[Traverso-devel] Accuracy of waveform drawing, and performance

[Remon]

Hello Traversers,

I wonder how important it is that the waveform reflect the _exact_ audio data, 
or if it's just enough  that the highest values are accurately displayed.

Traverso 0.40.0 uses polygons to draw the audio waveform, and as some of you 
noticed, enabling painting on a QGLWidget (OpenGL, see settings dialog) 
garbles the output.

Current cvs paints the waveform a little different, the polygon is only 
created from each even pixel, so it skips a pixel at a time.
The result is less calculation of course, and the opengl paint engine seems to 
have less trouble with it.

For a comparison:

http://traverso-daw.org/screenies/screen12.png

Topmost: polygon created with point for each pixel in the waveform
Mid: polygon created with point created for every second pixel in the waveform
Bottom: Traverso 0.30.0 line based painting


http://traverso-daw.org/screenies/screen13.png

Same, but the mid one is painting with OpgenGL enabled.
Please don't mind the fuzzy audioclip info area, I'm aware of that... 


Let me know please what your opinions are about accuracy of the painting, what 
you prefer (line based painting perhaps), and so on.

Thanks!

Remon

P.S.
Up to date binaries can be found here:
http://traverso-daw.org/download/binaries/