1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Compressing Voice .Wav Files

Discussion in 'Audio' started by Angbates, Mar 14, 2004.

  1. Angbates

    Angbates Guest

    What is the best compression form to send voice .wav files through email? I'm sending voice recordings that are about 1 to 1 1/2 hours long to a transcriptionist.
     
  2. jonss

    jonss Guest

    Is it necessary that they remain as .wav files? I don't know of any way to reduce a .wav file other than to convert it to mp3. If you obtain a copy of CDex (freebie) you'll be able to compress them to around 10% of their original size. Your ISP may have some restriction on file sizes for email, so find out from them what is allowed. If the person on the other end has CDex as well, they can reconvert them to .wav using the same program. There will be some quality loss but as it's voice there should'nt be any perceptible loss.
    Just reread your post. Files of that size, 1 to 1 1/2 hours are still going to be BIG. I'd edit them with a .wav editor and split the track at certain points and send them in sequence. Other than that, burn the whole thing to a CD-R and send them snail mail. It'll probably get there quicker than the time it takes to download.
     
    Last edited by a moderator: Mar 15, 2004
  3. tigre

    tigre Moderator Staff Member

    Joined:
    Feb 18, 2003
    Messages:
    790
    Likes Received:
    0
    Trophy Points:
    26
    There are much better choices than mp3 for this task.

    Basically it works like this:
    .wav file -> lossy compression -> compressed file -> email -> decompression to .wav -> transcription.

    I'd give speex a try as compression format. It's specialized on compressing/decompressing speech and gives usable quality at amazingly low bitrate (compared to mp3). Try "Speex drag'n'drop frontend" from here: http://www.rarewares.org/others.html
    On the same site there are commandline tools using newer versions of speex; you might want to try those additionally (not as comfortable to use as drag'n'drop frontend obviously, but you could use a .bat file/script for drag'n'drop encoding/decoding).

    I'd appreciate freedback in case you like it. :)
     
  4. Angbates

    Angbates Guest

    Thanks for everyone's suggestions. I'll look into rareware. But with speaking with some other colleagues they've suggested using whalemail.com or xdrive.com which would send those large files. Anyone has any feedback of those?

    Thanks!
     
  5. Angbates

    Angbates Guest

    Hi Tigre,
    I downloaded the speex drop and drag file, installed and "dropped adn dragged" one of my recorded .wav files, it converted it -- I guess, and I tried to play it back on Media Player and came back with error message that it did not support file format????
     
  6. tigre

    tigre Moderator Staff Member

    Joined:
    Feb 18, 2003
    Messages:
    790
    Likes Received:
    0
    Trophy Points:
    26
    To play it back, you have to
    - use an audio player that supports the format, e.g. foobar2000 or winamp (with plugin) or
    - just drag'n'drop the compressed file again on the speex frontent - the result will be a .wav file again that can be played back with any media player :)
    ! BE CAREFUL - IF YOU DON'T SELECT AN OUTPUT DIRECTORY ENCODING -> DECODING WILL OVERWRITE YOUR SOURCE FILE ! So either rename the encoded file after encoding, e.g. from test.spx to test_.spx, or move the encoded files to another folder for decoding (that would happen automatically if you e-mail them to someone else ;) ).
    This means the recipient of your mails must have "Speex drag'n'drop frontend" too.

    Another option would be using the frontent for playback directly:
    Rightclick -> Decoding options -> Check "Playback"

    While we're at it...
    some decoding options that might be important for you:

    - Pre-Amp: You can use it to adjust the volume of the decoded file / playback volume
    - Dither: It adds a small portion of noise to avoid truncation distortion. Recommendable if you choose "Unsigned 8 bit PCM" as output.
    - Output Wave Format Settings:
    16 bit PCM is CD resolution and recommendable quality-wise. It won't sound better with higher resolution. If you want smaller output files (1/2 size) use "Unsigned 8 bit PCM (default). Drawback: There'll be either some audible noise added (if Dither is used) or some distortion (without dither). If you run into playback problems with 8 bit try 16 bit.

    Encoding options:
    - Encoding quality: Lower: smaller files, Higher: better quality
    - Encoding complexity: Lower: faster en/decoding Higher: better quality
    - Downmix Stereo to Mono: Recommendable (If there's only one person speaking, stereo is useless and only wastes space).
    - ABR: "Average Bitrate", gives you total control over bitrate -> filesize. Try it if you need to reach a certain target filesize (e.g. your email attatchments must be smaller than 2 MB)
    - VBR: "Variable Bitrate" might increase quality/size ratio
    - Resample Output to: If your source .wav files are sampled at other sampling rates than 8000/16000/32000 Hz, e.g.44100Hz = CD quality, check this box and choose 8, 16 or 32. Lower will cut off high frequencies but give lower filesize/better compression. The best compromise will be 16000 IMO.

    Don't worry: If you don't want to mess with those options just try the default ones and only change something if you're not satisfied with the result.
     
    Last edited: Mar 15, 2004

Share This Page