4 ms·
Correct, last article I recall reading about deciphering VBR from packet size alone was something in the neighborhood of 50% success rate.
by CodeMichael 10y ago
Correct, last article I recall reading about deciphering VBR from packet size alone was something in the neighborhood of 50% success rate.
- Y_Y 10y agoThen why not compress really efficiently by just transmitting packet sizes?
- c22 10y agoBecause 50% of them won't be understood?
- e12e 10y agoOn the other hand, if you could do it, you'd probably have invented a convoluted speech-to-text (where text is a index into a dictionary of words). Note that you would also likely lose things like inflection, voice, accent etc - so while it might work as a texting system with voice input - it would be a poor substitute for voice chat..