PDA

View Full Version : Lossy or lossless?


RaelEarl
2007-10-25, 09:49 AM
Hi,
here's another one I'm unsure of. The beginning of the track looks like mp3 reconverted to wav, right? But the rest makes me think could it be OK?

I got this file along with the others in .ape-format which is said to be lossless.

I've uploaded a sample track here (http://amor.rz.hu-berlin.de/~heckersj/camel970319%20105.ape) .

Cheers

Five
2007-10-25, 10:49 AM
EAC is not reliable for spotting lossy sources (see here (http://www.thetradersden.org/forums/showthread.php?t=4288)) but that having been said doesn't look suspicious. if you can post a rapidshare link here I can take a look & listen on my end.

RaelEarl
2007-10-25, 03:11 PM
Didn't quite get the clue on EAC from the thread you linked to. I'm still not familiar with SoundForge functions for the same purposes...

Here is the rapidshare link:

http://rapidshare.com/files/65144876/camel970319_105.ape

Thank you!!

Five
2007-10-27, 06:37 PM
not certain about this one... I would not call it certainly lossy at this point but there's something very very strange going on at the higher frequencies :hmm: but then there's also noise up to the top (check the b&w pic) :hmm: :hmm:

(maybe) high quality lossy>analog>digital??

3rd screencap is from close to the beginning... very peculiar

would love to hear more opinions on this...

SundayDriver
2007-10-27, 08:03 PM
I love a good lossy or lossless thread. It's like trying to piece together a crime scene.

Now, I think this one looks lossless. But the high frequencies from 18kHz on up have been compensated some how. The audio has clearly seen some processing, as the levels have been boosted to within 0db, so I think the upper frequencies have been flattened as a result of this. That's the way it looks, but could that be possible? :hmm: I know hard limiting will do that.

That's my two cents or more like my "best guess".

Five
2007-10-27, 11:05 PM
hey thanks for the input

but I don't think its limiting... I just tried it out, cep hard limiting on the track in question, boosting it 20db with the max amplitude limited to -0.2. crushed the shit out of it. checked the crushed version's spectral, it has the exact same frequency cutoff.

then I tried the same limiter settings on another track that has 'normal' frequency response, and its the same, no effect on the frequency response at all.

I looked at some of the peaks they don't look crushed so I think normalization has been applied here or perhaps limiting as you suggested. but that's not causing the strange cutoff so far as I can tell. still confused about this one :(

paddington
2007-10-27, 11:22 PM
I don't see or hear anything wrong with this other than the flyback noise @<hidden> 16kHz. Isn't that stuff at the beginning some sort of synth noise / sfx?

SundayDriver
2007-10-27, 11:40 PM
I get it Five, makes sense- thanks for testing. But due to the lack of dyamnic range, I would lean more towards limiting.

Don't know what is causing the weirdness above 18 kHz though either.