• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Dithering is a Mathematical Process - NOT a psychoacoustic process.

JohnPM

Senior Member
Technical Expert
Joined
Apr 9, 2018
Messages
340
Likes
901
Location
UK
Any known issues or limitations with CoreAudio on macOS? That's where I'm using REW. So far I've not seen any problems with dither.
Nope, works very well and supports the full bit depth and all the channels the device offers.
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,201
Likes
16,983
Location
Riverview FL
What is ISP?

I keep reading it and can't assign a word to the "P", which makes me a bit unsure about the "I" and "S".
 

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,591
Likes
10,727
Location
Prague
Inter-Sample Peaks

Yes. Looks like this, happens when there is a fast amplitude change between samples and sample amplitude approaches FS.

1581695948871.png
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
Yes. Looks like this, happens when there is a fast amplitude change between samples and sample amplitude approaches FS.

Of course such signals can appear on recordings, but it maybe it should be kept in mind that's not a bandlimited square wave. :)

(Top is, bottom is a "naïvely" generated square wave.)

1581699274219.png
 
Last edited:

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
As for the question about old versions of Audition or the final version of Cool Edit Pro is reliable or not

Talking of "old" versions, the SRC Comparisons page is interesting.

Of course, I don't know/recollect if there's any settings that might have been missed in the old version of Audition. For Audacity version 2+ set to "Best Quality,", the sweep plot looks very similar to Audition 2019 Some of the results on that page are truly dire.

(N.B. Audition 5.5/CS6 SRC looks identical to Audition 2019.)


1581699748195.png
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
Talking of "old" versions, the SRC Comparisons page is interesting.

Of course, I don't know/recollect if there's any settings that might have been missed in the old version of Audition. For Audacity version 2+ set to "Best Quality,", the sweep plot looks very similar to Audition 2019 Some of the results on that page are truly dire.


View attachment 50188
Yes, Audition and Cool Edit (the first one I have used is 1.1) are good at these things, and good for simple destructive editing. After all, these are tools, people with talent and skills can do great things on Windows Paint, while others can create crap out out of Photoshop. It depends on how the tools are being used.
 

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
Yes, Audition and Cool Edit (the first one I have used is 1.1) are good at these things, and good for simple destructive editing. After all, these are tools, people with talent and skills can do great things on Windows Paint, while others can create crap out out of Photoshop. It depends on how the tools are being used.

I certainly don't have Paint pixel-by-pixel drawing skills but I use it all the time, it's actually pretty indispensable for rapid copy-pasting/cropping or basic text markup or quick diagrams. The power of Photoshop just gets in the way for that.

Microsoft wanted to "retire" it but there was a massive backlash.

OTOH, good luck with doing lens correction or auto perspective correction in Paint...

Use whatever gets the job done but one needs to watch out for DSP "boobie traps" in any and all software. :) Can't stand Audacity's UI, so I don't use it...
 
Last edited:

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Talking of "old" versions, the SRC Comparisons page is interesting.

Of course, I don't know/recollect if there's any settings that might have been missed in the old version of Audition. For Audacity version 2+ set to "Best Quality,", the sweep plot looks very similar to Audition 2019 Some of the results on that page are truly dire.

(N.B. Audition 5.5/CS6 SRC looks identical to Audition 2019.)


View attachment 50188
What is worse is some of those dire results are still in some DAW software and used to work on recordings. YIKES!

You can test any software you have and get graphs like this. Just do a log sweep, do the conversion of the sample rate and show the result in Audacity using spectrogram view and with the lower limit of the spectrogram set to -180 or -200 db.
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
You might not remember that CoolEdit came with a MS Visual C SDK back in the day where you just edited a plug-in "skeleton" and did a one click recompile. I wrote a plug-in to do a third order spline click repair that still works today.

Cool, but very different scenario to a commercial plug-in developer providing ongoing support for multiple platforms (VST, Audio Units, AAX, etc.) and various host software.

In case it's of any use to anyone, I've created a "bitshift" gain adjustment script for Reaper (JSFX.) By default, it turns out Reaper adds anti-denormal noise to the inputs of JSFX plug-ins, so since this is an exercise in bit-perfection, it is set to off.


Code:
desc: Volume - Bitshift

slider1:0<-48,48,6>Gain (dB)


in_pin:left input
in_pin:right input
out_pin:left output
out_pin:right output


@init
  ext_nodenorm = 1; // Input anti-denormal noise OFF

@slider
  // Convert dB to Bits
  bitshift = slider1 / 6;


  // Positive scale
  (bitshift > 0)?
  (
    splscl = (2 ^ bitshift);
  ) : (
      // Negative scale
      (bitshift < 0)?
      (
        splscl = 1 / (2 ^ - bitshift);
      ) : (
          // No scaling
          splscl = 1;
      );
  );


@sample
  // Scale
  spl0 = (spl0*splscl);
  spl1 = (spl1*splscl);
 
Last edited:

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK

bennetng

Major Contributor
Joined
Nov 15, 2017
Messages
1,634
Likes
1,692
Of course I remember your chip tune post. :)

View attachment 50274


And here it is again with brickwall dBTP limiting (threshold at 0dB, scaled on output by ~-0.1dB):

View attachment 50276

File:

https://we.tl/t-zsO9s9cswr

(Download link expires in one week.)

See whether or not the limiting has "ruined" the sound quality... or not. ;-)
But here is something I usually observed these days:
https://www.audiosciencereview.com/...lity-in-windows-using-wasapi.5272/post-117708
 

xr100

Addicted to Fun and Learning
Joined
Jan 6, 2020
Messages
518
Likes
237
Location
London, UK
e) what arbitrary frequency (~1KHz) hits the most 16 bit theoretical values (exercises all the levels @16/44.1) as that is the one I figure I want to use.

985.1Hz seems to work... well, at least over 8 seconds, the smallest absolute values are "exercised" twice, and the largest absolute values considerably more than 100 times. So that should give the LSB's a good workout? (Albeit even with a "multi-bit" DAC, the values would be changed by the "digital filter," incl. oversampling and noise shaping?)

8 second file.

[Download link expires in one week.]
 
Top Bottom