A Question about the sampling rate and range???

[jeronomije]

maybe is a bit weird, but here's the problem....

there's a FLAC file (in this case can be a .WAV or any other lossless format), so how one can determine is file genuine resolution?? For example, that is not just an MP3 convertet to FLAC or WAV!!!

i know that sounds strange, but is there a way/method to do a such thing??? (preferably on PC using some software).

thanks in advance.

[Pano]

Darn good question! Sometimes you can tell by running an FFT spectrum analysis on the file. If you see it cut off at 15Khz, you can suspect that it was once a low bit rate MP3.

But other than that, I don't know. It will be very interesting to see what other folks here come up with.

[jeronomije]

Thanks boy for a such express reply!

It did work, i've used CoolEdit's (2.1) Frequency Analysis, it shows a drastic drop around 15000hz.

Thanks again!!!

[Pano]

You're welcome!

Yes, that looks to me like it was once an MP3, low bitrate.

[OllBoll]

Another nice app which I'm personally using is Audacity, mostly because it's simple and can be found in the Ubuntu repositories =)

It has lots of analytic functions, amongst other a clipping finder to see if the mixer of the track was an enemy of hifi...

But to view bitrate... I don't know if it's possible to magically see that though unless there is some giveaway like the spectrum example

But it should be possible shouldn't it? I mean, there must be some program with some function that can analyze content by sample wandering and search for odd behaviour, possible missing information and such?

[CopperTop]

I imagine you might be able to spot the joins between the 'segments' in MP3 - can't say I know the method inside out, but one of the artefacts I hear in MP3 audio is 'sizzling' on certain content (consistent waveforms mixed with a proportion of random noise) which I think is due to the spectrum reduction being applied to each fresh segment (576 samples according to Wikipedia..?) and a different result being obtained each time. If you could spot 'spectral transitions' on regular 576 sample boundaries then that might be a good clue. (There's something very similar in JPEG-ed images, which show sharp transitions on 8x8 boundaries if you look very closely).

Presumably there are more sophisticated compression methods, such as wavelet, which avoid sharp boundaries, and this applies to images as well as audio.