|
|
Posted by ern on 12/15/05 23:55
Thanks for your interest. This is for a project at school. I have a
verstaile mp3 player (one which I can adjust various signals / input /
firmware). I want to have a test which will yield useful results
regarding the effects of different combinations of input / criteria
from the player. The approach right now is:
1. Use "FileX" mp3 and play it on the player under "normal"
conditions.
2. Record the output onto a host program in Windows.
3. Use "FileX" mp3 and play it on the player under "a wide variety of
input combinations" conditions.
4. Record the output onto a host program in Windows.
5. Compare both files and figure out which series of input (or other
criteria) CAUSED the audio imperfection.
The tricky part here is *timestamping* the sesssion, so you know at
what point the inputs changed, and at what point the corresponding
imperfection manifested.
The part of the project that will be the most difficult is automating
everything. I want the imperfections to be reported to a text file,
*why* they happened, and under what category of imperfection they fall
without any user interaction.
A good way of ridding many of the false positives would be to not
report anything outside of the time window when the changes in input
are occuring (since I only care about those imperfections which stem
from stress test).
Navigation:
[Reply to this message]
|