I'm not very familiar with mp3 conversion, but sometimes 320 kbps is cut off at 16 kHZ and other times the spectrum goes up to 21 kHz. Can anyone tell me what converter or converter settings I would need to use to convert a WAV to 320 kbps mp3 with the spectrum going up to the 20 kHz levels? When I use EAC and Lame 320 I get the cut off at 16. Any help is appreciated.

320 example 1 (cut off at 16):

(IMG:https://i.ibb.co/1dbSLPM/01.jpg)

320 example 2 (cut off above 20):

(IMG:https://i.ibb.co/GFPRnBs/02.jpg)

https://www.dbpoweramp.com/dmc.htm

Thanks! I downloaded this program and converted a track and I'm still getting a hard cut off at 16 kHz.

Here are the settings I used:

(IMG:https://i.ibb.co/mNBKzxR/03.jpg)

Here is the .wav:

(IMG:https://i.ibb.co/vmgZKZx/04.jpg)

Here is the converted 320 kbps mp3 using the settings in the first screenshot:

(IMG:https://i.ibb.co/9rJ3VWZ/05.jpg)

Any idea how I can retain the higher frequencies using this program?

Is the mp3 format so important? I think for personal use AAC 320 is mush better. With the same mp3 320 file size you don't have this spectrum cutoff. You can use iTunes for conversion.

legionz,
First of all, I'd take a look at the spectrum of the original WAV file. The frequency cutoff at 16 kHz is typical for the bitrate 192 kbps, definitely not 320. An encoder program cannot execute two different algorithms unless you make changes to its work.

Can you upload the original wav file please?
So everyone can try to encode it and tell us the results.

Here is the .wav:

(IMG:https://i.ibb.co/vmgZKZx/04.jpg)




Sure I'll do this later today.

Thank you. The mystery remains. I'll also wait for you to upload the original WAV file.

Just remove the -lowpass 1 argument.
(IMG:https://i.ibb.co/mNBKzxR/03.jpg)

Hmm, are you able to tell me how I do this in dBpoweramp? I'm not seeing a way to edit that section.


Here's the WAV file. Let me know if this doesn't work:

https://mega.nz/file/0TtTzSyQ#b8KvIrnJCecub...1_YXKu_SWYhFAtU

I think I've concluded that there isn't enough data above 16 kHz in the original WAV file for it to be retained after the conversion. I've tried removing that low pass filter and it still yields the same cutoff. However, if I take a different WAV file that I have that has a much stronger spectrum in the higher areas, the conversion to 320 retains the spectrum up to 21 kHz. I'm probably not using the correct terminology, but hopefully I'm getting my point across.

You're right, different genres have different-looking spectrals. A track that doesn't have a rich frequency spectral will look like in your example.

WAV
 NIMBIFER.wav.Spectrogram.png ( 63.92k ) Number of downloads: 98

mp3 -> FLAC transcode
 NIMBIFER.flac.Spectrogram.png ( 62.42k ) Number of downloads: 96

больших семь шапок из овцы не выкроить никак

Я крут, мой винт - велик. Я предпочитаю...

FLAC.

legionz,

Turning the low pass filter on and off has no effect.

I think this case has to do with insufficient data in the spectrum above 16 MHz. The recording actually sounds like the high frequencies have been cut off.

Well, I would say, it is not quite correct to say there was "insufficient data in the spectrum above 16 MHz"
Mathematically speaking, it's just the coefficients in the Fourier transformation for those frequencies are below a certain threshold,
which is considered to be inaudible for human ear by the conversion algorithm.
If there were a setting for the value of that threshold, we could decrease it and then certainly see the higher frequencies in the spectra.

kostja,
I mean, unimportant data for the encoder. For example, the author of the recording initially had, for some reason, WAV-file created from an MP3 with a bitrate of 192 kbps, then added some kind of instrument, or maybe vocals in the midrange. That is, high frequencies are represented in the mix, but minimally. And the encoder simply considers them unimportant and cuts them off.

Yes, "unimportant" for the encoder means "coefficients in the Fourier transformation for those frequencies are below a certain threshold".

I believe that these coefficients in the Fourier transformation are simply not true, not grim and gloomy enough.