What is the High end cut off for a 320 CBR mp3 Options

[oldtimer5]

I am just curious as to what the ball park high frequency cut off is for a generic mp3. I am not asking for a very tech answer. I am guessing the cut off would be about 20-22 kHz but I was never able to find even a soft number for any cut off value. I am assuming that is the reverse logic of the process. The actual cut off is 320 BR but does anyone have any clue as to what that translates to in terms of a high end cut off? I realize music complexity is the driving factor as to where the cutoff is.

Thanks for taking time to answer such a useless question. I am sure this is the only place where I could hope to get an answer to a question like this.

[Tahnru]

I don't have an exact figure, but I'm guessing the lowpass for lame at 320kbps is 20+khz. Based on this table: http://wiki.hydrogenaudio.org/index.php?ti...cal_information

But the lowpass will depend on which encoder you're using, which version of it, and which settings you're tossing at it.

[lvqcl]

For VBR modes, see here: http://wiki.hydrogenaudio.org/index.php?ti...cal_information
('high frequency cutoff' == 'lowpass')

[pdq]

If you are an "old timer" like myself then the cutoff is likely to be well above what we are able to hear.

[DVDdoug]

The actual cut off is 320 BR but does anyone have any clue as to what that translates to in terms of a high end cut off?

With lossy compression, I don't believe there is any theoretical or mathematical connection between bitrate and cutoff frequency. It's just one of the trade-offs. I think you could have a high cutoff frequncy with a very low bitrate. (It would still sound lousy.)

As an analogy, I can write "40kHz sine wave, 0dB, 1 hour", and those few bytes contain all of the information you need to perfectly re-create the signal. It's not really a question of frequency, it's a limitation of total information rate.

[kornchild2002]

I thought it was 20KHz and it looks like I am right:


I encoded the above file (which contains information all the way up to 22KHz) at 320kbps CBR with Lame 3.98.2. A lowpass cutoff point can be seen at 20KHz. As other have stated, make sure that you conduct blind ABX tests before worrying about this (I am sure you already know this).

[WonderSlug]

If one does a command line encode of a WAV file at 320kbps CBR with LAME 3.98.2, it can be seen in the output that LAME states it is using ~20KHz as the target for its polyphase lowpass filter.

[oldtimer5]

Yes, I am well aware of the over kill. The question was just one of curiosity. If I thought that info was critical I would have found out before and not after I ripped all my CDs. By the time you can afford a good stereo you can't hear 18kHz. I doubt that I can hear 16kHz if it was loud and isolated.

I do wish to take time to thank you all for such a thorough set of answers. You even asked my unanswered question. I can see the range of the cut off. There really isn't much range. The process is far more cut and dry than I expected.

This post has been edited by oldtimer5: Apr 24 2009, 13:20

[2Bdecided]

Of even more interest is the abrupt change in behaviour above 16kHz with mp3. That's well known and understood, but it's interesting how clear it is in that graph.

You can, of course, change Lame's low pass filter cut-off frequency, and transition band width. You shouldn't, but you can.

Cheers,
David.

[Martel]

I think that with recent releases of LAME, even if you set the cutoff to 20+kHz, the psychoacoustic filter can still drop everything above 16kHz. So the lowpass information in console is quite useless and this switch usually cannot be used to broaden the encoded spectrum much.

[[JAZ]]

I think that with recent releases of LAME, even if you set the cutoff to 20+kHz, the psychoacoustic filter can still drop everything above 16kHz. So the lowpass information in console is quite useless and this switch usually cannot be used to broaden the encoded spectrum much.

Because the switch is not intended to broaden the range, but to reduce it. if a filter is not applied, the psychoacoustic has to dedice what to keep. if the filter is applied, then some content is already discarded.