sample rate


    Mar 31 2009 | 9:41 pm
    For a project where we are using sound recognition to playback a story off a person who died in the great water flood in 1953 I need to record the name at a sample rate of 16000 Hz. I've tried to use the 'sr 16000' message into the buffer~, but that results in a very slow recording where the voice sounds like a monster!
    See www.deramp.nl go to 'meer info' and there is a you tube movie of an 3D impression.
    Can somebody point me in the right direction!
    Thanx,
    Peter

    • Apr 06 2009 | 3:12 pm
      MJ wrote on Tue, 31 March 2009 23:41I need to record the name at a sample rate of 16000 Hz.
      Could you explain why? It doesn't make sense and therefore needs explanation...
      I bet you can simply record at what sample rate you think sounds best. Then if you need as an end result a specific lofi quality, you can do a sample rate conversion with for example peak or whatever you like...
      In general, your interface determines the sample rate, not a Max object...
      Stefan
    • Apr 06 2009 | 3:33 pm
      Quote:Les Ondes Mémorielles---------x--
      --_____-----------|-----------|--
      --(_|_ ----|-----|-----()-------
      -- _|_)----|-----()--------------
      ----------()----------TJ Shredder
      Hey, you found it.
      _
      johan
    • Apr 06 2009 | 4:21 pm
      if one lives in utrecht it is a good idea to make a max patch about the great flood ...
      but seriously, there is no hardware which could play audio at 16 Khz samplingrate, so it is not exactly clear what you want to do.
      is it about _sounding like 16 khz samplingrate_ ? that could be done using [degrade] with a setting of (1./44100.*16000.)
    • Apr 06 2009 | 5:06 pm
      Quote:
      is it about _sounding like 16 khz samplingrate_ ? that could be done using [degrade] with a setting of (1./44100.*16000.)
      not necessarily. degrade would give you all the aliasing of energy above 8k folding back. that will sound very different from recording at 44k and then downsampling with a good sample rate converter, or (if you can find hardware that still supports this SR) recording at 16k with a properly implemented AA filter.
      it depends what is needed: a recording delivered (for whatever reason) at 16k (and 8 bit?) but still sounding as good as possible?
      OR
      a "lo-fi" sound with lots of aliasing, distortion and noise, delivered in whatever SR is appropriate.
      16k audio can sound better/cleaner/different from degrade at 1./44100.*16000.
      regards,
      klaas-jan