Jump to content

I think I came across an interesting bug


OddSox

Recommended Posts

Scenerio:

In one of my projects I direct two stereo tracks into a mono bus. I render audio in non real-time mode. After the render Cakewalk always crashes.

Render it in real time audible mode - it will not crash.

Switch the stereo tracks to mono - it will not crash.

I tried to recreate the scenario in a different project and everything worked just fine. Strange.

Link to comment
Share on other sites

I removed all plugins from both tracks and the bus for the sake of testing. Crashes happened nevertheless.

Thanks for the suggestion about buffer size. I'll have to experiment with it.

For now I've switched everything to stereo and I'm using panipulator to monoize the bus. It works!

Link to comment
Share on other sites

Something similar happens in the export dialog, if you try to export a mono clip with Channel Format "Split Mono" (2022.11 build 021). It results in a message saying:

"An internal error occurred while processing the mixdown. The operation was unsuccessful."

Some may wonder why I use "Split Mono" for a mono clip. But the reason I tried to use this, was that there is a +3 dB volume increase if I use channel format "Mono" only. Someone of the bakers said once that the cause is that the clip export runs thru a stereo bus (double mono).

EDIT: Strange, I just tried with Sonar (23.10 build 014) and there it worked (with the old export dialog)!

Edited by marled
Addition
Link to comment
Share on other sites

47 minutes ago, marled said:

... that the cause is that the clip export runs thru a stereo bus (double mono).

you could set the stereo buss to mono then drop -3db. the merging of stereo into a single mono is typically +3db (just the nature of db math...). i recently "re-discovered" this as i was doing some mono checks and accidentally exported as mono. then importing into Ozone the waveform was higher than expected (doubled effectively) and was purely mono... lol.

Link to comment
Share on other sites

1 hour ago, fossile said:

you could set the stereo buss to mono then drop -3db. the merging of stereo into a single mono is typically +3db (just the nature of db math...). i recently "re-discovered" this as i was doing some mono checks and accidentally exported as mono. then importing into Ozone the waveform was higher than expected (doubled effectively) and was purely mono... lol.

Yes, but the problem is that I wanted to export CLIPS, no tracks or busses. And if I use the export dialog it increases the volume by +3dB. If I drag the clip(s) to a folder, then it works as expected, but that way I cannot give the desired names and manual name changes on a couple of clips is very inconvenient.

Link to comment
Share on other sites

in the export dialog you can pick the export type as clips, i used {clipname} in the example - TRACK001 has CLIP001 and TRACK002 had CLIP002, and then exported. clips came out as named and types. note that the clip name is also in the underlying track lane so it may be possible to select a variety of clips from the lanes to export...

would presume that if i exported the stereo clip as mono i'd still see the +3db -- so maybe this doesn't solve the problem -- not sure the pan law settings would be applied but conceptually you could set it to -3db center and see. 

image.png.7827c0e69e81451255b1a62e023d2d11.png

image.png.846614f22b0709340d225d43e7568d2f.png

  • Thanks 1
Link to comment
Share on other sites

10 hours ago, fossile said:

would presume that if i exported the stereo clip as mono i'd still see the +3db

In my case I have only MONO clips! And they have also the +3dB volume increase if I did it like you described above! ?

That's why I tried to export with "Split Mono", because in Sonar and earlier versions of CbB that was the only way not to have the volume increase (you could just delete half of the exported clips). But with the new export dialog this works no longer, it fails with a message (see above).

Link to comment
Share on other sites

@Noel Borthwick just did an experiment  -- there is definitely a +3db increase when exporting the clips.

  • pan law set to 0db centered
  • created a sine @ 440hz -15db RMS (12-db peak) for 15 seconds .WAV
  • import into CW (drag and drop) -- check levels = -12db
  • export clip = -8.99db (this is the 3db increate)
  • imported the exported clip - as expect now -9db
  • pan law set to -3db sine/cosine
  • imported clip - now -12db
  • now export clip with pan law set to -3db
  • same - +3db added

pan law setting has no impact on the export or import of the actual clip.

source file

                                          
Cursor position (Time)            00:00:00.000        
Sample value at cursor (dB)       -Inf.               
Minimum sample position (Time)    00:00:00.019        
Minimum sample value (dB)         -12.000             
Maximum sample position (Time)    00:00:00.030        
Maximum sample value (dB)         -12.000             
RMS level (dB)                    -15.010             

Average value (dB)                -Inf.               
Zero crossings (Hz)               440.00              
 

exported clip content:

pan law set to 0db center

Cursor position (Time)            00:00:00.000        
Sample value at cursor (dB)       -Inf.               
Minimum sample position (Time)    00:00:00.019        
Minimum sample value (dB)         -8.990              
Maximum sample position (Time)    00:00:00.030        
Maximum sample value (dB)         -8.990              
RMS level (dB)                    -12.000             
Average value (dB)                -Inf.               
Zero crossings (Hz)               440.00              


pan law set to -3db center sine/cosine

Cursor position (Time)            00:00:00.000        
Sample value at cursor (dB)       -Inf.               
Minimum sample position (Time)    00:00:00.019        
Minimum sample value (dB)         -8.990              
Maximum sample position (Time)    00:00:00.030        
Maximum sample value (dB)         -8.990              
RMS level (dB)                    -12.000             
Average value (dB)                -Inf.               
Zero crossings (Hz)               440.00              
 

 

 

 

Edited by fossile
  • Like 2
Link to comment
Share on other sites

@fossile the actual issue isnt related to pan law. When you bounce with source category clips, it's running each selected clip through the entire signal flow and captures the "as heard" sound. However, when it encounters a mono clip with channel format mono (or follow source), the clip gets upconverted to stereo at the end of the bounce signal path and then it gets converted back to mono when writing the file. this is where the 3db gain comes into play.
I've fixed this by compensating for it for the next version.

  • Like 4
  • Thanks 1
Link to comment
Share on other sites

  • 4 weeks later...

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...