analysis repetition loop

[spoo333]

With certain files rganalysis (or some of the backends?) insists on re-analysing them even though they already have replaygain info in the tag. The analysis always completes without any issues, however next time rganalysis re-analyses those files. I've tried re-converting the audio files - always the same result. This is not a real problem, however it slows down the analysis process.

[DarwinAwardWinner]

Can you give me an example file along with the command that you used to run rganalysis so that I can test this bug on my own computer?

[spoo333]

The command I use is rganalysis -m --low-memory -b bs1770gain -j1 -g auto [directory]. I'll upload some files later and email the link to your private address. In one case - a flac album of 13 tracks with track 11 being a silent "gap" track - deleting track 11 solved the issue.

[spoo333]

Not comment but command ;-)

[spoo333]

I've sent you the link to the files

[DarwinAwardWinner]

I'm sorry, I don't seem to have gotten the email. Did you send it to the address listed in my Github profile?

[spoo333]

I've sent it to rct+github@thompsonclan.org now.

[DarwinAwardWinner]

For some reason I'm not receiving your emails. The last one I have is from your address is August 11.

[spoo333]

Can you check again? I've resent it

On 11 Sep 2016 16:07, "Ryan C. Thompson" notifications@github.com wrote:

For some reason I'm not receiving your emails. The last one I have is from your address is August 11.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DarwinAwardWinner/rganalysis/issues/1#issuecomment-246185149, or mute the thread https://github.com/notifications/unsubscribe-auth/AOIKCWgXgJQmg5s0g_rdQ1bS2s0oPXOYks5qpBkigaJpZM4J3EAB .

[DarwinAwardWinner]

Still nothing, I'm afraid.

[spoo333]

That's odd. I've sent it again from another address

On 11 Sep 2016 4:23 p.m., "Ryan C. Thompson" notifications@github.com wrote:

Still nothing, I'm afraid.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DarwinAwardWinner/rganalysis/issues/1#issuecomment-246185947, or mute the thread https://github.com/notifications/unsubscribe-auth/AOIKCR_nqT5J3cimU1uyu4WE0zwcFCj3ks5qpBzegaJpZM4J3EAB .

[spoo333]

Anything by now?

[DarwinAwardWinner]

Ok, I received that one, although I had to fish it out of the spam trap.

[spoo333]

Great, and shame about the spam issue. It's happened before...

On 11 Sep 2016 4:44 p.m., "Ryan C. Thompson" notifications@github.com wrote:

Ok, I received that one, although I had to fish it out of the spam trap.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DarwinAwardWinner/rganalysis/issues/1#issuecomment-246187024, or mute the thread https://github.com/notifications/unsubscribe-auth/AOIKCdCnO2Gkf4e9EgS9nfdExd5Laqebks5qpCHtgaJpZM4J3EAB .

[DarwinAwardWinner]

Ok, I've spotted the problem. When a track is especially quiet, it will get a gain value of exactly zero, which evaluates to false in Python, making my program believe that there is no replaygain tag. So I just need to add code to distinguish between a false value and an empty value.

[spoo333]

:-) great you could spot the issue that fast. I look forward to a future update of the script. Imagine I can remove the shared audio files online?

On 11 Sep 2016 17:38, "Ryan C. Thompson" notifications@github.com wrote:

Ok, I've spotted the problem. When a track is especially quiet, it will get a gain value of exactly zero, which evaluates to false in Python, making my program believe that there is no replaygain tag. So I just need to add code to distinguish between a false value and an empty value.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DarwinAwardWinner/rganalysis/issues/1#issuecomment-246189714, or mute the thread https://github.com/notifications/unsubscribe-auth/AOIKCeE2PkMxg-4u50ypI1dLzyJScIcmks5qpC6TgaJpZM4J3EAB .

[DarwinAwardWinner]

Can you leave them there until I've had a chance to test the fix on all of them? It's always possible there's a second bug lurking in the code.

[spoo333]

Sure. Also, if you need more files to test, I've got three more albums (albeit huge ones in size) that the script currently 'dislikes'. Or you let me know when the new code is ready and I'll run tests on my system. Thanks

On 11 Sep 2016 17:50, "Ryan C. Thompson" notifications@github.com wrote:

Can you leave them there until I've had a chance to test the fix on all of them? It's always possible there's a second bug lurking in the code.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DarwinAwardWinner/rganalysis/issues/1#issuecomment-246190341, or mute the thread https://github.com/notifications/unsubscribe-auth/AOIKCU65lCyvEMD3tr6NyRL_BHMbR1i5ks5qpDFdgaJpZM4J3EAB .

[DarwinAwardWinner]

Fixed by bf9e833 in version 3.4.

[spoo333]

Thank you. All files including the large ones are fine now.