Recherche avancée

Sur d’autres sites (6119)

• Blitted OpenGL Textures take less memory and CPU

  16 avril 2015, par Pedro H. Forli

  I’m making a game using pygame + pyopengl, and right now i’m trying to make a video player on this context. To do so I use ffmpeg to load different video formats, then convert each frame to an opengl texture, as designed below, and then play the video.

  class Texture(object):
  
      def __init__(self, data, w=0, h=0):
  
          """
  
          Initialize the texture from 3 diferents types of data:
  
          filename = open the image, get its string and produce texture
  
          surface = get its string and produce texture
  
          string surface = gets it texture and use w and h provided
  
          """
  
          if type(data) == str:
  
              texture_data = self.load_image(data)
  
  
  
          elif type(data) == pygame.Surface:
  
              texture_data = pygame.image.tostring(data, "RGBA", True)
  
              self.w, self.h = data.get_size()
  
  
  
          elif type(data) == bytes:
  
              self.w, self.h = w, h
  
              texture_data = data
  
  
  
          self.texID = 0
  
          self.load_texture(texture_data)
  
  
  
      def load_image(self, data):
  
          texture_surface = pygame.image.load(data).convert_alpha()
  
          texture_data = pygame.image.tostring(texture_surface, "RGBA", True)
  
          self.w, self.h = texture_surface.get_size()
  
  
  
          return texture_data
  
  
  
      def load_texture(self, texture_data):
  
          self.texID = glGenTextures(1)
  
  
  
          glBindTexture(GL_TEXTURE_2D, self.texID)
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST)
  
          glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST)
  
          glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA, self.w,
  
                       self.h, 0, GL_RGBA, GL_UNSIGNED_BYTE,
  
                       texture_data)

  Problem is that when i load all the textures of a given video, my RAM goes off the ceiling, about 800mb. But it’s possible to work around this by blitting each texture as it loads, like shown below.

  def render():
  
      glClear(GL_COLOR_BUFFER_BIT|GL_DEPTH_BUFFER_BIT)
  
      glLoadIdentity()
  
      glDisable(GL_LIGHTING)
  
      glEnable(GL_TEXTURE_2D)
  
      glBlendFunc(GL_SRC_ALPHA, GL_ONE_MINUS_SRC_ALPHA)
  
      glClearColor(0, 0, 0, 1.0)
  
  
  
  def Draw(texture, top, left, bottom, right):
  
      """
  
      Draw the image on the Opengl Screen
  
      """
  
      # Make sure he is looking at the position (0,0,0)
  
      glBindTexture(GL_TEXTURE_2D, texture.texID)
  
      glBegin(GL_QUADS)
  
  
  
      # The top left of the image must be the indicated position
  
      glTexCoord2f(0.0, 1.0)
  
      glVertex2f(left, top)
  
  
  
      glTexCoord2f(1.0, 1.0)
  
      glVertex2f(right, top)
  
  
  
      glTexCoord2f(1.0, 0.0)
  
      glVertex2f(right, bottom)
  
  
  
      glTexCoord2f(0.0, 0.0)
  
      glVertex2f(left, bottom)
  
  
  
      glEnd()

  def update(t) :
  render()
  Draw(t, -0.5, -0.5, 0.5, 0.5)

  # Check for basic Events on the pygame interface
  
  for event in pygame.event.get():
  
      BASIC_Game.QUIT_Event(event)
  
  
  
  pygame.display.flip()

  Although this reduces the RAM consumption to an acceptable value it makes the loading time bigger than the video length.

  I really don’t understand why opengl works this way, but is there a way to make a texture efficient without blitting it first ?

• ARM inline asm secrets

  6 juillet 2010, par Mans — ARM, Compilers

  Although I generally recommend against using GCC inline assembly, preferring instead pure assembly code in separate files, there are occasions where inline is the appropriate solution. Should one, at a time like this, turn to the GCC documentation for guidance, one must be prepared for a degree of disappointment. As it happens, much of the inline asm syntax is left entirely undocumented. This article attempts to fill in some of the blanks for the ARM target.
  

  Constraints

  Each operand of an inline asm block is described by a constraint string encoding the valid representations of the operand in the generated assembly. For example the “r” code denotes a general-purpose register. In addition to the standard constraints, ARM allows a number of special codes, only some of which are documented. The full list, including a brief description, is available in the constraints.md file in the GCC source tree. The following table is an extract from this file consisting of the codes which are meaningful in an inline asm block (a few are only useful in the machine description itself).

  f	Legacy FPA registers f0-f7.
  t	The VFP registers s0-s31.
  v	The Cirrus Maverick co-processor registers.
  w	The VFP registers d0-d15, or d0-d31 for VFPv3.
  x	The VFP registers d0-d7.
  y	The Intel iWMMX co-processor registers.
  z	The Intel iWMMX GR registers.
  l	In Thumb state the core registers r0-r7.
  h	In Thumb state the core registers r8-r15.
  j	A constant suitable for a MOVW instruction. (ARM/Thumb-2)
  b	Thumb only. The union of the low registers and the stack register.
  I	In ARM/Thumb-2 state a constant that can be used as an immediate value in a Data Processing instruction. In Thumb-1 state a constant in the range 0 to 255.
  J	In ARM/Thumb-2 state a constant in the range -4095 to 4095. In Thumb-1 state a constant in the range -255 to -1.
  K	In ARM/Thumb-2 state a constant that satisfies the I constraint if inverted. In Thumb-1 state a constant that satisfies the I constraint multiplied by any power of 2.
  L	In ARM/Thumb-2 state a constant that satisfies the I constraint if negated. In Thumb-1 state a constant in the range -7 to 7.
  M	In Thumb-1 state a constant that is a multiple of 4 in the range 0 to 1020.
  N	Thumb-1 state a constant in the range 0 to 31.
  O	In Thumb-1 state a constant that is a multiple of 4 in the range -508 to 508.
  Pa	In Thumb-1 state a constant in the range -510 to +510
  Pb	In Thumb-1 state a constant in the range -262 to +262
  Ps	In Thumb-2 state a constant in the range -255 to +255
  Pt	In Thumb-2 state a constant in the range -7 to +7
  G	In ARM/Thumb-2 state a valid FPA immediate constant.
  H	In ARM/Thumb-2 state a valid FPA immediate constant when negated.
  Da	In ARM/Thumb-2 state a const_int, const_double or const_vector that can be generated with two Data Processing insns.
  Db	In ARM/Thumb-2 state a const_int, const_double or const_vector that can be generated with three Data Processing insns.
  Dc	In ARM/Thumb-2 state a const_int, const_double or const_vector that can be generated with four Data Processing insns. This pattern is disabled if optimizing for space or when we have load-delay slots to fill.
  Dn	In ARM/Thumb-2 state a const_vector which can be loaded with a Neon vmov immediate instruction.
  Dl	In ARM/Thumb-2 state a const_vector which can be used with a Neon vorr or vbic instruction.
  DL	In ARM/Thumb-2 state a const_vector which can be used with a Neon vorn or vand instruction.
  Dv	In ARM/Thumb-2 state a const_double which can be used with a VFP fconsts instruction.
  Dy	In ARM/Thumb-2 state a const_double which can be used with a VFP fconstd instruction.
  Ut	In ARM/Thumb-2 state an address valid for loading/storing opaque structure types wider than TImode.
  Uv	In ARM/Thumb-2 state a valid VFP load/store address.
  Uy	In ARM/Thumb-2 state a valid iWMMX load/store address.
  Un	In ARM/Thumb-2 state a valid address for Neon doubleword vector load/store instructions.
  Um	In ARM/Thumb-2 state a valid address for Neon element and structure load/store instructions.
  Us	In ARM/Thumb-2 state a valid address for non-offset loads/stores of quad-word values in four ARM registers.
  Uq	In ARM state an address valid in ldrsb instructions.
  Q	In ARM/Thumb-2 state an address that is a single base register.

  Operand codes

  Within the text of an inline asm block, operands are referenced as %0, %1 etc. Register operands are printed as rN, memory operands as [rN, #offset], and so forth. In some situations, for example with operands occupying multiple registers, more detailed control of the output may be required, and once again, an undocumented feature comes to our rescue.

  Special code letters inserted between the % and the operand number alter the output from the default for each type of operand. The table below lists the more useful ones.

  c	An integer or symbol address without a preceding # sign
  B	Bitwise inverse of integer or symbol without a preceding #
  L	The low 16 bits of an immediate constant
  m	The base register of a memory operand
  M	A register range suitable for LDM/STM
  H	The highest-numbered register of a pair
  Q	The least significant register of a pair
  R	The most significant register of a pair
  P	A double-precision VFP register
  p	The high single-precision register of a VFP double-precision register
  q	A NEON quad register
  e	The low doubleword register of a NEON quad register
  f	The high doubleword register of a NEON quad register
  h	A range of VFP/NEON registers suitable for VLD1/VST1
  A	A memory operand for a VLD1/VST1 instruction
  y	S register as indexed D register, e.g. s5 becomes d2[1]

• Video files recorded in Google Chrome have stuttering audio

  4 juin 2018, par maxpaj

  Background

  I’m developing a platform where users can record videos of themselves or their screen and send them as video messages to customers / clients.

  I have limited users to only using my application with Google Chrome and I’m using the MediaRecorder API to record the video data from the users screen or webcamera. The codecs that are used for recording are VP8/OPUS (WEBM container).

  I need the videos to run in as many browsers as possible, so I’m using a 3rd party service to transcode videos from whatever format I’m getting from the users to a H.265/AAC MP4 container (caniuse MPEG-4/H.264).

  Issue

  Lately I’ve seen that some videos recorded on Mac OSX machines have the video and audio out of sync or that the video and audio stutters, depending on which player I’m using. I call these video files corrupt, for lack of a better word. Playing a corrupt file in Google Chrome renders smooth playing audio. Playing the video in VLC on my Windows machine renders stuttering audio.

  When I run the corrupt video files through the transcoding service I get video files with stuttering audio, no matter which player I’m using.

  This is an unwanted result and pretty much unacceptable since the audio needs to be smooth in order for the recipient of a video to not be bothered with the quality.

  Debugging

  According to the transcoding service support, this happens because of their mechanisms that try to sync up the audio and video from the corrupt file :

  Inspecting our encoding logs, I’ve noticed the following kind of
  warnings :

  [2018-05-16 14:08:38.009] [pcm_s16le @ 0x1d608c0] pcm_encode_frame :
  filling in for 5856 missing samples (122 ms) before pts 40800 to
  correct sync ! [2018-05-16 14:08:38.009] [pcm_s16le @ 0x1d608c0]
  pcm_encode_frame : dropping 2880 samples (60 ms) at pts 43392 to help
  correct sync to -3168 samples (-66 ms) !

  The problem here comes from the way that the audio in the original
  source file is encoded.

  -

  you should ensure that the audio is not out of sync (audio timestamps
  are correct) in your source file before submitting the job

  Running a corrupt file through ffmpeg on my own machine, re-encoding with the same codecs, produces the same kind of stuttering video. The logs produce an alarming amount of errors. Here is a sample of the log output :

  [libopus @ 0000029938e24d80] Queue input is backward in timeitrate= 194.8kbits/s dup=0 drop=5 speed=0.31x
  
  [webm @ 0000029938e09b00] Non-monotonous DTS in output stream 0:1; previous: 15434, current: 15394; changing to 15434. This may result in incorrect timestamps in the output file.
  
  [webm @ 0000029938e09b00] Non-monotonous DTS in output stream 0:1; previous: 15434, current: 15414; changing to 15434. This may result in incorrect timestamps in the output file.
  
  [libopus @ 0000029938e24d80] Queue input is backward in timeitrate= 193.3kbits/s dup=0 drop=5 speed=0.309x
  
  [webm @ 0000029938e09b00] Non-monotonous DTS in output stream 0:1; previous: 15539, current: 15499; changing to 15539. This may result in incorrect timestamps in the output file.
  
  [webm @ 0000029938e09b00] Non-monotonous DTS in output stream 0:1; previous: 15539, current: 15519; changing to 15539. This may result in incorrect timestamps in the output file.
  
  [libopus @ 0000029938e24d80] Queue input is backward in timeitrate= 192.0kbits/s dup=0 drop=5 speed=0.308x
  
  [webm @ 0000029938e09b00] Non-monotonous DTS in output stream 0:1; previous: 15667, current: 15627; changing to 15667. This may result in incorrect timestamps in the output file.
  
  [webm @ 0000029938e09b00] Non-monotonous DTS in output stream 0:1; previous: 15667, current: 15647; changing to 15667. This may result in incorrect timestamps in the output file.
  
  [libopus @ 0000029938e24d80] Queue input is backward in time

  I tried running the same inputs through another transcoding service and those outputs worked a lot better - video was still stuttering but the audio played smoothly, which is more important to the use case of my application.

  To my knowledge, this have so far only occurred for users on Mac OSX machines.

  Questions

  1. Is there anything I can do to have the files work better ? Or is this entirely a consequence of how encoding of video and audio in Google Chrome works ?

  2. One step in the right direction would be to just be able to detect when the video is corrupt. How can I do that ?