Could a full song be possible in only 94kb of audio data?
A full 1:36 long song in just 94kb that's tolerable to listen to and still sounds like a song? (Not good quality but not so bad you cant hear anything but distance rumbling?
Would the samplerate have to be decreased and treble increased to compensate for samplerate loss?
A full 1:36 long song in just 94kb that's tolerable to listen to and still sounds like a song? (Not good quality but not so bad you cant hear anything but distance rumbling?
Would the samplerate have to be decreased and treble increased to compensate for samplerate loss?
Last edited:
Can it be done with just a simple program like Audacity alone?
And would the song sound fuzzy/buzzy/ringy or garbly from the compression method? Which would be more tolerably more pleasing to hear?
The file can also be then placed inside a .zip folder compressed even further shrinking the file
And would the song sound fuzzy/buzzy/ringy or garbly from the compression method? Which would be more tolerably more pleasing to hear?
The file can also be then placed inside a .zip folder compressed even further shrinking the file
I've done this before. You can go down to 64kbps with OGG, and that will give you a nice audio file with the sound quality equivalent of MP3 128kbps.
But to go deeper down Its honestly best going with Joint Stereo on MP3 and using these settings in lame: lame -m j --resample 11.025 -b 16 -r -q 0 --verbose track01.cdda.raw track01.cdda.mp3
Those settings will get a 1:30 song down to 200kbytes.
At these settings you really need to record in 8000khz or 11.025khz, you could do 22khz but it will result in larger file sizes.
There was an older codec I remember from back in the 1990s which could do a song in 100kbytes but it was a mobile phone codec, I forget its name now. It alludes to the fact that it can be done with a proprietary codec made specifically for an industry.
I think it was a part of the codec package which was supplied with that Voice Recorder program that was bundled with Windows 95 & 98.
If anyone still has their copy around maybe you could give it a whirl, you need to go into the preferences and specifically tell it to use that codec.
You get essentially the same audio quality as you would on a 2G mobile phone call. Great for voice no good for audio.
Its one of these codes anyway: https://en.wikipedia.org/wiki/Comparison_of_audio_coding_formats
I would also strongly recommend that you do a low pass filter step aswell, no point in encoding higher frequencies that you cannot hear anyway due to the low bitrate.
It sounded good though.
But to go deeper down Its honestly best going with Joint Stereo on MP3 and using these settings in lame: lame -m j --resample 11.025 -b 16 -r -q 0 --verbose track01.cdda.raw track01.cdda.mp3
Those settings will get a 1:30 song down to 200kbytes.
At these settings you really need to record in 8000khz or 11.025khz, you could do 22khz but it will result in larger file sizes.
There was an older codec I remember from back in the 1990s which could do a song in 100kbytes but it was a mobile phone codec, I forget its name now. It alludes to the fact that it can be done with a proprietary codec made specifically for an industry.
I think it was a part of the codec package which was supplied with that Voice Recorder program that was bundled with Windows 95 & 98.
If anyone still has their copy around maybe you could give it a whirl, you need to go into the preferences and specifically tell it to use that codec.
You get essentially the same audio quality as you would on a 2G mobile phone call. Great for voice no good for audio.
Its one of these codes anyway: https://en.wikipedia.org/wiki/Comparison_of_audio_coding_formats
I would also strongly recommend that you do a low pass filter step aswell, no point in encoding higher frequencies that you cannot hear anyway due to the low bitrate.
It sounded good though.
Last edited:
What i done was boost the bass by 7.5db and the treble by 15dB a couple times in sync (treble + bass boosted at same time) (Yes a bit extreme i know.) then de-amplify the music till it's at a modest volume and non-clipping. then i done the rest of the low-sample rate and low bit rate stuff I had it encoded at 8kbp/s MP3 and 8000 samples/sec and used a single mono channel in "Stereo" mp3 mode seemed to give a few bytes smaller than joint stereo with identical audio
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.