Announcement

Collapse
No announcement yet.

RIAA DIfferent in 96 KHz vs 44.1 KHz

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Guys,

    It does look like there is a +6dB gain difference in the VPA at 96kHz vs 44.1 and 48k. It seems to be just an overall gain offset, not a different curve. I can fix this in this upcoming release.

    96k Click image for larger version

Name:	image_524.png
Views:	129
Size:	19.4 KB
ID:	53065


    48k Click image for larger version

Name:	VPP48k.png
Views:	118
Size:	19.7 KB
ID:	53066



    44.1k Click image for larger version

Name:	VPP441k.png
Views:	122
Size:	19.4 KB
ID:	53067

    The sample rate conversions are pretty darn good, especially if you do a master quality conversion. Almost no detectable difference except a slight roll off about 20kHz. The CD quality conversion does have more roll off.

    As for the time difference, there will be some difference. I have not looked into the numbers you have here to see if that is excessive or not.




    Rick
    Last edited by rickc; 02-19-2017, 12:12 PM.
    Rick

    Comment


    • #17
      Rick,

      ​Thanks for checking into the bug. By the way, how did you get those spectrum plots ?

      Marc

      Comment


      • #18
        Thanks Rick
        Would it be possible to save all three traces using DC10Forensics and show all 3 on one display? A direct comparison there is where I see the differences in the traces.
        D
        Dub: to mix (recorded sound or videotape from different sources) into a single recording

        Comment


        • #19
          To answer for Rick as to how he got those plots, it appears to me that he made a set of white noise files with the Make Waves Generator at various sample rates and then just loop played them through the VPP and into the Spectrum Analyzer in Averaging mode until it built the graphs. When I tested the VPP system many a year ago, I used the Make Waves generator using one tone at a time across each sample rates and each EQ curves (5 primary curves). My method is more accurate, but takes a lot more time. As said before, the 6 dB overall error is just an offset error. I am not in favor of fixing it because it will produce a huge amount of testing work and a good possibility for introducing new bugs into that algorithm.

          On a relative frequency basis, it is basically as accurate as any system could be. Absolute gain is not relevant and can be adjusted with the gain control.

          Look at it this way; your stereo has an RIAA phono input. Nobody knows its exact gain but for its relative gain at various inflection points on the EQ curve. So, who cares about the overall gain of your stereo so long as you can turn it up loud enough to fill your room with sound?

          Craig
          Last edited by Craig Maier; 02-19-2017, 07:08 PM.
          "Who put orange juice in my orange juice?" - - - William Claude Dukenfield

          Comment

          Working...
          X