ryan/mozjpeg
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

16-bit lossless JPEG support

Browse Source
This commit is contained in:
DRC
2022-12-15 12:20:50 -06:00
parent ad4e2ad86f
commit 2241434eb9
46 changed files with 906 additions and 148 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
structure.txt
View File

@@ -519,9 +519,9 @@ shown are:
is not used for full-color output. Works on one pixel row at a time; may
require two passes to generate a color map. Note that the output will
always be a single component representing colormap indexes. In the current
design, the output values are JSAMPLEs or J12SAMPLEs, so the library cannot
quantize to more than 256 colors when using 8-bit data precision. This is
unlikely to be a problem in practice.
design, the output values are JSAMPLEs, J12SAMPLEs, or J16SAMPLEs, so the
library cannot quantize to more than 256 colors when using 8-bit data
precision. This is unlikely to be a problem in practice.
* Color reduction: this module handles color precision reduction, e.g.,
generating 15-bit color (5 bits/primary) from JPEG's 24-bit output.
@@ -621,8 +621,16 @@ Arrays of 12-bit pixel sample values use the following data structure:
typedef J12SAMPROW *J12SAMPARRAY; ptr to a list of rows
typedef J12SAMPARRAY *J12SAMPIMAGE; ptr to a list of color-component arrays
The basic element type JSAMPLE (8-bit sample) will be unsigned char, and the
basic element type J12SAMPLE (12-bit sample) with be short.
Arrays of 16-bit pixel sample values use the following data structure:
typedef something J16SAMPLE; a pixel component value, 0..MAXJ16SAMPLE
typedef J16SAMPLE *J16SAMPROW; ptr to a row of samples
typedef J16SAMPROW *J16SAMPARRAY; ptr to a list of rows
typedef J16SAMPARRAY *J16SAMPIMAGE; ptr to a list of color-component arrays
The basic element type JSAMPLE (8-bit sample) will be unsigned char, the basic
element type J12SAMPLE (12-bit sample) will be short, and the basic element
type J16SAMPLE (16-bit sample) will be unsigned short.
With these conventions, J*SAMPLE values can be assumed to be >= 0. This helps
simplify correct rounding during downsampling, etc. The JPEG standard's
@@ -633,7 +641,8 @@ decompression the output of the IDCT step will be immediately shifted back to
When 8-bit samples are in use, the code uses MAXJSAMPLE and CENTERJSAMPLE,
which are defined as 255 and 128 respectively. When 12-bit samples are in use,
the code uses MAXJ12SAMPLE and CENTERJ12SAMPLE, which are defined as 4095 and
2048 respectively.)
2048 respectively. When 16-bit samples are in use, the code uses MAXJ16SAMPLE
and CENTERJ16SAMPLE, which are defined as 65535 and 32768 respectively.)
We use a pointer per row, rather than a two-dimensional J*SAMPLE array. This
choice costs only a small amount of memory and has several benefits: