To explore the process of audion compression.
Reality Check 11
1) Create modified discrete cosine transformation (MDCT) matrix
2) Quantize and dequantize
3) (Optional) Apply windowing function
4) Apply inverse of MDCT
Steps:
1) Set bits to be 4 and window size to be 32.
2) Input signal of frequency \(64Hz\).
3) Compare the original and decoded signal and the difference (RMSE) between them.
4) Try different even and odd frequencies and compare the result.
Matlab code: here
Result:
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Steps:
1) Adjust code to apply the windowing funciton.
2) Set bits to be 4 and window size to be 32.
3) Input signal of frequency \(128Hz\).
4) Compare the original and decoded signal and the difference (RMSE) between them.
5) Try different even and odd frequencies and compare the result.
Matlab code: here
Result:
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Steps:
1) Set bits to be 4 and window size to be 32.
2) Input a chord signal 261.63Hz, 329.63Hz, 392.00Hz and 523.25Hz (C)
3) Compare the original and decoded signal and the difference (RMSE) between them.
4) Try different bits and compare the result.
Matlab code: here
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
Steps:
1) Input an audio signal.
3) Compare the original and decoded signal and the difference (RMSE) between them.
4) Try different bits with and without the windowing function, then compare the result.
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |
![]() |