[creyc]

I see what you're saying, and yes you may be hearing "something", but don't jump to conclusions that it's automatically 25k. That I highly doubt. Most likely what you and the other head-fi'ers are hearing is noise. As you push the limits of your source some funky things can happen.

If you can hear 25k, and you're POSITIVE of it, at what frequency can you no longer hear?

(Oh btw I did read the thread, I just find it a bit..messy. People often don't know what they're hearing, just that they hear "something" be it noise or the dishwasher in the other room. You know what a sine wave sounds like though.)

And just for the sake of this thread I'm hearing 21.9k. Oddly when I go above this the soundcard displays that it's no longer outputting anything, which seems to be in consensus with my perceptions.

 

[amartignano]

Boys... those mp3s are the worst thing you can do for a proper test...

Please listen to me:

[size=x-large]They are mp3 (!!!) at 56kbs (!!!!!!!!!) at 32 kHz (!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!).[/size]

32 kHz of sampling frequency means that the last frequency such a file can (in theory, very theory) store is 32/2 = 16000 Hz. It's simply the sampling theory: the max frequency you can store is half the frequency sampling.

A proper test will be with uncompressed PCM file at (at least) cd quality (i.e. 44100 kHz frequency clock = 22050 kHz freq max).

Please don't post digital audio test if you don't know what digital audio is.

bye
Andrew

 

[tourmaline]

you're right. I even don't like mp3 for listening, let alone a hearing test! Quote:

Originally Posted by amartignano Boys... those mp3s are the worst thing you can do for a proper test... Please listen to me: [size=x-large]They are mp3 (!!!) at 56kbs (!!!!!!!!!) at 32 kHz (!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!).[/size] 32 kHz of sampling frequency means that the last frequency such a file can (in theory, very theory) store is 32/2 = 16000 Hz. It's simply the sampling theory: the max frequency you can store is half the frequency sampling. A proper test will be with uncompressed PCM file at (at least) cd quality (i.e. 44100 kHz frequency clock = 22050 kHz freq max). Please don't post digital audio test if you don't know what digital audio is. bye Andrew

 

[hozo]

Quote:

Originally Posted by amartignano Boys... those mp3s are the worst thing you can do for a proper test... Please listen to me: [size=x-large]They are mp3 (!!!) at 56kbs (!!!!!!!!!) at 32 kHz (!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!).[/size] 32 kHz of sampling frequency means that the last frequency such a file can (in theory, very theory) store is 32/2 = 16000 Hz. It's simply the sampling theory: the max frequency you can store is half the frequency sampling. A proper test will be with uncompressed PCM file at (at least) cd quality (i.e. 44100 kHz frequency clock = 22050 kHz freq max). Please don't post digital audio test if you don't know what digital audio is. bye Andrew

Oh, wow. Guess I should read the thread myself before posting. Don't worry about comb filter effects, soundcard noise floor, etc. It won't matter if you're feeding it through an mp3 encoder. The encoder will cut off after at about 16kHz. It'll introduce ringing which will make even "16kHz" audible. And that's not to mention aliases from the downsampling (though the encoder might filter the higher freqs to prevent).

 

[Febs]

Quote:

Originally Posted by amartignano 32 kHz of sampling frequency means that the last frequency such a file can (in theory, very theory) store is 32/2 = 16000 Hz. It's simply the sampling theory: the max frequency you can store is half the frequency sampling.

Not only that, but I am virtually certain that when using encoding rates as low as 56kbps, LAME employs some heavy bandpass filtering, so it's likely that the lowpass filter is cutting off frequencies even lower than 16kHz.

 

[sxr71]

Quote:

Originally Posted by hozo Oh, wow. Guess I should read the thread myself before posting. Don't worry about comb filter effects, soundcard noise floor, etc. It won't matter if you're feeding it through an mp3 encoder. The encoder will cut off after at about 16kHz. It'll introduce ringing which will make even "16kHz" audible. And that's not to mention aliases from the downsampling (though the encoder might filter the higher freqs to prevent).

Well the same thing happens using Audacity's built in tone generator which generates it as a .wav.

 

[SgtE]

hmm, i can hear them all except 16000, thats really strange...

 

[Lisa]

Quote:

Originally Posted by SgtE hmm, i can hear them all except 16000, thats really strange...

I think you're the fourth person to say this. There is something strange about that 16khz file.

 

[amartignano]

Quote:

Originally Posted by SgtE hmm, i can hear them all except 16000, thats really strange...

The file are sampled at 32kHz (not mentioning the very low quality compression factor of the mp3), so the max frequency the files can record is 32/2= 16 kHz. For upper frequencies you hear aliasing positioned at frequency lower than 16 kHz.

bye
Andrew