8bit or 16bit

Discussion in 'Working with Sound' started by dcd715, Jul 17, 2016.

  1. dcd715

    dcd715 Newbie

    Joined:
    Jul 17, 2016
    Messages:
    2
    Likes Received:
    2
    Hi All, please excuse this somewhat dumb question but I hope someone can clarify for me.

    If an audio file is allegedly recorded at 16bit 44.1khz stereo, I have noticed with numerous software packages that when inspecting the separate tracks of the file it states that the frequency for each channel is 22.050khz, this I understand as 22.050 x 2 = 44.100khz, however what happens to the bit rate? as if the file is 32bit or 32bit floating (24bit padded in reality) the software shows 16bit for each channel. Does this therefore mean that if a file is 16bit 44.1khz that each channel is in reality 8bit 22.050khz. If so I am correct in thinking that in reality each channe sample only has 256 combinations and not 65,536 or have I completely missed something. Thanks in advance.
     
  2.  
  3. kimikaze

    kimikaze Platinum Record

    Joined:
    Aug 20, 2014
    Messages:
    346
    Likes Received:
    158
    No, It is a Nyquist–Shannon sampling theorem. Which in short mean, sampling frequency need to be more than twice of frequency you want to reproduce. So we want to reproduce around 22khz, which is limit for human hearing, so by theorem we need at least 44.1khz sample rate. High resolution audio converters for example can reproduce 48khz of audio frequency spectrum, which mean sample rate must be at least 96khz. So when you see 44.1, 48, 96, 192khz this is sampling rate frequency and like in your case, when you see 22khz, software show you actual frequency range of audio not sampling rate of file. And this is for both chanel, so for example 44.1khz, 16 bit mean 44.1, 16bit for each chanel if is stereo.
     
    Last edited: Jul 17, 2016
  4. junh1024

    junh1024 Rock Star

    Joined:
    Jul 28, 2011
    Messages:
    1,396
    Likes Received:
    433
    If you look at the FFT of a 44k file, sensible software will display up to 22k, because you need 44k, to store a 22k signal. Everything above 22k is not useful for analysis.

    http://whatis.techtarget.com/definition/Nyquist-Theorem

    This doesn't apply to bitdepth.

    44k/16b/2ch = 44k/16b for EACH channel.
     
  5. Baxter

    Baxter Audiosexual

    Joined:
    Jul 20, 2011
    Messages:
    3,915
    Likes Received:
    2,755
    Location:
    Sweden
    If it is 44.1kHz and 16bit then BOTH channels of a stereo file are 44.1kHz and 16bit. :)

    To have a cycle at 22050Hz you need twice the frequency (positive and negative value), so in theory it's not a 22050Hz sine wave but rather a square wave (hence the aliasing/degradation/harmonics, etc).
     
    Last edited: Jul 17, 2016
    • Agree Agree x 3
    • Like Like x 1
    • List
  6. dcd715

    dcd715 Newbie

    Joined:
    Jul 17, 2016
    Messages:
    2
    Likes Received:
    2
    Thank you everyone for your speedy response, I have a better understanding now, best regards.
     
  7. LEGEND

    LEGEND Guest

    You're not welcome
     
  8. Daz

    Daz Guest

    Making suggestions to yourself is the first sign of madness LEGEND...



    And why all the dislikes :dunno:
     
    Last edited by a moderator: Jul 17, 2016
    • Like Like x 1
    • Agree Agree x 1
    • List
  9. foster911

    foster911 Guest

    What's exactly your problem? Not joking! Why are you trying to antagonize all of the members? Are you writhing in some pains? People here may leech and remedy your incurable illness.
     
Loading...
Loading...