mp3 encoder?

[Rusty K]

Maybe I should have asked this question before now, but which mp3 encoder does AA use?

Thanks,
Rusty K

Originally Posted by Rusty K

Maybe I should have asked this question before now, but which mp3 encoder does AA use?

Thanks,
Rusty K

Adobe Audition 1.5 uses MP3Pro which uses the Fraunhoffer codec.

From what I've heard, some (very few) players don't support MP3Pro encoded MP3 files but these were online flash players.

[LogicDeLuxe]

Originally Posted by danny.guitar

From what I've heard, some (very few) players don't support MP3Pro encoded MP3 files but these were online flash players.

There is a difference between "support" and "can play". MP3pro is designed compatible to regular mp3. Though, if it is played on a player which doesn't support MP3pro (which there are a lot of, especially hardware players), you'll get a frequency response up to about 10 kHz only.

Also MP3pro doesn't work that well with some styles of music, although it can be quite good at some.

You can just disable MP3pro encoding and have regular MP3, so that's not that big a problem.

The majority consider LAME a much more sophisticated encoder, thus you should give that one a try, too. There is even a plug-in: http://www.rarewares.org/mp3-lame-libraries.php
Alternatively, you can just save your song to 24 bit wav, and let a standalone LAME encode it, which sure is useful, if you master with another software.

I use Razorlame to save my settings and batch encode files (when I'm outside AA).

[Rusty K]

LogicDeluxe,

I downloaded and worked with WinLame last night and for some reason it didnt' work. Visually the program went thourgh the whole process but when I checked the overwritten files with AA. Nothing had changed. Also even though I selected 128/bps it kept converting to 192/bps (default) which by the way I noticed was a much better sounding mp3.

I'm going to check out RazorLame

Thanks for the Logic.....Deluxe!

[Rusty K]

something interesting....

I just encoded the same wav file with WinLame and then again with AA for comparison. Both at 128bps. I then analyzed both...

WinLame.....showed possible clip L ch-7/R ch-1.

AA.............showed possible clip L ch-9/R ch-11.

I checked with "Clip Restoration" and found there was no actual clipping in either instance.


One thing about WinLame is it doesn't seem to change the name of the file after processing. It remained a wav file in name so I had to rename it by hand. That could be very confusing and messy. The program also doesn't seem to "overwrite" even when that option is checked.


Rusty K

[LogicDeLuxe]

Originally Posted by Rusty K

I checked with "Clip Restoration" and found there was no actual clipping in either instance.

Then you loaded as 32 bit float, I guess. If you load as 16 bit, or if you convert to 16 bit after loading, it will clip.

One thing about WinLame is it doesn't seem to change the name of the file after processing.

All my software which uses lame saves mp3 files. No wav files. I didn't even know that there is such an option. Does this happen with RazorLame as well?

The program also doesn't seem to "overwrite" even when that option is checked.

If it is in fact writing a wav file, and if you chose to write in the same directory, it is probably because it can't overwrite the file which it is currently reading, obviously.

[Rusty K]

Then you loaded as 32 bit float, I guess. If you load as 16 bit, or if you convert to 16 bit after loading, it will clip.

Nope the 16 bit file went directly into WinLame. The file I did with AA I did open in AA so you tell me was that 32bit float? The only way to bypass 32 bit float in AA is to use the "batch conversion" I guess?

All my software which uses lame saves mp3 files. No wav files. I didn't even know that there is such an option. Does this happen with RazorLame as well?

Yes I thought this was weird as well. I just downloaded RazorLame so I haven't had time to check it out.

If it is in fact writing a wav file, and if you chose to write in the same directory, it is probably because it can't overwrite the file which it is currently reading, obviously.

Yes this dysfuntion seems to be a bug in the program. I hope RazorLame works better.

This may be another stupid question but I don't have to download Lame itself in order for these programs to work do I? I mean Lame is already in the program download right? When a Lame update comes along what do I do to update the program?

Rusty K

[LogicDeLuxe]

Originally Posted by Rusty K

Nope the 16 bit file went directly into WinLame. The file I did with AA I did open in AA so you tell me was that 32bit float? The only way to bypass 32 bit float in AA is to use the "batch conversion" I guess?

As I told you in the other thread, there is a "decode to 32-bit"-option in the mp3 options in the "save"-dialog. Disable it, and then load your mp3's. They will be quantized to 16 bit then, and clip at 0 dB.

This may be another stupid question but I don't have to download Lame itself in order for these programs to work do I? I mean Lame is already in the program download right? When a Lame update comes along what do I do to update the program?

Usually those GUI's come with LAME. You certainly will notice, when LAME is missing, as the software will complain about it. If that is the case, you just download a precompiled LAME as well and put it along with the GUI. Updating LAME works the same way.

I use cDex. It's free. Download the installer and install it:
http://cdexos.sourceforge.net/?q=download

Then go to Options -> Settings and change your settings to look like this.



Click OK to close the settings window. Then just drag & drop the MP3 file into the empty list on the main window. A window will pop up. Click "Encode". It will save it to: \My Documents\MP3\

Edit: Oops I mean \My Documents\My Music\MP3\

[sikter]

Which settings are "The best" for an acoustic guitar project if the limit would be 10 MB for 4 minutes song?

I don't have any monitors right now so I would appreciate any help from anyone...
I appreciate if you guys could check my acoustic project (follow my sig) and tell me what to do in the mix.

[LogicDeLuxe]

Originally Posted by sikter

Which settings are "The best" for an acoustic guitar project if the limit would be 10 MB for 4 minutes song?

That should do it, and is unlikely to exceed the 10 MB on any style:
LAME -b 32 -m j -h -V 0 -B 320 -q 0 -k

[Rusty K]

I've been unable to use RazorLame. Whenever I try to "encode" I get an error message....

The File "C:ProgramFiles/RazorLame/Lame.exe" cannot be found.


There is no Lame.exe file in this program or in the 3.97 version of Lame that I thought I downloaded from SourceForge. The only file in the Lame download was Lame-3.97.tar.tar.

This prompted my question about whether or not Lame was already in these program downloads.



Rusty K

[LogicDeLuxe]

You can get the lame.exe here: http://www.rarewares.org/mp3-lame-bundle.php

[Rusty K]

LogicDeluxe,

Once again you come to the rescue. Thanks so much for all your help. I am confident that you are as big a help to others as you have been to me. You are doing a good thing bro.

That last conversion with Razor yielded the same result as with WinLame. That was straight from a wav file at 16bit 44,100. Possible clips 7-L, 1-R with no actual clipping although the AA vu's show red.

Rusty K

[LogicDeLuxe]

Originally Posted by Rusty K

Possible clips 7-L, 1-R with no actual clipping although the AA vu's show red.

Could well be, as the indication methods are differently.

- The clip indicator at the peak meter detects a clip when two or more consecutive samples are at 0 dB or above (either positive or negative side). If "adjust for DC" is checked (which I don't recommend, as I can not imagine any usefulness of that function), the sample is high pass filtered before the detection, which usually results in more clips.

- The clip detection of the "Clip Restoration" doesn't take absolute level into account, but only counts samples which are close to each other. The result depends on the thresholds "Overhead" and "Minimum run size". This one can detect more clips when a mp3 is loaded as 16 bit, as clips can actually occur during the loading.

- And at last, we have the "Analysis"-tool which also includes a clip detection. It counts every situation where one or more consecutive samples are at 0 dB or above (either positive or negative side). This method of detection is considered overcautious, this is why it is properly labeled "Possibly Clipped Samples", ie. it can't be said for certain.

In general, less then 3 consecutive samples at the limit can be safely ignored, and most peak meters do so.

[Rusty K]

Jeeze...the further I go with this the more confused I'm becoming. Nothing is acting as described.

I re-ripped the tracks from CD. I first tried converting the files as is to mp3 with Razor. When analyzed they showed +.01db both channels with actual clipping when analyzed with clip restoration. So I then tried to convert the files in AA with "decode at 32bit" checked and the mp3 turned out to be clipping even harder. I then tried batch converting the files to 32bit in AA and then encoding. First, when analyzed at 32bit (wav) the level was below the actual level of the original wav file but then after encoding to mp3 they were back up to +.01db.

Can I just use the "amplify" effect in AA to bring the level of the mp3 down below clipping? "Normalizing" was mentioned but "amplify" seems easier.

Hopelessly confused,

RustyK

[LogicDeLuxe]

Originally Posted by Rusty K

I first tried converting the files as is to mp3 with Razor. When analyzed they showed +.01db both channels with actual clipping when analyzed with clip restoration. So I then tried to convert the files in AA with "decode at 32bit" checked and the mp3 turned out to be clipping even harder.

Did you turn of "Account for DC" in the analysis tool? You should, as only then, you'll get the peaks which real represent what you are seeing in the waveform editor.

If the measuring is set up correctly, you never will get bigger than 0 dB peaks on a 16 bit wave.

[Rusty K]

Did you turn of "Account for DC" in the analysis tool? You should, as only then, you'll get the peaks which real represent what you are seeing in the waveform editor.

Yes it is always checked. the analysis in the left channel was up to +.58db L and +.64 R. This is almost a +.90db increase in both channels.


Rusty K

[Rusty K]

I've encoded the same wav file using WinLame with even worse results. +1.26db L and R....this is a +1.5db increase and most definitely clipping.

I think I'm going to take this thread/question over to the mp3 mixing clinic.


Rusty K