Lossy?

[april20]

Hi!

Need your opinion regarding this nasty dip on Audacity_FA characteristic... EAC's FA is asymptomatic; SA shows that thick dark line... It's a recent (2006) audience recording, received in .wav files on DVDR. Any suggestions?

[saltman]

I moved the thread here. You should get a better response here.

[ameyer17]

Frequency analysis looks like it could be a FM but the spectral analysis doesn't seem to agree with that.
Any chance you could upload a 30 second or so FLAC sample somewhere?

[Five]

could you upload a 10-20 second sample to rapidshare/megaupload or similar ftp replacement site? can't see a whole lot from what you posted.

ps you should set the upper limit of the spectral in audacity to 22050 in the preferences!

[april20]

Here is a sample.

Originally Posted by Five

...
ps you should set the upper limit of the spectral in audacity to 22050 in the preferences!

THANKS, very helpful, now I can see it all! Put a few tracks together (in a correct order), like a continuous recording should go, and now it looks even more suspicious... don't know of what though...

[Five]

you're welcome... why 22050 isn't the default boggles my mind, really

I'm not seeing the dip showing up in the SA like it does in the FA for some reason. also would like to see SA zoomed to 2 seconds & get a hold of a sample to look/listen to if at all possible. doesn't look obiously lossy from what I've seen so far, but the FA is still a little suspicious (nr?).

[april20]

I don’t know what to make out of it, that just under 20kHz peak obviously coincides with the noise or whatever line on SA, and disappears where music stops and audience cheering starts, as if it was separately processed, though I can't hear any transition.

Okay, the sample is here (9.81 MB WAV, can post FLAC if preferred), and the spectrum zoomed is below, the shaded 2 sec area is right where the sample starts.

[april20]

can't edit the above, here is the mentioned SA zoomed

[Five]

for some reason I'm not seeing what I need to see from your screencap... however from the sample it is simple to determine that this is classic atrac (md) which is lossy but still considered tradeable.

would have been better to have it in flac format (same file, half the size) but no matter!

[ameyer17]

The non-zoomed SA almost makes it look like there could be 3 sources there, the first one ending at about 2:00, the second one from 2:00 or so to about 6:00, and the third one from about 6:00 on.

[april20]

Originally Posted by ameyer17

The non-zoomed SA almost makes it look like there could be 3 sources there, the first one ending at about 2:00, the second one from 2:00 or so to about 6:00, and the third one from about 6:00 on.

Yeah, the first part are applause, the second - some talk, and the third is the music, interrupted by short applauses at 11:00, 16:15 and 21:30. And all transitions are very smooth... The question is where those 20kHz music-related noises came from

Thanks Five for MiniDisc reference, that'd at least explain the pre-16kHz dip.

[ameyer17]

Originally Posted by april20

Yeah, the first part are applause, the second - some talk, and the third is the music, interrupted by short applauses at 11:00, 16:15 and 21:30. And all transitions are very smooth... The question is where those 20kHz music-related noises came from

Thanks Five for MiniDisc reference, that'd at least explain the pre-16kHz dip.

Analog transfer through a crappy soundcard, perhaps?

[Five]

yeah for sure it went thru the analog ins of a soundcard (not direct digital) and appears to have been processed (check out how tidy the peaks are) which adds a bit of noise in the highs beyond the usual slight haze. Pretty good job somebody did on it, too. Also worth mentioning, its a near-mono source, but the two channels aren't 100% identical (maybe 95%).

[april20]

Originally Posted by Five

... Pretty good job somebody did on it, too.

So... it should be worth keeping then ... Thanks folks!