How to calculate the coefficients of the lowest eight filters which belong to the MPEG-Audio 32-band filter bank.

46 views
Skip to first unread message

Yu Shi

unread,
May 5, 2016, 5:51:58 AM5/5/16
to echoprint
Dear all,
           
            I want to improve the performance of the short time detection, such as the samples less than 5s. So I want to improve numbers of the band filter. 
      But I am not sure how to calculate. Although I already know the 512 coefficients of the MPEG-Audio 32-band filter.

      Help! Help! It makes me feel so bad.


     Thank you everyone  

Eduardo Avaria

unread,
May 6, 2016, 8:28:37 AM5/6/16
to echoprint
I had the same problem some years ago and finally I developed my own algorithm to match on 3 seconds. You can find my thread for more info. Let me know if I can help you with anything

Yu Shi

unread,
May 9, 2016, 8:02:05 AM5/9/16
to echoprint
Thank you for your answer which make me know it's possible to match on short times. Please forgive me, I don't know how to find your thread for more info. I am looking forward to learning from you.

Eduardo Avaria

unread,
May 9, 2016, 9:46:10 AM5/9/16
to echo...@googlegroups.com

--
You received this message because you are subscribed to a topic in the Google Groups "echoprint" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/echoprint/AEI5Lqj3aaQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to echoprint+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Yu Shi

unread,
May 11, 2016, 7:28:14 AM5/11/16
to echoprint
I have read the paper, do you think it's possible to recognize 5s length audio by detecting onsets? I try to increase 8 bands to 16 bands, but the result was not very good.

Eduardo Avaria

unread,
May 11, 2016, 8:51:20 AM5/11/16
to echo...@googlegroups.com
Yes, in my implementation I designed de algorithm to fingerprint every 11.5 ms, and work with a 256 samples vicinity, so you can recognize 3 seconds of audio. I have a demo prepared to work on big data systems, aiming to monitor radio and tv all over the world (with the proper FM interface). 
Contact me about the details of your project and maybe we can work together using my implementation.
Reply all
Reply to author
Forward
0 new messages