Some of my IEMs only go to 16khz for each channel (or phone). So I was thinking to encode my wavs to 320kbps cbr bitrate 32khz mp3 for use on my portable mp3 player (which only plays mp3, aac, and wav).

What way would get me the best quality (closest to original)?

1. Downsampling the 44.1khz wav to 32khz and then coverting to 32khz mp3.
2. Downsampling the 44.1khz wav to 32khz wav then upsampling it to 44.1khz wav then converting it to 44.1khz mp3 (leaving a black area above 16khz).
3. Converting the original 44khz wav straight to 32khz mp3.

Now when I do both 1. and 2. I get the same file size for each. But looking at 2.'s spectral graph in Adobe Audition I see a black area above 16khz going to 22khz in each channel. So I'm thinking there is wasted bits in this black area. Am I right? If so then in 1. is all the bits concentrated from 0-16khz and thus giving me the better quality than 2.? Is there a 4th option that is better than those three above?


Thanks!

Well, you can obtain the best possible quality @ 32KHz going to MONO.

Suggestions:

• resample with a good resampler (such as SSRC, PPHS or even r8brain);
• downmix to mono with Stereo Tool;
• keep in consideration Vorbis or AAC in place of MP3;

320 kbps and 32 kHz sampled mp3 isn't a perfect match.
32 kHz sampled mp3 means you have a frequency limit ~ 15.5 kHz and more serious pre-echo problems than when using 44.1 kHz.
I wouldn't accept these restrictions with only your current earphone in mind. Your audio system can vary, and maybe you will want to share your music with someone else within the legal restrictions.
A better way is sticking to 44.1 kHz and use Lame with a lowpass, for instance --lowpass 16.0, or use Helix which discards everything beyond 16 kHz.
I guess Lame 3.98.2 -V0 --lowpass 16.0 is an excellent choice for your needs. Of course you can use -b 320, but I suggest to use mp3packer afterwards which losslessly will bring bitrate down by ~ 5..10%.

In both cases you've instructed the MP3 encoder to use 320 kilobits for each second of sound.

Given the bitrate and playing time you can calculate file size:
File Size in MB = (Bitrate in kbps x Playing Time in minutes) / 140
(There is some rounding in that formula, and it doesn't include tags, headers, or any other file overhead.)

Thanks! I have adobe audition as well as nero wave editor.

1. So if I use the lowpass feature then the spectral graph of the music file will reveal a black cuttoff area from 16-22khz. Now say I encode in 320kbps CBR. Now in any given second will any of the 320bits be allocated towards the black cuttoff area?

If yes then does 16/44(320)= 233 bits go towards then non black area (music area) and 320-233=87 bits goes toward the black cuttoff area?

What I'm trying to do is pack all the bits into the 0-16khz range and non in the 16-22khz range so I can get greater quality than 320kbps with same file size or same quality as a 320kbps file but with smaller file size.

2. I'm kinda a noob. So what program is not to hard to use to do encode with a Low pass filter in Lame? And exactly (list complete commands) how do I do it with this program (although you listed the partial commands above)?

If getting 'closest to the original' is a factor, I don't see why you would want to resample at all.

As you have noticed, it won't save you any space with cbr.

I suggest you try V2, or even V4 or 5, and compare the quality to your 320cbr on whatever device you listen to, then compare the file size savings and find the balance you desire.

I can't abx V0 or even V2 from original wav or from 320cbr. Yet the placebo effect is very real for me though. If I know I'm listening to something of higher quality then it will sound better to me and make me feel more satisfied. So I operate 100% on the placebo effect and not reality. I guess than is why some people listen to flacs instead of V0. But thanks!

Only a tiny fraction of an overall bitrate. Nothing to worry about.


Definitely not.


Do you really need the highest possible bitrate? (Edit: was answered in prev. post)

For Lame CBR 320 lowpass 16 kHz use: -b 320 --lowpass 16.0
This way Lame encodes only frequencies up to 16 kHz, and all the 320 kbps quality goes into this frequency range.
As it's about peace of mind qualitywise and as you don't have to care about very high bitrate: in case your DAP supports FLAC you can use lossyWAV + FLAC which at the superb 'portable' quality level takes ~ 380 kbps on average.
Not so much more than 320 kbps.

Well in that case, only full original wave files should do you.

My player doesn't have enough room plus my iems don't go beyond 16khz. So I figured 32khz mp3 was best for my needs. I'm considering nero aac but have no clue how to do a low pass filter encode with 320kbps CBR.