Syntrillium Backup Forums :: View topic

Braunyaur · Posted - Wed Jul 24, 2002 11:29 am

As a precursor to this, let me say that I am a total newbie to audio file manipulation. I have been tasked with converting our current audio format for listening via our Intranet. Please review my information, and give any advice, suggestions, etc... that you feel is applicable.

After some research and suggestions from co-workers, I am evaluating Cool Edit 2000 to perform these functions. I do not need to totally remix the audio, just convert and compress it.

In order to save space, and provide a format compatible with Windows Media Player (this is not an option), I need to convert .VOX files to .MP3.

All .VOX files are 4-bit Dialogic ADPCM, and range in size from 80kb to 9,000kb.

The only sample rate that the .VOX files sound normal in is 6000, MONO, 16-bit. When I attempt to save to .MP3, the program gives me the following message...
____________________________________________________________________________
Title - "This wave format does not support the current sample type."

"MPEG3 Evaluation (FhG) can not handle the current sample type. Would you like to convert to 8000Hz, Mono 16-bit and continue saving?"
____________________________________________________________________________

If I tell it YES and convert the sample to 8000Hz, Mono, 16-bit, the file sounds like it's in a very deep barrel. The quality of these files is not very good to begin with, but they are understandable. I need to retain as much of the original quality as possible.

I have tried the following in an attempt to retain the original sound...

1 - Opening file using the "Open As..." and setting the Sample Rate to various settings from 6000 up to 441000Hz.
2 - Setting the "Interpret Sample Format As..." to 6000.
3 - Previewed file to verify correct sound.
4 - Saved file as .MP3 with Options set at
VBR - 10 (55-65Kbps)
Maximum Bandwidth - 4960Hz.
VBR Quality - 10
Padding - ISO Padding

The quality is acceptable, but the file compresison is non-existent. I may save 5 or 10k.

I am thinking of using the following settings...

Open the file with Sample Rate of 16000, and an Interpret Rate of 6000.
Save file as .MP3 with Options set at...
CBR - 16kbps, 16000Hz, Mono
Maximum Bandwidth - 4000Hz.
Codec - Fast Codec (High Quality)
Padding - ISO Padding

This seems to give the best quality, and acceptable compression (about 30% on a 150k file). I will not know fully until our company decides to purchase the software (1 minute limitation).

Is there anything that I am missing that can give me better performance or compression? Am I on the right track, or way off base?

Any help or advice is appreciated.

Braunyaur

Daan van Rooijen · Location: Netherlands

Dialogic ADPCM is used mostly for voice response systems, and is highly compressed. I don't think that MP3 will squeeze them much tighter.

Your procedure for opening and interpreting the files is correct. VOX files do not contain any information as to the sampling rate at which they were recorded, and if you can't ask anyone who knows, you'll have to find the correct rate by ear.

I really don't know if this will help, but two things that I'd try are..:

1. After reading the files, try the Convert Sample Type option under Edit to increase the sampling rate to 44.100. This won't improve the sound but it should give you more options for MP3 compression.

2. Maybe the MP3 encoder doesn't take well to your already heavily processed (filtered) recordings. You could try creating a same length WAV file that contains inaudible (ie very low volume) white noise (Generate > Noise), and mix-paste this with your speech file. Then save as MP3 and see if it sounds any better than without the filler noise.

post78 · Location: USA