• State
  • Locked Locked
  • Replies 4 replies
  • Subscribers 30 subscribers
  • Views 288 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

PCM3070 - log2 details

David Alexandrou
Intellectual 410 points
Other Parts Discussed in Thread: PCM3070, TAS3204

Hey there, 

With the Home Audio version of Purepath Studio, the help file for the LOG2 component doesn't show what the accuracy is for the PCM3070, only for other chips.

Additionally, it suggests typically an approx 12.bit accuracy, denoting E-04, but is that a true 4 decimal places, or is it only accurate to .0125 like the previous DSP architectures?

Cheers

David

  • 0
    TI__Prodigy 20 points

    David,

    The acuracy for the is exactly as for the AIC3254 which is noted on the help file. It is accrurate upto 4 decimal places for the given input data range of (0.25 to 1).

     

    Best Regards,

    Mithra

  • 0 in reply to Mithra Weerakoon
    Intellectual 410 points

    Heya, 

    Thanks for the reply - just to absolutely clarify because I had this problem with the tas3204 - It's four decimal places - is the resolution of that accuracy .0001 steps?

    Cheers

    David.

  • 0 in reply to David Alexandrou
    TI__Prodigy 20 points

     Hi David,

    The accuracy of the result is up to 1.1E-4. The log2 is calculated in software using approximations, and the accuracy is good up to the above specified number, based on the valid input values as specified in the help file.  However the resolution of the convergence of the answer may not be exactly 0.0001 steps from one input value to the other.  It will depend on the input value you are tying to find the log2 of.

    Best Regards,

    Mithra

  • 0 in reply to Mithra Weerakoon
    Intellectual 410 points

    Mithra, 

    My hope is just that it has smaller steps than the TAS chips - the log approximation on TAS chips had a minimum step value of 0.0125.I found that .0125 was still large enough to generate ripples of 0.3dB in the signal, which were audible as sparkle in some conditions.

    So really, as long as the log function is CAPABLE of 0.0001 differences, not that it's guaranteed from value to value, I will be happy.

    Thanks

    David.