Recherche avancée

Sur d’autres sites (8010)

• Anatomy of an optimization : H.264 deblocking

  26 mai 2010, par Dark Shikari — H.264, assembly, development, speed, x264

  As mentioned in the previous post, H.264 has an adaptive deblocking filter. But what exactly does that mean — and more importantly, what does it mean for performance ? And how can we make it as fast as possible ? In this post I’ll try to answer these questions, particularly in relation to my recent deblocking optimizations in x264.

  H.264′s deblocking filter has two steps : strength calculation and the actual filter. The first step calculates the parameters for the second step. The filter runs on all the edges in each macroblock. That’s 4 vertical edges of length 16 pixels and 4 horizontal edges of length 16 pixels. The vertical edges are filtered first, from left to right, then the horizontal edges, from top to bottom (order matters !). The leftmost edge is the one between the current macroblock and the left macroblock, while the topmost edge is the one between the current macroblock and the top macroblock.

  Here’s the formula for the strength calculation in progressive mode. The highest strength that applies is always selected.

  If we’re on the edge between an intra macroblock and any other macroblock : Strength 4
  If we’re on an internal edge of an intra macroblock : Strength 3
  If either side of a 4-pixel-long edge has residual data : Strength 2
  If the motion vectors on opposite sides of a 4-pixel-long edge are at least a pixel apart (in either x or y direction) or the reference frames aren’t the same : Strength 1
  Otherwise : Strength 0 (no deblocking)

  These values are then thrown into a lookup table depending on the quantizer : higher quantizers have stronger deblocking. Then the actual filter is run with the appropriate parameters. Note that Strength 4 is actually a special deblocking mode that performs a much stronger filter and affects more pixels.

  One can see somewhat intuitively why these strengths are chosen. The deblocker exists to get rid of sharp edges caused by the block-based nature of H.264, and so the strength depends on what exists that might cause such sharp edges. The strength calculation is a way to use existing data from the video stream to make better decisions during the deblocking process, improving compression and quality.

  Both the strength calculation and the actual filter (not described here) are very complex if naively implemented. The latter can be SIMD’d with not too much difficulty ; no H.264 decoder can get away with reasonable performance without such a thing. But what about optimizing the strength calculation ? A quick analysis shows that this can be beneficial as well.

  Since we have to check both horizontal and vertical edges, we have to check up to 32 pairs of coefficient counts (for residual), 16 pairs of reference frame indices, and 128 motion vector values (counting x and y as separate values). This is a lot of calculation ; a naive implementation can take 500-1000 clock cycles on a modern CPU. Of course, there’s a lot of shortcuts we can take. Here’s some examples :

  • If the macroblock uses the 8×8 transform, we only need to check 2 edges in each direction instead of 4, because we don’t deblock inside of the 8×8 blocks.
  • If the macroblock is a P-skip, we only have to check the first edge in each direction, since there’s guaranteed to be no motion vector differences, reference frame differences, or residual inside of the macroblock.
  • If the macroblock has no residual at all, we can skip that check.
  • If we know the partition type of the macroblock, we can do motion vector checks only along the edges of the partitions.
  • If the effective quantizer is so low that no deblocking would be performed no matter what, don’t bother calculating the strength.

  But even all of this doesn’t save us from ourselves. We still have to iterate over a ton of edges, checking each one. Stuff like the partition-checking logic greatly complicates the code and adds overhead even as it reduces the number of checks. And in many cases decoupling the checks to add such logic will make it slower : if the checks are coupled, we can avoid doing a motion vector check if there’s residual, since Strength 2 overrides Strength 1.

  But wait. What if we could do this in SIMD, just like the actual loopfilter itself ? Sure, it seems more of a problem for C code than assembly, but there aren’t any obvious things in the way. Many years ago, Loren Merritt (pengvado) wrote the first SIMD implementation that I know of (for ffmpeg’s decoder) ; it is quite fast, so I decided to work on porting the idea to x264 to see if we could eke out a bit more speed here as well.

  Before I go over what I had to do to make this change, let me first describe how deblocking is implemented in x264. Since the filter is a loopfilter, it acts “in loop” and must be done in both the encoder and decoder — hence why x264 has it too, not just decoders. At the end of encoding one row of macroblocks, x264 goes back and deblocks the row, then performs half-pixel interpolation for use in encoding the next frame.

  We do it per-row for reasons of cache coherency : deblocking accesses a lot of pixels and a lot of code that wouldn’t otherwise be used, so it’s more efficient to do it in a single pass as opposed to deblocking each macroblock immediately after encoding. Then half-pixel interpolation can immediately re-use the resulting data.

  Now to the change. First, I modified deblocking to implement a subset of the macroblock_cache_load function : spend an extra bit of effort loading the necessary data into a data structure which is much simpler to address — as an assembly implementation would need (x264_macroblock_cache_load_deblock). Then I massively cleaned up deblocking to move all of the core strength-calculation logic into a single, small function that could be converted to assembly (deblock_strength_c). Finally, I wrote the assembly functions and worked with Loren to optimize them. Here’s the result.

  And the timings for the resulting assembly function on my Core i7, in cycles :

  deblock_strength_c : 309
  deblock_strength_mmx : 79
  deblock_strength_sse2 : 37
  deblock_strength_ssse3 : 33

  Now that is a seriously nice improvement. 33 cycles on average to perform that many comparisons–that’s absurdly low, especially considering the SIMD takes no branchy shortcuts : it always checks every single edge ! I walked over to my performance chart and happily crossed off a box.

  But I had a hunch that I could do better. Remember, as mentioned earlier, we’re reloading all that data back into our data structures in order to address it. This isn’t that slow, but takes enough time to significantly cut down on the gain of the assembly code. And worse, less than a row ago, all this data was in the correct place to be used (when we just finished encoding the macroblock) ! But if we did the deblocking right after encoding each macroblock, the cache issues would make it too slow to be worth it (yes, I tested this). So I went back to other things, a bit annoyed that I couldn’t get the full benefit of the changes.

  Then, yesterday, I was talking with Pascal, a former Xvid dev and current video hacker over at Google, about various possible x264 optimizations. He had seen my deblocking changes and we discussed that a bit as well. Then two lines hit me like a pile of bricks :

  <_skal_> tried computing the strength at least ?
  <_skal_> while it’s fresh

  Why hadn’t I thought of that ? Do the strength calculation immediately after encoding each macroblock, save the result, and then go pick it up later for the main deblocking filter. Then we can use the data right there and then for strength calculation, but we don’t have to do the whole deblock process until later.

  I went and implemented it and, after working my way through a horde of bugs, eventually got a working implementation. A big catch was that of slices : deblocking normally acts between slices even though normal encoding does not, so I had to perform extra munging to get that to work. By midday today I was able to go cross yet another box off on the performance chart. And now it’s committed.

  Sometimes chatting for 10 minutes with another developer is enough to spot the idea that your brain somehow managed to miss for nearly a straight week.

  NB : the performance chart is on a specific test clip at a specific set of settings (super fast settings) relevant to the company I work at, so it isn’t accurate nor complete for, say, default settings.

  Update : Here’s a higher resolution version of the current chart, as requested in the comments.
  

  http://git.videolan.org/?p=x264.git ;a=commit ;h=b9ec3ea27812288485329ac7771143021cd4917b

• Of ctors and dtors

  18 février 2011, par Multimedia Mike — Programming, Sega Dreamcast

  I haven’t given up on the Sega Dreamcast programming. I was able to compile a bunch of homebrew code for the DC many years ago and I can’t make it work anymore. Again, I was working with a purpose-built, open source RTOS named KallistiOS (or KOS). I can make the programs compile but not run. I had ELF files left over from years ago which still executed. But when I tried to build new ELF files, no luck— the programs crashed before even reaching my main() function.

  I found the problem : ELF files are comprised of a number of sections and 2 of these sections are named ’.ctors’ and ’.dtors’ which stand for constructors and destructors. The KOS RTOS performs a manual traversal of .ctors section during program initialization and this is where things go bad. The traversal code doesn’t seem to account for a .ctors section that only contains a single entry. I commented out the function that does the traversal and programs started to work, at least until it was time to exit the program and return control to the program loader. That’s when the counterpart .dtors section traversal code ran and demonstrated the same problem. I’ll exhibit the problematic code at the end of this post.

  So I’m finally tinkering with Sega Dreamcast programming once again and with a slightly better grasp of software engineering than the first time I did this.

  Portable and Compatible C ?
  If nothing else, this low-level embedded stuff exposes you to some serious toolchain arcana, the likes of which you will likely never see working strictly in the desktop arena.

  Still, this exercise makes me wonder why C code from a decade ago doesn’t compile reliably now. Part of it is because gcc has gotten stricter about the syntax it will accept. In the case of this specific crashing problem, I suspect it comes down to a difference in the way the linker generates the final ELF file. I’ve written a list of items I have had to modify in the KOS codebase in order to get it to compile on more recent gcc versions. I wonder if it would be worth publishing the specifics, or if anyone would ever find the information useful ? Oh, who am I kidding ? Of course I’ll write it up, perhaps publish a new version of the code, if only because that’s the best chance I have of finding my own work again some years down the road.

  Problematic C Code
  See if this code makes any sense to you. It somehow traverse a list of 32-bit function pointers (in different directions, depending on constructors or destructors), executing each in turn. However, it appears to fall over if the list of pointers consists of a single entry.
  

  PLAIN TEXT

  C :

  1. typedef void (*fptr)(void) ;
  2.  
  3. static fptr ctor_list[1] __attribute__((section(".ctors"))) = { (fptr) -1 } ;
  4. static fptr dtor_list[1] __attribute__((section(".dtors"))) = { (fptr) -1 } ;
  5.  
  6. /* Call this to execute all ctors */
  7. void arch_ctors() {
  8.     fptr *fpp ;
  9.  
  10.     /* Run up to the end of the list (defined by crtend) */
  11.     for (fpp=ctor_list + 1 ; *fpp != 0 ; ++fpp)
  12.          ;
  13.  
  14.     /* Now run the ctors backwards */
  15.     while (—fpp> ctor_list)
  16.         (**fpp)() ;
  17. }
  18.  
  19. /* Call this to execute all dtors */
  20. void arch_dtors() {
  21.     fptr *fpp ;
  22.  
  23.     /* Do the dtors forwards */
  24.     for (fpp=dtor_list + 1 ; *fpp != 0 ; ++fpp )
  25.         (**fpp)() ;
  26. }

• concatenate encrypted m3u8 ts video segments into a whole

  20 octobre 2016, par iMath

  The m3u8 content

  #EXTM3U
  
  #EXT-X-VERSION:3
  
  #EXT-X-ALLOW-CACHE:YES
  
  #EXT-X-TARGETDURATION:5
  
  #EXT-X-MEDIA-SEQUENCE:0
  
  #EXT-X-KEY:METHOD=AES-128,URI="http://www.gaiamount.com/keys/3958/4k",IV=0x440796ac5ac10cc577f9bec05cd042ef
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0000.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0001.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0002.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0003.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0004.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0005.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0006.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0007.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0008.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0009.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0010.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0011.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0012.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0013.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0014.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0015.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0016.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0017.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0018.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0019.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0020.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0021.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0022.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0023.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0024.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0025.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0026.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0027.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0028.ts
  
  #EXTINF:5.000000,
  
  /seg/u116/v810_mp4_0029.ts
  
  #EXTINF:4.760000,
  
  /seg/u116/v810_mp4_0030.ts
  
  #EXT-X-ENDLIST

  I’ve already download all the ts video segments using another Downloader , then try to follow the method to concatenate these ts video segments into a whole , however the ffmpeg complains the following

  C:\Users\i>ffmpeg -f concat -i mylist.txt -c copy output.ts
  
  ffmpeg version N-81475-gdc7e5ad Copyright (c) 2000-2016 the FFmpeg developers
  
    built with gcc 5.4.0 (GCC)
  
    configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-dx
  
  va2 --enable-libmfx --enable-nvenc --enable-avisynth --enable-bzlib --enable-lib
  
  ebur128 --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --ena
  
  ble-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libfree
  
  type --enable-libgme --enable-libgsm --enable-libilbc --enable-libmodplug --enab
  
  le-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-lib
  
  openh264 --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschr
  
  oedinger --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libtheor
  
  a --enable-libtwolame --enable-libvidstab --enable-libvo-amrwbenc --enable-libvo
  
  rbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --ena
  
  ble-libx265 --enable-libxavs --enable-libxvid --enable-libzimg --enable-lzma --e
  
  nable-decklink --enable-zlib
  
    libavutil      55. 29.100 / 55. 29.100
  
    libavcodec     57. 54.100 / 57. 54.100
  
    libavformat    57. 48.100 / 57. 48.100
  
    libavdevice    57.  0.102 / 57.  0.102
  
    libavfilter     6. 57.100 /  6. 57.100
  
    libswscale      4.  1.100 /  4.  1.100
  
    libswresample   2.  1.100 /  2.  1.100
  
    libpostproc    54.  0.100 / 54.  0.100
  
  [concat @ 03e30fe0] Impossible to open 'C:/Users/i/AppData/Roaming/iMath/01.ts'
  
  mylist.txt: Invalid data found when processing input
  
  
  
  C:\Users\i>

  so (1)what should I do to concatenate these encrypted m3u8 ts video segments into a whole ?
  (2) Any way to get the video byte size an m3u8 url referenced to ?