Contrast Analysis not Working

[Gale Andrews]

It's also unfortunate that for most of the examples WCAG actually show a selection in a stereo pair, which now won't work and which scenario wasn't considered by them.

So it appears that they were working on the assumption that stereo tracks are fairly similar in both channels?

Maybe, I don't think anyone even considered or raised it.

If that's the case, what was the reason for the "fix"?

Probably because you raised a bug report for it ( http://bugzilla.audacityteam.org/show_bug.cgi?id=454 ). I haven't officially "approved" the fix yet, as you can see, but you can say that about other bugs too because I am short of time.

Martyn did not answer any of my questions about his fix in comment 3 so I guess we'll close that bug, raise new bugs for the other issues that were found in testing the fix, and decide if the loss of analysing the bottom half of an unsplit stereo pair is worth a new bug. I think I would favour a message (when a single stereo pair is selected) "Stereo Pair selected. Contrast can only measure one selected channel at a time. Press "Cancel" then use the Track Drop-Down Menu to "Split Stereo Track", or press OK to analyze the bottom channel only." Or forget the second sentence and the "Cancel" button and just analyze the bottom track when they press OK.

I didn't get any reply when I asked David MacDonald about wrongly stating that you can select track using the TimeText controls in Contrast. However I've now also asked him about the implications of only analyzing half a stereo pair or doing both.



Gale

[steve]

If that's the case, what was the reason for the "fix"?

Probably because you raised a bug report for it ( http://bugzilla.audacityteam.org/show_bug.cgi?id=454 ). I haven't officially "approved" the fix yet, as you can see, but you can say that about other bugs too because I am short of time.

The original bug report was that the "wrong" track was sometimes analysed (the bottom track, regardless of which track is selected).

WCAG 2.0 does not appear to specify what should happen in the case of stereo tracks. Neither is it clear whether rms weighting should be used, though the linked documentation suggests C weighting. The current effect measures unweighted rms.

In the absence of a clearly defined standard I would suggest that the best option would be that, in the case of stereo tracks, the plug-in should measure the root of the average squared sample values (all selected samples).

I have raised bugs 511, 512 and 513 regarding the known problems so Bug 454 can probably be closed now.

[Gale Andrews]

In the absence of a clearly defined standard I would suggest that the best option would be that, in the case of stereo tracks, the plug-in should measure the root of the average squared sample values (all selected samples).

Probably, but if that is possible do we then retain the restriction on only analyzing one track at a time? A selected left and right channel would be the same as a stereo pair. Does it then become meaningful to analyze multiple tracks before mixing them, and what if the user has moved the gain or pan sliders?

I have raised bugs 511, 512 and 513 regarding the known problems so Bug 454 can probably be closed now.

Thanks for that. I resolved-fixed #454 and commented in http://bugzilla.audacityteam.org/show_bug.cgi?id=511 .



Gale

[steve]

Probably, but if that is possible do we then retain the restriction on only analyzing one track at a time? A selected left and right channel would be the same as a stereo pair. Does it then become meaningful to analyze multiple tracks before mixing them, and what if the user has moved the gain or pan sliders?

For a single track the gain slider is not relevant because the pass/fail is only concerned with the relative rms level of the foreground level to the background level. The absolute level is unimportant.

For multiple tracks the situation becomes much more complicated and I see no clear case for what the effect should do. The user may want per track readings or they may want one reading for the mix. I would suggest that if a user wants a reading for a multi-track mix, then they should use Ctrl+shift+M to create a mix track, then select and analyse the mix track (then Ctrl+Z to undo the mix track).