Bit depth to MP3

  • Thread starter Thread starter Alimbuyuguen
  • Start date Start date
Sleepy, the bit-rate is just defines the minimum network bandwidth required to stream the file in real-time. This doesn't mean that the information contained in these containers is accurate. You're probably familiar with the quality/speed option when creating a CBR MP3 file, it's very audible. This setting increases the precision of all calculations at the expense of increased processing work during encoding - but it doesn't touch the bit-rate.

The same thing happens when you feed the encoder with more information, the results of all these calculations will be more accurate. Most DAWs and plugins also use a higher bit-depth (24bit or more) than the target bit-depth (usually 16bit CD format), for very good reasons. ;)

MP3 encoding involves a lot calculations, the more accurate the input, the more accurate the output. I'll do some experiments and post the resulting wavs, because it's still theory. :)


On a side note, I began to develop a realtime vst format "monitor" several years ago, but never found the time to bring it further. The idea was to easily preview several mp3 encoding/decoding settings (and more formats) during mastering. Maybe i should give it another try ;)
 
Last edited:
That makes more sense. So; the shell is 320kbps, but inside, that entire 320kbps may, or may not be used.
 
So, I am still trying to understand something...

Does mp3 actually have a samplerate and/or bit depth that relates to the 44.1/48 and 16/24 as we know it from wav/aif?

Actually a bit curious about this.
 
Why was this locked? One of the most interesting discussions of FP for a while...now open again!
 
Thank you krushing. Finally I can add my 2 cents. A .wav file has a sample rate of 1411kb/s, and the top for mp3 is 320kb/s. this is kiloBITS, not bytes....for those who are wondering. There are many other factors when it comes to the quality of an mp3 file or encoding. First and foremost is the audio codec being used. Fraunhofer II is the best - bar none. Secondly are the settings when encoding a file to mp3. There are a multitude of options. If file size is a concern (which it shouldn't be in my mind - a 5 minute song would have a file size of 10-12MB in top mp3 configuration), you have the option to adjust things such as the high frequency cutoff, whether or not the sample rate is CONSTANT or VARIABLE (the bitrate will vary - on passages with less sound 'information', the sample rate will drop, and it will come back up when there is more going on), and options such as true stereo or MS-joint stereo (which takes similar information that produces sounds in both speakers and only keeps one copy, thereby saving space). For the best quality mp3 available, I would suggest Fraunhofer II codec, constant bitrate of 320kb/s, true stereo and no joint stereo options, and a high frequency cutoff of 22050Hz (which is the highest it can go - if you do the math 44.1k divided by 2 is 22050 - one for each speaker in the stereo spectrum). If your file is 48K instead of 44.1 (as some newer CDs are released 48k), you can put the HF cutoff at 24000Hz.
 
Last edited:
RyanFields said:
and a high frequency cutoff of 22050Hz (which is the highest it can go - if you do the math 44.1k divided by 2 is 22050 - one for each speaker in the stereo spectrum).
Click to expand...

The reason is not actually "one for each speaker"... it is a number based on the Nyquist theorem which essentially states that in order to reproduce audio that goes up to a given frequency, you must have a sample rate of twice that frequency to achieve that. This is the case regardless of whether the audio is mono or stereo.
 
OK, So what everyone's saying is...

.wav and .mp3 are really the same thing, but .mp3 has a far lower Bitrate; 320kb vs 1411kb

is that right? In my earlier comment I mentioned how my .mp3s list a samplerate depending on what the original .wav they were made from was. So really .mp3s have a bit depth and sample rate, it's just held in a 320kb wrapper rather than a 1411kb wrapper.

Yeah? Am I getting this??
 
Bryst said:
OK, So what everyone's saying is...

.wav and .mp3 are really the same thing, but .mp3 has a far lower Bitrate; 320kb vs 1411kb

is that right? In my earlier comment I mentioned how my .mp3s list a samplerate depending on what the original .wav they were made from was. So really .mp3s have a bit depth and sample rate, it's just held in a 320kb wrapper rather than a 1411kb wrapper.

Yeah? Am I getting this??
Click to expand...

no. i don't believe so.
 
This was a mod-only discussion until krushing ruined it for all of us. jp Didn't notice it had been locked. I had some problems editing/deleting one of my posts a few days ago, so may have been my fault. Apologies to anyone that was interested in the discussion and couldn't contribute due to my error.

dvyce said:
So, I am still trying to understand something...

Does mp3 actually have a samplerate and/or bit depth that relates to the 44.1/48 and 16/24 as we know it from wav/aif?

Actually a bit curious about this.
Click to expand...

As far as this part goes, yes, they do have sample rates just like we have with .wav files. You can even export lower sample rates with .mp3's, and the loss of high frequency content is evident just like with .wav files. I'm just not really convinced about the bit depth part, yet. I have had times when I needed to make the smallest .mp3 file possible, and at least with Cubase, lowering the sample rate lessened the audio quality, but not the file size. But I think I was using CBR and not VBR for compatibility reasons. I generally use CBR anyways. I do believe Cubase used the Fraunhoffer codec but they may have switched to LAME. For a while, Fraunhoffer sounded better with smaller files.

Bryst said:
OK, So what everyone's saying is...

.wav and .mp3 are really the same thing, but .mp3 has a far lower Bitrate; 320kb vs 1411kb

is that right? In my earlier comment I mentioned how my .mp3s list a samplerate depending on what the original .wav they were made from was. So really .mp3s have a bit depth and sample rate, it's just held in a 320kb wrapper rather than a 1411kb wrapper.

Yeah? Am I getting this??
Click to expand...

No. Image file formats are different cups. A .wav file is like filling that cup up to to the top no matter if you're going to drink it all or not. The point being, you always have enough water regardless. Whereas an .mp3 file, that cup is only going to be filled up to whatever is enough for you and mostly it's enough but never as much as the .wav file.

Not a perfect analogy, but because .wav's can contain a higher bit rate so you can look at it as being a bigger cup, also.
 
sleepy said:
No. Image file formats are different cups. A .wav file is like filling that cup up to to the top no matter if you're going to drink it all or not. The point being, you always have enough water regardless. Whereas an .mp3 file, that cup is only going to be filled up to whatever is enough for you and mostly it's enough but never as much as the .wav file.

Not a perfect analogy, but because .wav's can contain a higher bit rate so you can look at it as being a bigger cup, also.
Click to expand...

No that helps. That's kinda what I was thinking. Just couldn't express it properly.

This is all really good stuff to know.

:cheers:
 
You must log in or register to reply here.
Back
Top