"Not Enough Space Left On Destination Drive..." Partition Size Limitation?

Hi-

I'm doing a largish project for a local low power FM station, normalizing all of the MP3s they use for broadcast.  Not too far into the process I started getting the

"Error: Out of disk space Not enough space left on destination drive to save entire file. Do you wish to keep the portion that has saved so far?"

message.  The C: drive has 8.5G on a 74G drive, the F: drive (external USB drive) where the files and temp folder is located has 40.3G on a 600G drive.  Both are NTFS partitions.

Does Cool Edit 2000 have some kind of limit between 40 and 50G on the amount of data it can see on a drive?

Windows XP SP3 with all updates as of today, very recently rebuilt, 1G RAM, 2.0G Celeron processor.

Richard

When an mp3 file is opened in Cool Edit it is internally converted to wave which must be stored somewhere. If the mp3 files are large, this may be the problem.

Having said that, Cool Edit is the wrong tool for your task. Although it may not be obvious, the file is being converted to wave and then back to mp3, which results in a further loss in quality. This will also be true of any audio editor.

Have a look at mp3gain which is designed to do exactly what you want without any quality loss. It will also operate on many files in one batch, which will allow you to process large proportions of your job unattended, e.g. overnight.

When an mp3 file is opened in Cool Edit it is internally converted to wave which must be stored somewhere. If the mp3 files are large, this may be the problem.

Having said that, Cool Edit is the wrong tool for your task. Although it may not be obvious, the file is being converted to wave and then back to mp3, which results in a further loss in quality. This will also be true of any audio editor.

Have a look at mp3gain which is designed to do exactly what you want without any quality loss. It will also operate on many files in one batch, which will allow you to process large proportions of your job unattended, e.g. overnight.

Thanks very much for your response!

It doesn't seem to be a function of file size per se, because these are basically pop songs, with a maximum size of under (usually well under) 10M, so unless Cool Edit is trying to read all the .WAV versions of the files into RAM at once I wouldn't think that was the problem.

MP3Gain does seem to be very promising, although I find their approach a bit different.  Performing the tasks without having to re-compress is perfect, but the whole idea of mixing up the concepts of compression and peak normalization is counter-intuitive to me.

Compression is actually not a bad thing at all, considering that I'm processing these files for radio broadcast, but is there a way to batch the process of 1) compressing all the files to a given perceived loudness level, then 2) doing a peak normalization to a given percentage of maximum digital volume level?  I find their model of targeting an arbitrary peak dB level odd.  I understand that their target audience is mostly people trying to create tracks suitable for playing on a portable device in high-noise environments, so it may work for that purpose.

Thanks for your help!

Richard

Hydrogen Audio Forums is where the developers hang out. You may find more insights on the most economical route to your goal over there.
http://www.hydrogenaudio.org/forums

Compression is actually not a bad thing at all, considering that I'm processing these files for radio broadcast...

If the broadcaster concerned has something like an Optimod in the transmission chain, then you may well find that material that has already been compressed is severely compromised by it. Not only will it be compressed twice, it will be by completely different algorithms. Pretty much all transmission chains are designed to produce the maximum levels they can from untreated material, without it sounding 'strained' - and that's what Optimods excel at doing well.

I hadn't picked up from the first post that you were compressing as well as normalising. mp3gain will only normalise to a chosen level.

As to whether you should be normalising, Steve has addressed that perfectly.