I'm currently working on the EAQUAL(based on ITU-R BS.1387) code for evaluating mp3 encoders(LAME and ISOMP3 to be specific). But I'm facing problems because the encoders use a 1152 sample framesize whereas PEAQ/EAQUAL uses 1024 sample framesize. Besides, PEAQ/EAQUAL uses 109 critical bands per frame, unlike mp3 encoders that use 21. So, how do I ensure a fair comparison of NMR values on a frame-by-frame basis?
Can anyone also tell me abt the psychoacoustic modelling done in EAQUAL, especially how is the masking threshold computed?