Phil Leigh;572649 Wrote: > No need to be humble when you are correct! :-) > > The point is it really isn't that hard. > > Say you've got a 24/96 file with significant content between 22-40k > > First, apply a savage high-pass filter and takes out everything below > (say) 22k. > Then apply a constant-time pitch shift to bring the residual content > down to between 1-2k (That's quite a big shift - 5 octaves or so). > > What's left will be pretty obvious - either it's uncorrelated noise or > it's the upper harmonics of instruments (mostly cymbals and other > percussive instruments) in the original recording, which will follow > the "beat" of the music. You can easily see & hear which it is. > If you play the two files side by side it should still sound like > music... > > If you want to experiment, get hold of some DVD-A's... > > You can do all this with Audacity (which is free). > > Another method is to compare a DVD-A with its upsampled redbook > equivalent using AudioDiffMaker - this will null out the common > content, leaving just the extra information on the DVD-A - again, since > this will mostly be above the fringe of peoples hearing you need to > pitch-shift it down to understand it - but you can see straight away on > the screen if it's just noise or something that correlates to the > music. > > > If you are imagining that someone would maliciously or creatively pass > a 16/44 track through a software or hardware "harmonics generator" to > create false correlated info above 22k - yes, that is essentially a > variation on the reverse of the process I've just described. > > That's broadly how an Aphex Aural Exciter used to work - albeit not at > the supersonic frequencies - (Rumours/Fleetwood Mac, anyone?). > > However, if you just add white/pink/otherwise shaped noise to a track > it will be obvious. > > If you have what appear to be correlated harmonics of the music it's > impossible to tell if they were on the original recording "multi-track" > masters or artificially generated afterwards, since that is a legitimate > "effect" used in pop/rock mastering frequently anyway! (I'be never heard > of it being done on classical but that's not my field - I'd be > surprised.) > > This is where common sense comes in. What age is the original master? > What equipment was likely to have been used to capture the original > sound? What effects were available then? > > If you can find a real-world example of a DVD-A, SACD or high-rez > download where you suspect that the recording has been "remastered" by > adding HF noise rather than musically correlated upper harmonics we can > pursue this further...
Well, what makes me think it wouldn't be that hard to cheat you, is that you seem to believe that the difference between high and low resolution file is what's happening above 22 Khz. I'am sorry, but I am convinced that what is above 22 KHz has almost no importance as far as sound quality is concerned. I am quite sure I couldn't hear the difference between a real high resolution 24/96 file, and the exact same file where everything above 22 KHz is suppressed. I still call it high resolution 24/96 file, and I doubt you can easily hear the difference (I mean listening to it, not looking at the spectrum ...) So a high resolution file is certainly not defined as a file with high frequencies (above 22 KHz) content. For me, the real difference between high and low resolution file, is that samples have 24 bits versus 16 bits. That is the real difference as far as sound quality is concerned, and that has nothing to do with high or low frequencies. Transforming a low resolution file in a high resolution one is everything but simply adding information above 22 KHz ... And if I would want to cheat you, I would certainly not use such a simple process than the ones you suggest ... -- nicolas75 ------------------------------------------------------------------------ nicolas75's Profile: http://forums.slimdevices.com/member.php?userid=15823 View this thread: http://forums.slimdevices.com/showthread.php?t=74688 _______________________________________________ Touch mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/touch
