This is the official homepage of Rawshack, a raw image analyzer utility. Rawshack reports a variety of satsitical information about a raw image file, including pixel count data that can be used to generate your own histogram using external programs like Microsoft Excel. Rawshack is an open source application, licensed under GPL v3 and uses the LibRaw Library. It supports most bayer-filter sensor cameras.
| Date | Ver | Change List | Source Code | LibRaw Source | My LibRaw changes | Windows Binary (32-bit) | Windows Binary (64-bit) | Mac OSX Binary (64-bit) | Linux Binary (64-bit) |
| 06/14/2017 | 1.06 | changelist | source (zip) | LibRaw-0.18.2.zip | LibRaw Patch File | rawshack32.exe | rawshack.exe | rawshack | rawshack |
| 06/09/2017 | 1.05 | changelist | source (zip) | LibRaw-0.18.2.zip | LibRaw Patch File | rawshack.exe | rawshack | rawshack | |
| 06/05/2017 | 1.04 | changelist | source (zip) | LibRaw-0.18.2.zip | LibRaw Patch File | rawshack.exe | rawshack | rawshack | |
| 01/22/2014 | 1.03 | changelist | source (zip) | LibRaw-0.16.0.zip | LibRaw Patch File | rawshack.exe | N/A | N/A | |
| 04/19/2001 | 1.02 | changelist | source (zip) | LibRaw-0.13.4.zip | LibRaw Patch File | rawshack.exe | N/A | N/A | |
| 04/18/2011 | 1.01 | changelist | source (zip) | LibRaw-0.13.4.zip | LibRaw Patch File | rawshack.exe | N/A | N/A | |
| 04/16/2011 | 1.00 | Initial Release | rawshack.cpp | LibRaw-0.13.4.zip | LibRaw Patch File | rawshack.exe | N/A | N/A |
rawshack [optional args] filename(s)
You can specify one or more filenames to be processed. You can also use a wildcard specifier, which directs Rawshack to process every file whose name matching the specification. For example, "c:\mypics\*.cr2" will process all files in the c:\mypics directory that have a .cr2 extension. You can use multiplie wildcard specifications, like "c:\mypics\*.cr2 c:\mypics\*.nef", to process all .cr2 and .nef files in the c:\mypics directory.
--help
This help display
--blacksubtraction
Performs black-level bias subtraction of image data. Some cameras, like Canons, apply a black offset to the analog signal data prior to quantization through the ADC (Analog-to-Digital converter). This allows for the recreation of signal data at the very bottom of the noise floor, which is important for certain applications like astrophotography. If --blacksubtraction is not specified, the original ADU (Analog-to-Digital Unit) values will be analyzed, which includes the analog bias. If --blacksubtraction is specified, the ADU values above the black level will be reduced by the black level, and all ADU values equal to or below the black level will be reset to zero.
--clippinglevels=default | channelmax | bitdepth | r,g,b,g2
Set raw clipping levels. The clipping levels determine which pixel values are interpreted by Rawshack as "clipped" or fully saturated. Having the correct clipping levels is essential for the proper reporting of Rawshack's clipping statistics. If the --clippinglevels option is not specified, Rawshack will use the clipping levels it has defined for the camera in its internal database of cameras. If there is no database entry for that camera, or if the channel maximums for the specific raw image are higher than the levels specified in the database entry then Rawshack will use the channel maximums. "Channel maximums" simply means the highest pixel value that was found for each of the 4 bayer color channels.
You can direct Rawshack to always ignore the internal database by using "channelmax" (always use the channel maximums for clipping), "bitdepth" (use the maximum value that can be represented by the bitdepth of the raw as the clipping value, provided the bitdepth was available for the raw). You can also directly specify the absolute clipping values for the channels. For example, "--clippinglevels=15580,15580,15575,15400" will set the clipping level of the red and green channels to 15580, the blue channel to 15575, and the the green_2 channel to 15400.
--dimensions=x,y,width,height
Set dimensions of image subset to analyze. By default Rawshack will analyze the entire contents of the image, net of any camera-specific border/masked areas. This option lets you direct Rawshack to analyze only a subset of the image area.
--nostatsfile
By default, Rawshack generates the analyzed statistics of each raw into a file named <rawfilename>.txt. For example, if your raw was named dsc_0092.nef, the statistics file will be named dsc_0092.txt. Specifying --nostatsfile disables the generation of this file, directing the output to stdout instead (console screen).
--histogramdata
Rawshack can optionally generate a separate output file that contains the count of pixel values at each ADU. This data can then be used via an external program like Microsoft Excel to generate histograms on a per-channel or channel-combined basis. The name of the generated file is <rawfilename>.hist.txt. The output is a comma-delimited ASCII file with a legend on the first row, followed by the pixel values for each ADU. Here is a sample output:
ADU, Red, Green, Blue, Green_2
00000,00000000,00000000,00000001,00000000
00001,00000015,00000008,00000013,00000005
00002,00000027,00000012,00000033,00000009
...
16383,00002694,00000000,00003480,00000000
The first row the legend. Each subsequent row has the ADU in the first column, which starts from 0 and ends at either the higher of all channel maximums or configured clipping levels, whicheveri s greater. The next 4 columns describe the count of pixels found for that ADU in each color channel.
--outputinsrcdir
By default, Rawshack stores the output files (statistics, histogram data, etc...) in the current working directory. Specify the --outputinsrcdir option if you'd like the output files to be placed in the same directory as the raw file.
--sonycompdata=levels (V1.03)
Generates a comma-delimited ASCII file describing all 16x2 pixel blocks that are encoded with delta tonal lossy compression levels matching the levels specified for this option. There are 5 compression levels and any combintation of them can be specified in the levels option. Level 0 represents pixels encoded with lossless compression, Level 4 represents pixels with the most lossy compression (and varying tiers of lossy compression in between). For example, to report only pixels with the highest level of lossy compression (Level 4), use "--sonycompdata=4". To report pixels using the two most lossy compression levels use "--sonycompdata=3,4". If you'd like to report pixels for all levels (every pixel in raw), use "--sonycompdata=0,1,2,3,4". The default if no levels are specified is 3,4. See my dpreview thread here for more details on Sony's lossy compression and how this option can be used to detect image artifacts that may result from it.
The name of the generated file is <rawfilename>.sonycomp.csv and is formatted as follows:
First Line:
Sony Camera Model, LibRaw image width, LibRaw image height, Hex value describing --sonycompdata levels value (encoded bits)
All Subsequent lines:
Pixel X-coordinate, Pixel Y-coordinate, Compression Level of Pixel
Sample Data:
NEX-3N,4928,3276,0x18
3104,523,4
3105,524,3
3104,525,4
3105,526,4
A PNG of the resulting CSV file can be generated using my createsonypng PHP script (download here - save it as createsonypng.php from your browser). View the comments at the top of the script for instructions on how to install and configure PHP. The script will color-code each pixel based on its compressoin level. Level 4 = White, Level 3 = Red, Level 2 = Yellow, Level 1 = Blue, Level 0 = Green). Here is a sample PNG generated from the script.
--columngroupstats=#cols,... (V1.05)
--rowgroupstats=#rows,... (V1.05)
Generates a seperate set of stddev stats (mean, stddev, and COV) by binning pixel stats into groups of the specified number of columns (--columngroupstats) or rows (--rowgroupstats). For example, if the raw is 6000x4000 and you specify --columngroupstats=100, rawshack will geneate stats for 60 seperate column bins, each 100 pixels wide (pixels in columns 0..99, pixels columns 100..199, etc...). The resulting stats for each specified column or row group will be written into 4 files, named <rawfilename>.cols-xxx-yyy.txt (--columngroupstats) or <rawfilename>.rows-xxx-yyy.txt (--rowgroupstats), where 'xxx' is the nmber of pixels in the group and 'yyy' is the color channel (Red, Green, Blue, and Green_2). You can specify multiple column and row groups to be generated at the same time (ex: --rowgroupstats=64,32,16,8). Here is a sample of the output for one of the group stat files:
Mean,StdDev,COV
512.06,3.72,0.73%
512.01,3.55,0.69%
512.04,3.50,0.68%
--recurse (V1.06)
Recurses through subdirectories when performing a wildcard search. Subdirectories will only be searched for file arguments that contain a wildcard, such as *.NEF. When using this option on OSX and Linux, put quotes around the path, otherwise the shell will perform a wildcard replacement without passing the wildcards to rawshack. Example: "/MyVol/MyPics/*.arw" --recurse
--maxthreads (V1.06)
The maximum number of image processing threads. Default is the number of processor cores detected on the system (default is capped at 4 on the 32-bit Windows executable due to the 3GB memory constraint). Be sure you have sufficient physical memory to support the thread count you specify otherwise performance will degrade from OS virtual memory paging file thrashing - each raw file can require hundred megabytes of memory, so the total memory consumed will be #threads x #files.
--maxconcurrentreads (V1.06)
The maximum number of files that will be read simultaneously across image processing threads. The default is two, which prevents performance degradation from disk thrashing on spinning disk drives.
File: C:\develop\testraw\D7000_ISO100.NEF
Camera: NIKON D7000
Exposure/Params: ISO 100 f/8 1/500s 24mm
Bit Depth: 14-bit
Image dimensions: 4948x3280 (pixel count = 16,229,440)
Analyzed image rect: 0000x0000 to 4948x3280 (pixel count = 16,229,440)
Clipping levels src: Channel Maximums
Clipping levels: Red=16,383; Green=15,779; Blue=16,383; Green_2=15,779
Black point: 0
Min pixel value:
Red: pixVal= 3, count = 1, first coord: 4094x0018
Green: pixVal= 12, count = 1, first coord: 1175x2882
Blue: pixVal= 0, count = 1, first coord: 0001x1525
Green_2: pixVal= 13, count = 2, first coord: 1178x2883
Min non-zero pixel value:
Red: pixVal= 3, count = 1, first coord: 4094x0018
Green: pixVal= 12, count = 1, first coord: 1175x2882
Blue: pixVal= 5, count = 1, first coord: 0763x2439
Green_2: pixVal= 13, count = 2, first coord: 1178x2883
Max pixel value:
Red: pixVal= 16,383, count = 2,694, first coord: 4258x0420
Green: pixVal= 15,779, count = 2, first coord: 1151x0018
Blue: pixVal= 16,383, count = 3,480, first coord: 1151x0019
Green_2: pixVal= 15,779, count = 1, first coord: 1152x0017
Max non-clipped pixel value:
Red: pixVal= 16,381, count = 1, first coord: 2536x1646
Green: pixVal= 15,778, count = 4,540, first coord: 4247x0414
Blue: pixVal= 16,379, count = 1, first coord: 2537x1627
Green_2: pixVal= 15,778, count = 4,543, first coord: 2510x0167
Clipped pixel counts (individual channels):
Red: count = 2,694, first coord: 4258x0420
Green: count = 2, first coord: 1151x0018
Blue: count = 3,480, first coord: 1151x0019
Green_2: count = 1, first coord: 1152x0017
Clipped pixel counts (channel combinations) for 2x2 bayer pixel GRBG matrix:
None: 4053489, R: 389, G: 1, R+G: 0
B: 1174, R+B: 2305, G+B: 1, R+G+B: 0
G2: 1, R+G2: 0, G+G2: 0, R+G+G2: 0
B+G2: 0, R+B+G2: 0, G+B+G2: 0, R+G+B+G2: 0
Lowest pixel value bit set:
Red: bit= 0, count = 2,037,231, first coord: 0000x0000
Green: bit= 0, count = 2,021,157, first coord: 0001x0000
Blue: bit= 0, count = 2,037,761, first coord: 0001x0001
Green_2: bit= 0, count = 2,020,726, first coord: 0004x0001
Highest non-clipped pixel value bit set:
Red: bit=13, count = 1,680, first coord: 1150x0016
Green: bit=13, count = 9,570, first coord: 1151x0016
Blue: bit=13, count = 2,827, first coord: 1151x0017
Green_2: bit=13, count = 9,525, first coord: 1150x0017
Pixel average and standard deviation:
Red: Mean= 401.93, StdDev= 713.17, COV= 177.44%, Unique values= 7,481
Green: Mean= 768.60, StdDev= 1179.69, COV= 153.49%, Unique values= 11,461
Blue: Mean= 550.96, StdDev= 913.96, COV= 165.89%, Unique values= 8,861
Green_2: Mean= 770.56, StdDev= 1180.79, COV= 153.24%, Unique values= 11,484
Pixel counts for each raw stop:
Stop #/Range Red Green Blue Green_2
--------------- ---------- ---------- ---------- ----------
0 [00000-00001] 0 0 1 0
1 [00002-00003] 1 0 0 0
2 [00004-00007] 41 0 1 0
3 [00008-00015] 3,548 13 451 5
4 [00016-00031] 44,801 5,817 24,940 5,491
5 [00032-00063] 182,871 64,706 147,800 63,880
6 [00064-00127] 890,952 231,445 615,279 229,885
7 [00128-00255] 1,616,572 965,524 1,109,004 960,707
8 [00256-00511] 531,739 1,418,037 1,181,971 1,419,200
9 [00512-01023] 362,200 643,914 437,993 648,758
10 [01024-02047] 326,237 320,510 293,485 321,209
11 [02048-04095] 89,345 308,370 221,420 308,597
12 [04096-08191] 4,679 89,452 18,708 90,102
13 [08192-16383] 4,374 9,572 6,307 9,526
14 [16384-32767] 0 0 0 0
15 [32768-65535] 0 0 0 0
Rawshack - Raw Image Analyzer
Copyright (C) 2011; adam@testcams.com
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
If you have a feature suggestion or bug to report (which is likely, considering it's brand new!), please contact me at adam@testcams.com