Sorenson Video 3: Difference between revisions

From MultimediaWiki
Jump to navigation Jump to search
(Some more info)
(Some info about codes decoding)
Line 18: Line 18:
* 1-7 - inter 16x16 blocks, type specifies prediction direction
* 1-7 - inter 16x16 blocks, type specifies prediction direction
* 8 and 33 - intra 4x4 block (type 33 means that there are no coefficients for this block)
* 8 and 33 - intra 4x4 block (type 33 means that there are no coefficients for this block)
* all other values are for inter 4x4 blocks, type is used to set prediction direction and coded block pattern
* 9-32 - intra 16x16 blocks, type is used to set prediction direction and coded block pattern
In any case coefficients are stored in 4x4 (sub)blocks.


Dezigzag pattern (from [[H.264]]):
Dezigzag pattern (from [[H.264]]):
Line 29: Line 30:
     /
     /
   o-->o-->o-->o
   o-->o-->o-->o
=== Coefficient decoding ===
Each coefficient is stored as Golomb codeword, last bit is coefficient sign, code = 0 means end of nonzero coefficients.
Codes for block type 3:
{| border="1" cellpadding="5" style="border-collapse: collapse; border-style: dashed; border-color: #2f6fab;"
|- bgcolor="#f0f0f0" |
! code !! run !! value
|-
| 0-2
| 0
| '''code'''
|-
| 3
| 1
| 1
|-
| 4-...
| '''code''' & 0x3
| (('''code''' + 9) >> 2) - '''run'''
|}
Codes for intra block types:
{| border="1" cellpadding="5" style="border-collapse: collapse; border-style: dashed; border-color: #2f6fab;"
|- bgcolor="#f0f0f0" |
! code !! run !! value
|-
| 0 || 0 || 0
|-
| 1 || 0 || 1
|-
| 2 || 1 || 1
|-
| 3 || 0 || 2
|-
| 4 || 2 || 1
|-
| 5 || 0 || 3
|-
| 6 || 0 || 4
|-
| 7 || 0 || 5
|-
| 8 || 3 || 1
|-
| 9 || 4 || 1
|-
| 10 || 1 || 2
|-
| 11 || 1 || 3
|-
| 12 || 0 || 6
|-
| 13 || 0 || 7
|-
| 14 || 0 || 8
|-
| 15 || 0 || 9
|-
| 16-...
| '''code''' & 0x7
| ('''code''' >> 3) - intra_run['''run''']
|}
Codes for inter block types:
{| border="1" cellpadding="5" style="border-collapse: collapse; border-style: dashed; border-color: #2f6fab;"
|- bgcolor="#f0f0f0" |
! code !! run !! value
|-
| 0 || 0 || 0
|-
| 1 || 0 || 1
|-
| 2 || 1 || 1
|-
| 3 || 2 || 1
|-
| 4 || 0 || 2
|-
| 5 || 3 || 1
|-
| 6 || 4 || 1
|-
| 7 || 5 || 1
|-
| 8 || 0 || 3
|-
| 9 || 1 || 2
|-
| 10 || 2 || 2
|-
| 11 || 6 || 1
|-
| 12 || 7 || 1
|-
| 13 || 8 || 1
|-
| 14 || 9 || 1
|-
| 15 || 0 || 4
|-
| 16-... (inter)
| '''code''' & 0xF
| ('''code''' >> 4) - inter_run['''run''']
|}
Run correction values:
  intra_run = { 8, 2, 0, 0, 0, -1, -1, -1, [minus  ones] };
  inter_run = { 4, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, [zeroes] };


=== Macroblock transform and dequantization ===
=== Macroblock transform and dequantization ===

Revision as of 01:49, 29 April 2007

Video codec apparently based on an early H.264 draft.

The decoder in FFmpeg produces heavy artifacting for some uncommon samples.

Decoding Process

This codec extensively uses Golomb coding.

Slice Header

TODO

Macroblock layer

Each macroblock starts with Golomb code signalling MB type:

  • 0 - skip block
  • 1-7 - inter 16x16 blocks, type specifies prediction direction
  • 8 and 33 - intra 4x4 block (type 33 means that there are no coefficients for this block)
  • 9-32 - intra 16x16 blocks, type is used to set prediction direction and coded block pattern

In any case coefficients are stored in 4x4 (sub)blocks.

Dezigzag pattern (from H.264):

 o-->o-->o   o
         |  /|
 o   o   o / o
 | / |   |/  |
 o   o   o   o
   /
 o-->o-->o-->o

Coefficient decoding

Each coefficient is stored as Golomb codeword, last bit is coefficient sign, code = 0 means end of nonzero coefficients.

Codes for block type 3:

code run value
0-2 0 code
3 1 1
4-... code & 0x3 ((code + 9) >> 2) - run

Codes for intra block types:

code run value
0 0 0
1 0 1
2 1 1
3 0 2
4 2 1
5 0 3
6 0 4
7 0 5
8 3 1
9 4 1
10 1 2
11 1 3
12 0 6
13 0 7
14 0 8
15 0 9
16-... code & 0x7 (code >> 3) - intra_run[run]

Codes for inter block types:

code run value
0 0 0
1 0 1
2 1 1
3 2 1
4 0 2
5 3 1
6 4 1
7 5 1
8 0 3
9 1 2
10 2 2
11 6 1
12 7 1
13 8 1
14 9 1
15 0 4
16-... (inter) code & 0xF (code >> 4) - inter_run[run]

Run correction values:

 intra_run = { 8, 2, 0, 0, 0, -1, -1, -1, [minus  ones] };
 inter_run = { 4, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, [zeroes] };

Macroblock transform and dequantization

Transform coefficients:

 13  17   1   7
 13   7  -1 -17
 13  -7  -1  17
 13 -17   1  -7

Dequantization is performed by multiplying every coefficient by the same value determined by quantizer. In case if inter blocks first coefficient may be quantized slightly differently:

For some intra blocks (TODO: condition):

   dc = 169 * 1538 * block[0]

For chroma blocks (?):

   dc = (svq3_dequant_coeff[Q] * (block[0] >> 3)) >> 1;

Quantizer table (from svq3.c)

 static const uint32_t svq3_dequant_coeff[32] = {
   3881,  4351,  4890,  5481,  6154,  6914,  7761,  8718,
   9781, 10987, 12339, 13828, 15523, 17435, 19561, 21873,
  24552, 27656, 30847, 34870, 38807, 43747, 49103, 54683,
  61694, 68745, 77615, 89113,100253,109366,126635,141533
 };

Dequantization formula (dc=0 if not defined otherwise):

 out = (coeff * svq3_dequant_coeff[Q] + dc + 0x80000) >> 20;