Mitch Capper
eb005e8b9c
FIFO support for zoneminder zone debugging ( #2594 )
...
Adds fifo options for diagnostic images for much lower impact diagnostics mode. Diagnostic images are only written when there is a client listening for them (otherwise they are skipped). Also added a json stream for the detection data so you can see in real time the pixels or blobs detected for the motion. This allows for easy real time stream of both delta and reference images (as video streams) along with the detection numbers.
2019-05-16 15:37:03 -04:00
Isaac Connor
443fd35d99
Fix imagesize requirements ( #2404 )
...
* introduce non-loop-unrolled version of function and use them when the image size is not a multiple of 12 or 16
* Remove tests for imagesize being a multiple of 16 or 12 to handle functions with unrolled loops
* Use non-unrolled functions when image size is not a multiple of 12 or 16
* use std_blend when image is odd size
2019-02-22 09:44:57 -05:00
Isaac Connor
0f14df64e2
google code style and whitespace
2018-10-22 10:41:27 -04:00
Andrew Bauer
ba10c431f7
check for __ARM_PCS_VFP instead of __armel__
2018-10-12 09:34:16 -05:00
Andrew Bauer
ad70648cb3
don't build neon instructions on armel ( #2246 )
2018-10-10 14:10:02 -04:00
Isaac Connor
89cd7ceaf8
Whitespace, extra parentheses, google code style
2018-09-23 16:47:06 -04:00
Isaac Connor
62f8f5cfdb
whitespace
2018-05-03 13:54:18 -04:00
Isaac Connor
8518278476
coverity scan fixes/cleanups
2018-04-12 11:29:35 -04:00
Isaac Connor
a1def1dfe9
whitespace
2018-04-03 12:23:54 -07:00
Isaac Connor
6cfb550cf4
add cast to remove compile warning
2018-03-20 14:49:01 -04:00
Isaac Connor
599848346e
close infile on error condition. Make failure to read not be fatal.
2017-12-12 13:35:19 -05:00
Isaac Connor
7278448b01
remove test for less than zero of uints
2017-12-12 13:30:58 -05:00
Isaac Connor
b246ae4d19
remove less than zero tests for unsigned parameters
2017-12-12 13:25:02 -05:00
Isaac Connor
46094fe6cf
close outfile on error
2017-12-12 13:18:57 -05:00
Isaac Connor
43e4421368
fix strncpy not leaving null terminator space
2017-12-12 12:52:20 -05:00
Isaac Connor
64816aa005
fixes from cppcheck.
2017-11-18 16:00:10 -05:00
Isaac Connor
159c0a7165
moer cppcheck fixes
2017-11-17 07:52:26 -05:00
Isaac Connor
3227599e01
google code style
2017-09-25 16:22:52 -04:00
Isaac Connor
08bdf5a729
fix build on trusty
2017-08-24 12:19:57 -04:00
Isaac Connor
a81ff85fbb
add ffmpeg decoding of a .mp4 to get the frames
2017-08-23 15:05:44 -04:00
Isaac Connor
536ccc0cec
maybe fix memleak, maybe fix crash
2017-08-10 15:44:20 -04:00
Isaac Connor
9bd6c99586
Merge branch 'master' into storageareas
2017-05-15 12:52:35 -04:00
Andrew Bauer
1e45146db8
Merge pull request #1881 from mastertheknife/aarch64_neon
...
Add AArch64 ARM Neon code (complements #1823 )
2017-05-14 07:49:19 -05:00
Kfir Itzhak
5df7499da4
Move ARM Neon data prefetches to be after the loads
2017-05-13 19:08:18 +03:00
Kfir Itzhak
c66deb564f
Added ARM Neon AArch64 versions of delta
2017-05-13 09:19:13 +03:00
Andrew Bauer
8315be69de
Update zm_image.cpp
...
remove extra }
2017-05-12 18:46:19 -05:00
Kfir Itzhak
ec2b6235a0
Added ARM Neon AArch64 version of fastblend
2017-05-12 22:59:08 +03:00
Isaac Connor
1ddf4e32cf
Merge branch 'master' into storageareas
2017-05-11 08:55:42 -04:00
Kfir Itzhak
0d864b2c6c
Fix delta self-test introduced in #1823 failing ( #1878 )
...
* Fix self-test introduced in #1823 failing and improve logging about the failure
* Remove unnecessary newline added by previous commit
2017-05-11 07:13:04 -04:00
Isaac Connor
fd44ff4e60
Merge branch 'master' into storageareas
2017-05-10 13:16:08 -04:00
Kfir Itzhak
0584454aa8
Fix conflicts in Image::Initialise() caused by merging #1810 and #1823
2017-05-10 17:37:25 +03:00
Andrew Bauer
c4f471d967
Merge branch 'master' of https://github.com/abishai/ZoneMinder into abishai-master
2017-05-10 08:06:48 -05:00
Andrew Bauer
ccc2cc832f
Merge branch 'armv7_neon' of https://github.com/mastertheknife/ZoneMinder into mastertheknife-armv7_neon
2017-05-10 07:45:58 -05:00
Andrew Bauer
747583e9f4
Merge pull request #1810 from mastertheknife/ssse3_improvements
...
Minor zm_image SSSE3 improvements
2017-05-10 07:06:33 -05:00
Isaac Connor
6827e0412c
white and google style
2017-04-28 10:34:48 -04:00
Kfir Itzhak
71e6735b60
Neon32 functions now work on 64 bytes at a time. This results in 4-6x performance increase over standard functions
...
Memory allocations and image size requirements changed to be as needed for 64 byte alignment.
Self-test code for Blend modified accordingly and added Self-test for the delta functions.
2017-04-16 10:57:37 +03:00
Isaac Connor
4215f227da
Merge branch 'mastertheknife-ssse3_improvements' into storageareas
2017-04-07 13:43:09 -04:00
Isaac Connor
6ec340d525
Merge branch 'ssse3_improvements' of https://github.com/mastertheknife/ZoneMinder into mastertheknife-ssse3_improvements
2017-04-07 13:38:50 -04:00
Isaac Connor
f2dcd18297
Merge branch 'armv7_neon' of https://github.com/mastertheknife/ZoneMinder into storageareas
2017-03-30 13:46:30 -04:00
Kfir Itzhak
7162c62d26
Implemented a fix to get the Neon code to compile successfully with GCC < 6.0
2017-03-25 21:49:15 +03:00
Isaac Connor
a33a8402ef
Merge branch 'armv7_neon' of https://github.com/mastertheknife/ZoneMinder into neon
...
Conflicts:
src/zmf.cpp
src/zms.cpp
2017-03-24 12:14:16 -04:00
Kfir Itzhak
3431cf7732
Reduce code duplication in the SSSE3 functions: a single function for delta and a single function for RGB32->grayscale convert
2017-03-22 11:38:15 +02:00
Kfir Itzhak
29885135a6
Corrected issues in last two ARM Neon commits
2017-03-19 13:39:24 +02:00
Kfir Itzhak
510f65c414
Added ARM Neon (AArch32) FastBlend function
2017-03-19 13:31:44 +02:00
Kfir Itzhak
c5c9b52d18
Added ARM Neon (AArch32) Delta functions for 32bit color and 8bit grayscale
2017-03-19 13:02:42 +02:00
Kfir Itzhak
f7fcab24c4
Fixed a mistake in recent commit e7a681b8ff
2017-03-14 22:46:36 +02:00
Kfir Itzhak
f30192e5da
Use standard memcpy() on x86-64, as it behaves similar performs the equal or better
2017-03-14 22:41:35 +02:00
Kfir Itzhak
67f7ad40ae
Removed SSSE3 deinterlacing functions, as they were typically equal or slower than the standard code (when compiled with -O2 or better)
...
The function is too complicated to be vectorized efficiently
2017-03-14 22:36:26 +02:00
Kfir Itzhak
e7a681b8ff
Added BGRA, ARGB and ABGR variants of SSSE3 32bit color to grayscale conversion and made the code actually use them all
2017-03-14 22:29:55 +02:00
Kfir Itzhak
20604c2150
SSSE3 rgba->gray8 convert function now uses less instructions and is 10-20% faster
2017-03-14 16:54:00 +02:00
