Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.
The cepstrum is defined as the inverse Fourier transform of the log-magnitude Fourier spectrum. The Mel scale is roughly linear with Hertz scale to 1kHz then with increasing spacing approx. 50% per octave above this, according to human hearing perception.
The MFCC scaled filter banks then become wider at higher frequency vs equal spaced for linear filter or LFCC.
Overall, MFCC makes a better feature set as compared to cepstral coefficients ( LFCC ).
But when the speaker and listener both have stronger signals above 1kHz, then LFCC will outperform MFCC for error and recognition rates such as in women's voices.
LFCC also is slightly better with interference from mumbling noise but not with white noise.
Perhaps an adaptive method will perform better , moving somewhere between the two methods.