Pink Noise Dithering - filmgrain of audio

Discussion in 'Audio Science' started by cameng318, Aug 18, 2023.

  1. cameng318

    cameng318 Friend

    Pyrate
    Joined:
    Jun 14, 2020
    Likes Received:
    330
    Trophy Points:
    63
    Location:
    Wyoming
    Got higer order noise shaper figured out. It's actually a lot easirer than I thought for PCM signals. For 5th order noise shaper, its truncation resistance is about 5bit:
    ns5.jpg
    So the pattern is, as long as the white noise portion doesn't get above the slope noise, whether the slope is at high or low frequency, the quantization noise won't be distinguishable.

    However, I still didn't like the sound of this 5th order noise shaper on my Gungnir Multibit. At 4x OS, the shaper demands a 5th order filter at 30kHz, which I don't think my DACs are that aggressive. The result sound is a little digitus and robotic. 9th order noise shaper sounds even worse on my setup.

    Therefore, it's another advantage of pink noise dithering, that it is much less demanding on the analog filters. Aliasing doesn't show up until 1.5~2x of the nyquist frequency.
     
  2. Garns

    Garns Friend

    Pyrate
    Joined:
    Jul 9, 2016
    Likes Received:
    2,484
    Trophy Points:
    93
    Location:
    Sydney, AUS
    Nice, I'll download this and give it a listen at the weekend!
     

Share This Page