In the mood 4 sharing. My 4 VST as OSM's 4 U :)

[tester]

Interesting thing about these oversampling methods. Tor - do you know how to implement these shaping methods to not oversampled signals? And do you know other noise shaping methods and how to implement them?

[tester]

Oversample x4 inside. I must say that there are audible differences between OS4 and others, I hear them also on polyphase resampling.

Also found an interesting talk on dithering and some noise shaping:
http://synthmaker.co.uk/forum/viewtopic.php?f=4&t=6568

[tester]

And one more thing. In peak filter, it seems that Q should be equal Q/4. Whatever comparisons I do with external filters - all SM/FS implementations of biquads have that difference in Q factor. Have no idea why.

[tester]

I wonder how will sound 8x oversampled monster.
I asume that the same type of cascading US/DS should follow?

[tor]

I never noticed the Q factor you mention but I will check it out. One more thing your 'e' limiter can be optimized a little.

Code: Select all

e = ((b-d)>=0)&(b-d)+((b-d)<0)&1e-10;

Code: Select all

bd=b-d;
e = ((bd)>=0)&(bd)+((bd)<0)&1e-10;

[tester]

Unoptimized version of 8x oversampled filters. Can someone tell if this is correct setup?

As for optimization - the up/downsamplers can be pushed through mono4 in a different way. Right now - the 2nd pass of filtering just goes through all, but the up/down sampling can be processed separately, thus - I guess - saving some CPU I guess. Need to check that.

Meanwhile I think I add 16x oversampled version, to allow full tests.

[tor]

looks all good to me. have not tested though.

[tor]

Dozius made this one. Maybe used to save CPU

[tester]

...and how to use it correctly in OS higher than 2?
I mean - how to get these all A,B,C... nodes for filter?
And more important - what comes out of these nodes? (in terms of spectral range? some sort of splits of spectra?)

[tor]

There you have it