Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Do we have dither + noise shaping dsp?  (Read 1602 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Do we have dither + noise shaping dsp?

Hi, at times, in order to keep file sizes reasonable, I reduce bit depth from my source (24b) flacs. So e. g. they end up with 20 bits per sample. Now we have ditherers with eg triangular pdf, but do we have noise shaping? Eg like that consortiums algos pow-r1 - 3? If we had it, I'd probably use r3 for my mostly classical high dynamic range content.
Any plugin I'm not aware of?

I use mda dither for the dithering.

Thinking about looking into the vst domain, but I'm not sure how appropriate it is for the converter (thinking of clicks at start/end, or maybe problems with gapless)

thx

Re: Do we have dither + noise shaping dsp?

Reply #2 –
Built-in "Dither" feature in converter is dither+noise shaping actually. But it can output 16 bit only, not 20. It uses shibata noise shaping algorithm. You can see its curve here - https://web.archive.org/web/20220524120829/https://sox.sourceforge.net/SoX/NoiseShaping

POW-R is patented, so you won't find it in free software like fb2k

Quote from: Squeller on
Maybe tpdf dither is just enough.
I doubt it is possible to hear differences between simple TPDF and different noise shaping algorithms in 16 bit under normal listening conditions. But if you want compare their "sound", just convert to 8 bit and listen.

Re: Do we have dither + noise shaping dsp?

Reply #3 –
TPDF dither should be the recommended method. At 20 or even 16 bits per sample noise shaping doesn't offer audible benefits, only causes heavy noisiness that makes compression harder.
And I'd recommend using my Smart Dither. No point in turning digital silence into noise IMO, unless you want to simulate cassette.

Re: Do we have dither + noise shaping dsp?

Reply #4 –
@Bogozo, yes inaudible, probably. This is about best possible outcome under given conditions.

@Case. My understanding is, dithering already results in white noise and shaping moves it into areas of less sensitive hearing... And right, noise increasing file size in lossless compression is a point.

Gotta add, usually I'm not dealing with digital silence 😁 I see, the smart part of smartdither is the detection of digital silence, right?
 

Re: Do we have dither + noise shaping dsp?

Reply #5 –
Quote from: Bogozo on
But if you want compare their "sound", just convert to 8 bit and listen.
Just noticed that a) mda dither has a noise shaping setting and b) allows reviewing the result easily via its "zoom" feature.

Noise shaped noise is indeed less audible. I will make some flac file size comparisons and probably just use that.

Re: Do we have dither + noise shaping dsp?

Reply #6 –
Perhaps your source is DSD then, otherwise digitally sourced tracks almost always have silence and if it's intentionally added, it's perfect silence without noise.
I remembered dither has been a topic on Hydrogenaudio. Googling for "optimal dither 2Bdecided" brought me this thread where old time experts also say TPDF is the best choice.

Re: Do we have dither + noise shaping dsp?

Reply #7 –
After auditioning (in Reaper, in fb2k's mda dither plugin) different scenarius, I decided to use Dithering (TPDF + HP) followed by noise shaping, where the latter makes noise much less audible (in inaudible levels of course, the higher the bit depth).

Example, noise freq. distribution and volume at 20 bps after dither+noise shaping (of a silent source):

X