View Issue Details
| ID | Project | Category | View Status | Date Submitted | Last Update |
|---|---|---|---|---|---|
| 0005527 | MMW v4 | Conversion/Leveling | public | 2009-04-21 19:26 | 2009-05-08 17:45 |
| Reporter | rusty | Assigned To | |||
| Priority | urgent | Severity | major | Reproducibility | always |
| Status | closed | Resolution | fixed | ||
| Product Version | 3.1 | ||||
| Fixed in Version | 3.1 | ||||
| Summary | 0005527: Automatic volume analysis re-analyzes podcasts with each new episode | ||||
| Description | If 'automatically analyze volume of unanalyzed tracks' is enabled, then each time a new episode downloads for a podcast, MM re-analyzes all of the podcast episodes (presumably in order to calculate Album Leveling coefficient). This is a waste of cpu cycles. Isn't there a way of calculating album gain without re-analyzing all of the tracks individually? | ||||
| Tags | No tags attached. | ||||
| Fixed in build | 1239 | ||||
|
|
I guess that album gain should not be auto-analyzed for podcasts at all. If a user wants album gain for podcasts to be analyzed (I doubt) then he can do so manually. |
|
|
I agree--album gain isn't really required for podcasts (though I still don't understand the mechanics of why everything must be recalculated in order to calculate album gain). |
|
|
I have not studied how exactly calculating of album gain works, but from what I see from our code there is a common album handle when processing each track via Gain.dll. So it seems that it would be a high risk to do anything about it for MM 3.1. In the Gain.dll it uses Root Mean Square (RMS) algorithm. It implies that finally there needs to be square root for the sum (x1^2+x2^2+x3^2...) and therefore all samples of all tracks needs to be read? Details here: http://en.wikipedia.org/wiki/Root_mean_square Nevertheless the original issue is fixed in build 1239. i.e. Album gain is not auto-analyzed for podcasts. |
|
|
Verified 1244 |
