Recherche avancée

Sur d’autres sites (4996)

• Inside WebM Technology : The VP8 Alternate Reference Frame

  15 juin 2010, par noreply@blogger.com (John Luther) — inside webm, vp8

  Since the WebM project was open-sourced just a week ago, we’ve seen blog posts and articles about its capabilities. As an open project, we welcome technical scrutiny and contributions that improve the codec. We know from our extensive testing that VP8 can match or exceed other leading codecs, but to get the best results, it helps to understand more about how the codec works. In this first of a series of blog posts, I’ll explain some of the fundamental techniques in VP8, along with examples and metrics.

  The alternative reference frame is one of the most exciting quality innovations in VP8. Let’s delve into how VP8 uses these frames to improve prediction and thereby overall video quality.

  Alternate Reference Frames in VP8

  VP8 uses three types of reference frames for inter prediction : the last frame, a "golden" frame (one frame worth of decompressed data from the arbitrarily distant past) and an alternate reference frame. Overall, this design has a much smaller memory footprint on both encoders and decoders than designs with many more reference frames. In video compression, it is very rare for more than three reference frames to provide significant quality benefit, but the undesirable increase in memory footprint from the extra frames is substantial.

  Unlike other types of reference frames used in video compression, which are displayed to the user by the decoder, the VP8 alternate reference frame is decoded normally but is never shown to the user. It is used solely as a reference to improve inter prediction for other coded frames. Because alternate reference frames are not displayed, VP8 encoders can use them to transmit any data that are helpful to compression. For example, a VP8 encoder can construct one alternate reference frame from multiple source frames, or it can create an alternate reference frame using different macroblocks from hundreds of different video frames.

  The current VP8 implementation enables two different types of usage for the alternate reference frame : noise-reduced prediction and past/future directional prediction.

  Noise-Reduced Prediction

  The alternate reference frame is transmitted and decoded similar to other frames, hence its usage does not add extra computation in decoding. The VP8 encoder however is free to use more sophisticated processing to create them in off-line encoding. One application of the alternate reference frame is for noise-reduced prediction. In this application, the VP8 encoder uses multiple input source frames to construct one reference frame through temporal or spatial noise filtering. This "noise-free" alternate reference frame is then used to improve prediction for encoding subsequent frames.

  You can make use of this feature by setting ARNR parameters in VP8 encoding, where ARNR stands for "Alternate Reference Noise Reduction." A sample two-pass encoding setting with the parameters :

  --arnr-maxframes=5 --arnr-strength=3

  enables the encoder to use "5" consecutive input source frames to produce one alternate reference frame using a filtering strength of "3". Here is an example showing the quality benefit of using this experimental "ARNR" feature on the standard test clip "Hall Monitor." (Each line on the graph represents the quality of an encoded stream on a given clip at multiple datarates. The higher points on the Y axis (PSNR) indicates the stream with the better quality.)

  
  

  The only difference between the two curves in the graph is that VP8_ARNR was produced by encodings with ARNR parameters and VP8_NO_ARNR was not. As we can see from the graph, noise reduced prediction is very helpful to compression quality when encoding noisy sources. We’ve just started to explore this idea but have already seen strong improvements on noisy input clips similar to this "Hall Monitor." We feel there’s a lot more we can do in this area.

  Improving Prediction without B Frames

  The lack of B frames in VP8 has sparked some discussion about its ability to achieve competitive compression efficiency. VP8 encoders, however, can make intelligent use of the golden reference and the alternate reference frames to compensate for this. The VP8 encoder can choose to transmit an alternate reference frame similar to a "future" frame, and encoding of subsequent frames can make use of information from the past (last frame and golden frame) and from the future (alternate reference frame). Effectively, this helps the encoder to achieve results similar to bidirectional (B frame) prediction without requiring frame reordering in the decoder. Running in two-pass encoding mode, compression can be improved in the VP8 encoder by using encoding parameters that enable lagged encoding and automatic placement of alternate reference frames :

  --auto-alt-ref=1 --lag-in-frames=16

  Used this way, the VP8 encoder can achieve improved prediction and compression efficiency without increasing the decoder’s complexity :

  
  

  In the video compression community, "Mobile and calendar" is known as a clip that benefits significantly from the usage of B frames. The graph above illustrates that the use of alternate reference frame benefits VP8 significantly without using B frames.

  Keep an eye on this blog for more posts about VP8 encoding. You can find more information on above encoding parameters or other detailed instructions to use with our VP8 encoders on our site, or join our discussion list.

  Yaowu Xu, Ph.D. is a codec engineer at Google.

  

• Beware the builtins

  14 janvier 2010, par Mans — Compilers

  GCC includes a large number of builtin functions allegedly providing optimised code for common operations not easily expressed directly in C. Rather than taking such claims at face value (this is GCC after all), I decided to conduct a small investigation to see how well a few of these functions are actually implemented for various targets.

  For my test, I selected the following functions :

  • __builtin_bswap32 : Byte-swap a 32-bit word.
  • __builtin_bswap64 : Byte-swap a 64-bit word.
  • __builtin_clz : Count leading zeros in a word.
  • __builtin_ctz : Count trailing zeros in a word.
  • __builtin_prefetch : Prefetch data into cache.

  To test the quality of these builtins, I wrapped each in a normal function, then compiled the code for these targets :

  • ARMv7
  • AVR32
  • MIPS
  • MIPS64
  • PowerPC
  • PowerPC64
  • x86
  • x86_64

  In all cases I used compiler flags were -O3 -fomit-frame-pointer plus any flags required to select a modern CPU model.
  

  ARM

  Both __builtin_clz and __builtin_prefetch generate the expected CLZ and PLD instructions respectively. The code for __builtin_ctz is reasonable for ARMv6 and earlier :

  rsb     r3, r0, #0
  and     r0, r3, r0
  clz     r0, r0
  rsb     r0, r0, #31
  

  For ARMv7 (in fact v6T2), however, using the new bit-reversal instruction would have been better :

  rbit    r0, r0
  clz     r0, r0
  

  I suspect this is simply a matter of the function not yet having been updated for ARMv7, which is perhaps even excusable given the relatively rare use cases for it.

  The byte-reversal functions are where it gets shocking. Rather than use the REV instruction found from ARMv6 on, both of them generate external calls to __bswapsi2 and __bswapdi2 in libgcc, which is plain C code :

  SItype
  __bswapsi2 (SItype u)
  
    return ((((u) & 0xff000000) >> 24)
            | (((u) & 0x00ff0000) >>  8)
            | (((u) & 0x0000ff00) <<  8)
            | (((u) & 0x000000ff) << 24)) ;
  
  DItype
  
  __bswapdi2 (DItype u)
  
  
  
     return ((((u) & 0xff00000000000000ull) >> 56)
  
            | (((u) & 0x00ff000000000000ull) >> 40)
  
            | (((u) & 0x0000ff0000000000ull) >> 24)
  
            | (((u) & 0x000000ff00000000ull) >>  8)
  
            | (((u) & 0x00000000ff000000ull) <<  8)
  
            | (((u) & 0x0000000000ff0000ull) << 24)
  
            | (((u) & 0x000000000000ff00ull) << 40)
  
            | (((u) & 0x00000000000000ffull) << 56)) ;
  
  
  

  While the 32-bit version compiles to a reasonable-looking shift/mask/or job, the 64-bit one is a real WTF. Brace yourselves :

  push    r4, r5, r6, r7, r8, r9, sl, fp
  mov     r5, #0
  mov     r6, #65280 ; 0xff00
  sub     sp, sp, #40 ; 0x28
  and     r7, r0, r5
  and     r8, r1, r6
  str     r7, [sp, #8]
  str     r8, [sp, #12]
  mov     r9, #0
  mov     r4, r1
  and     r5, r0, r9
  mov     sl, #255 ; 0xff
  ldr     r9, [sp, #8]
  and     r6, r4, sl
  mov     ip, #16711680 ; 0xff0000
  str     r5, [sp, #16]
  str     r6, [sp, #20]
  lsl     r2, r0, #24
  and     ip, ip, r1
  lsr     r7, r4, #24
  mov     r1, #0
  lsr     r5, r9, #24
  mov     sl, #0
  mov     r9, #-16777216 ; 0xff000000
  and     fp, r0, r9
  lsr     r6, ip, #8
  orr     r9, r7, r1
  and     ip, r4, sl
  orr     sl, r1, r2
  str     r6, [sp]
  str     r9, [sp, #32]
  str     sl, [sp, #36] ; 0x24
  add     r8, sp, #32
  ldm     r8, r7, r8
  str     r1, [sp, #4]
  ldm     sp, r9, sl
  orr     r7, r7, r9
  orr     r8, r8, sl
  str     r7, [sp, #32]
  str     r8, [sp, #36] ; 0x24
  mov     r3, r0
  mov     r7, #16711680 ; 0xff0000
  mov     r8, #0
  and     r9, r3, r7
  and     sl, r4, r8
  ldr     r0, [sp, #16]
  str     fp, [sp, #24]
  str     ip, [sp, #28]
  stm     sp, r9, sl
  ldr     r7, [sp, #20]
  ldr     sl, [sp, #12]
  ldr     fp, [sp, #12]
  ldr     r8, [sp, #28]
  lsr     r0, r0, #8
  orr     r7, r0, r7, lsl #24
  lsr     r6, sl, #24
  orr     r5, r5, fp, lsl #8
  lsl     sl, r8, #8
  mov     fp, r7
  add     r8, sp, #32
  ldm     r8, r7, r8
  orr     r6, r6, r8
  ldr     r8, [sp, #20]
  ldr     r0, [sp, #24]
  orr     r5, r5, r7
  lsr     r8, r8, #8
  orr     sl, sl, r0, lsr #24
  mov     ip, r8
  ldr     r0, [sp, #4]
  orr     fp, fp, r5
  ldr     r5, [sp, #24]
  orr     ip, ip, r6
  ldr     r6, [sp]
  lsl     r9, r5, #8
  lsl     r8, r0, #24
  orr     fp, fp, r9
  lsl     r3, r3, #8
  orr     r8, r8, r6, lsr #8
  orr     ip, ip, sl
  lsl     r7, r6, #24
  and     r5, r3, #16711680 ; 0xff0000
  orr     r7, r7, fp
  orr     r8, r8, ip
  orr     r4, r1, r7
  orr     r5, r5, r8
  mov     r9, r6
  mov     r1, r5
  mov     r0, r4
  add     sp, sp, #40 ; 0x28
  pop     r4, r5, r6, r7, r8, r9, sl, fp
  bx      lr
  

  That’s right, 91 instructions to move 8 bytes around a bit. GCC definitely has a problem with 64-bit numbers. It is perhaps worth noting that the bswap_64 macro in glibc splits the 64-bit value into 32-bit halves which are then reversed independently, thus side-stepping this weakness of gcc.

  As a side note, ARM RVCT (armcc) compiles those functions perfectly into one and two REV instructions, respectively.

  AVR32

  There is not much to report here. The latest gcc version available is 4.2.4, which doesn’t appear to have the bswap functions. The other three are handled nicely, even using a bit-reverse for __builtin_ctz.

  MIPS / MIPS64

  The situation MIPS is similar to ARM. Both bswap builtins result in external libgcc calls, the rest giving sensible code.

  PowerPC

  I scarcely believe my eyes, but this one is actually not bad. The PowerPC has no byte-reversal instructions, yet someone seems to have taken the time to teach gcc a good instruction sequence for this operation. The PowerPC does have some powerful rotate-and-mask instructions which come in handy here. First the 32-bit version :

  rotlwi  r0,r3,8
  rlwimi  r0,r3,24,0,7
  rlwimi  r0,r3,24,16,23
  mr      r3,r0
  blr
  

  The 64-bit byte-reversal simply applies the above code on each half of the value :

  rotlwi  r0,r3,8
  rlwimi  r0,r3,24,0,7
  rlwimi  r0,r3,24,16,23
  rotlwi  r3,r4,8
  rlwimi  r3,r4,24,0,7
  rlwimi  r3,r4,24,16,23
  mr      r4,r0
  blr
  

  Although I haven’t analysed that code carefully, it looks pretty good.

  PowerPC64

  Doing 64-bit operations is easier on a 64-bit CPU, right ? For you and me perhaps, but not for gcc. Here __builtin_bswap64 gives us the now familiar __bswapdi2 call, and while not as bad as the ARM version, it is not pretty :

  rldicr  r0,r3,8,55
  rldicr  r10,r3,56,7
  rldicr  r0,r0,56,15
  rldicl  r11,r3,8,56
  rldicr  r9,r3,16,47
  or      r11,r10,r11
  rldicr  r9,r9,48,23
  rldicl  r10,r0,24,40
  rldicr  r0,r3,24,39
  or      r11,r11,r10
  rldicl  r9,r9,40,24
  rldicr  r0,r0,40,31
  or      r9,r11,r9
  rlwinm  r10,r3,0,0,7
  rldicl  r0,r0,56,8
  or      r0,r9,r0
  rldicr  r10,r10,8,55
  rlwinm  r11,r3,0,8,15
  or      r0,r0,r10
  rldicr  r11,r11,24,39
  rlwinm  r3,r3,0,16,23
  or      r0,r0,r11
  rldicr  r3,r3,40,23
  or      r3,r0,r3
  blr
  

  That is 6 times longer than the (presumably) hand-written 32-bit version.

  x86 / x86_64

  As one might expect, results on x86 are good. All the tested functions use the available special instructions. One word of caution though : the bit-counting instructions are very slow on some implementations, specifically the Atom, AMD chips, and the notoriously slow Pentium4E.

  Conclusion

  In conclusion, I would say gcc builtins can be useful to avoid fragile inline assembler. Before using them, however, one should make sure they are not in fact harmful on the required targets. Not even those builtins mapping directly to CPU instructions can be trusted.

• Revision 15938 : Canoniser les dates de l’agenda (32/12/2009 => 1/1/2010) dans les URL ...

  14 août 2010, par esj@… — Log

  Canoniser les dates de l’agenda (32/12/2009 => 1/1/2010) dans les URL produites plutôt que de le faire après, et du coup refuser ces mauvaises dates, inévitablement produites par un robot. Partage de code entre les appels public et privé de l’agenda, et correction d’un bug d’un cas rare de l’agenda (...)