fuhrmi]

Hello Phillipe,

I have a problem with Wavelab version 11 (or 11.1 as well).

The following workflow I have:
When producing voice recordings, I let the playback run and meanwhile -before the cursor is at the respective spot- lower the level of the breathers by -6dB via a keyboard command (sometimes also with 2x key click by a total of -12dB).
Since version 11 Wavelab sometimes generates a digital noise with 0dB at this edited position, which is also visible in the waveform.
I have not yet managed to generate this error specifically to find out the cause.
I also tried replacing all my own presets and preferences with the factory settings, which unfortunately didn’t help either.
The problem appears on different computers.

What can I do?

Best regards,

Andreas

[S-EHansson]

Hi!

with Audio File Window or Montage ?
and what key command ,
with zero-crossings active ?

regards S-EH

[fuhrmi]

Hi Sven-Erik,

with Audio File Window
Key Command “Q”
Yes, zero-crossings is active

best regards, Andreas

[PG1]

@fuhrmi

lower the level of the breathers by -6dB via a keyboard command (sometimes also with 2x key click by a total of -12dB).

Lower the level using what processor or plugin?

[fuhrmi]

Hi Phillipe,

I use the tool “Amplification” integrated with Wavelab used (see picture).

Best regards,

Andreas

[PG1]

@fuhrmi

with 0dB at this edited position, which is also visible in the waveform.

Please show me such a zoomed noise. Even ideally, save the file in 32 bit float, to send me such a result.

@fuhrmi

I let the playback run and meanwhile -before the cursor is at the respective spot-

Do you mean, you lower a time selection that is just ahead of the playback cursor?

[fuhrmi]

@PG1

Do you mean, you lower a time selection that is just ahead of the playback cursor?

Yes, but not only directly in front of it, but often 10 to 60 seconds before it

As soon as the error appears again, I will send you a short file in 32bit float.

[fuhrmi]

Here is a picture with a zoomed noise:

and here after normalized the noise:

and finally after normalized - zoomed:

[fuhrmi]

It actually seems that the phenomenon has changed slightly between 11.0 and 11.1:

Since 11.1, the noise is not always visible in the waveform. Also, the noise disappears from the processed area once the file has been saved.
However, if a short silence with a fade of, for example, 10ms in length was inserted in this area (via silence generator), the noise becomes visible in the area of the 10ms fades and remains even after saving.

[PG1]

@fuhrmi

the noise disappears from the processed area once the file has been saved.

Are you really sure? With “Save” or “Save As” ?

[fuhrmi]

Yes, with both “save” options.
But only if the noise is not visible in the waveform.
If it is visible, it remains even after saving.

[PG1]

Do you have standard file paths (no network) for these 3 paths? This is where the processed samples are stored during editing. And BTW, what is the typical time selection duration when this happens (the noise duration)

[fuhrmi]

@PG1

Do you have standard file paths (no network) for these 3 paths? This is where the processed samples are stored during editing

Yes, there are no network paths:
#1: C:\temp
#2: C:\Users\Andreas\AppData\Local\Temp
#3: E:\ temp

(E:\ is a normal internal SSD, no network)

The typical time selection is between ca. 300 and 700 ms.
In the 11.0 version, the noise was then also just as long.
However, I believe that since the 11.1 version, I have only had shorter noise, around 10 ms.

[PG1]

@fuhrmi

Yes, there are no network paths:

All normal stuff. I am scratching my head, as I can’t reproduce.
If you happen to save a noisy file, please send it to me (the noisy part and its surrounding is enough). Mayve the sample analysis would give me some hints.

[fuhrmi]

You can download the short files here:

https://we.tl/t-c7qlHPwUch

[PG1]

Real noise indeed. Is the noise region equal to the full process region, or only to part of it?

[fuhrmi]

In version 11.0 it was exactly always in the length of the process region.
In version 11.1 it is sometimes only at the beginning of the process region.

[S-EHansson]

Have to ask

Do you find this digital noise in older version as well ?

Thanks!

regards S-EH

[PG1]

At which approximate frequency do you estimate to encounter the problem? 1 every 10 processes? 1 every 100 processes?..

Have you kept a WaveLab 10 copy on the same system, to see if this happens too?

[fuhrmi]

@S-EHansson

Do you find this digital noise in older version as well ?

No, the problem appeared only from version 11.0.
Version 10 ran fine