Hybrid Series Matrices

[*.mp4 guy]

Second version of my previous matrix, preliminary testing shows a compressability gain plz give me your feedback, thank you.

8,8,10,12,16,19,22,27
8,10,12,16,19,22,27,33
10,12,16,19,22,27,33,37
12,16,19,22,27,33,37,41
16,19,22,27,33,37,41,45
19,22,27,33,37,41,45,61
22,27,33,37,41,45,61,100
27,33,37,41,45,61,99,173

17,18,20,21,22,24,26,30
18,22,23,23,25,27,30,32
20,23,24,26,28,30,33,34
21,23,26,29,32,34,35,38
22,25,28,32,35,37,40,43
24,27,30,34,37,43,46,48
26,30,33,35,40,46,52,53
29,32,34,38,43,48,53,64

Hybrid8AQ-V2 Beta (http://uploadhut.com/view.php/93910.xcm)

[niamh]

I'll try it today, results tomorrow ... sounds good!

[*.mp4 guy]

The Reduced ringing matrix blocked to much, So it is no longer the reduced ringing matrix, it is now the kind of reduced ringing matrix, that doesn't block as much and will probably work well for 1 and 2 CD encodes.

8,9,13,19,27,32,36,41
9,12,18,26,30,34,39,44
13,18,26,30,34,39,42,47
19,26,30,34,39,42,45,50
27,30,34,39,42,45,48,54
32,34,39,42,45,48,52,59
36,39,42,45,48,52,57,64
41,44,47,50,54,59,64,74

17,18,20,21,22,24,26,30
18,22,23,23,25,27,30,32
20,23,24,26,28,30,33,34
21,23,26,29,32,34,35,38
22,25,28,32,35,37,40,43
24,27,30,34,37,43,46,48
26,30,33,35,40,46,52,53
29,32,34,38,43,48,53,64

Hybrid LBR-V3 (http://uploadhut.com/view.php/94201.cqm)

[niamh]

Yes actually, I tried it and forgot to post back, it blocks :)
But it doesn't ring ;)

[Didée]

Originally posted by *.mp4 guy
it is now the kind of reduced rining matrix, that doesn't block as much
Ermh ... you improved on the blocking by letting the inter matrix unchanged, plus forcing stronger quantization on the intra matrix? Interesting.

I'd rather try to change the "convex" distribution of coefficients in intra's first quadrant into a more "flat" or even "concave" one, but well ...