Psyk's stuff

Further to my old (and probably obsolete) entry here I thought it worthwhile to add an update in regards to encoding to MP4 format playable on Windows Media Centres and PS3's, although it should work on other platforms such as the Xbox but can't be sure as I don't have one for testing.

One of the biggest issues I had with the previous method was the reliability in muxing it into an MP4 via MP4Box and especially when trying to get the audio, video sync right which I think had to do with it not being able to detect the FPS correctly.  In any case, the procedure involves converting everthing to an AVI (mencoder's preferred format) and then ripping/converting them raw and muxing it back into an MP4 using MP4box, mencoder can do MP4's but it's broken.

First things first, and I can't take all the credit for this.  Credit goes to the h264enc project which uses mencoder heavily and has some great built-in profiles.  Excellent display of show-casing mencoder's power and versatility.

Lets, start...

Add a new profile entry in ~/.mplayer/mencoder.conf  

 [H264ENC]  
 profile-desc="taken from h264enc, PS3 compatible encode with high quality."  
 vf=pp=al:c,softskip,scale=720:-10,harddup   
 aspect=16:9  
 af=volnorm=2,lavcresample=48000:16:1   
 srate=48000  
 oac=faac=yes   
 faacopts=mpeg=4:br=160:tns=yes:object=2  
 ovc=x264=yes  
 x264encopts=crf=19:me=umh:me_range=24:nodct_decimate:nointerlaced:8x8dct:nofast_pskip:trellis=1:partitions=p8x8,b8x8,i8x8,i4x4:mixed_refs:keyint=240:keyint_min=24:psy_rd=0.8,0.2:frameref=3:bframes=3:b_adapt=2:b_pyramid:weight_b:direct_pred=auto:subq=7:chroma_me:cabac:aud:aq_mode=1:deblock:vbv_maxrate=20000:vbv_bufsize=20000:level_idc=41:threads=auto:ssim:psnr  

Obviously you can add your own parameters here especially in the "vf" section where you can adjust the scale, add some additional features (e.g. pullup,softskip, yadif etc.)

I won't go into all the details here, but basically this scales the image down to a 720p type format, normalizes the volume and resamples it to 48kHz, encodes it to AAC at 160kb and uses a constant rate factor for h264 of '19' for higher-end quality without much regard for final file size (i.e. it focuses on maximizing the video quality).

Once you have your profile you can run:

 mencoder -profile H264ENC "source_movie_file.ext" -o "source_movie_file.avi"  

This is the encoding process and the lengthiest bit...

If you want to hardsub an SRT file you can specify the -sub "name_of_srt_file.srt" and the -subfont-autoscale 1 parameters to the mencoder line above.

For example:

 mencoder -profile H264ENC "source_movie_file.ext" -o "source_movie_file.ext.avi" -sub "name_of_srt_file.srt" -subfont-autoscale 1  

Once this is finished, you'll have a h264 encoded video and aac encoded audio muxed into an AVI format, which is not very friendly format and container combo, only mencoder/mplayer/vlc player friendly :)

Next process requires us to split the video and audio into 'raw' input files that MP4Box can handle. We'll do the video first:

 mencoder "source_movie_file.avi" -nosound -ovc copy -of rawvideo -o h264_video.h264  

This will output a file called "h264_video.h264" in a raw stream that MP4Box can understand. As we've already done the encoding in the first step this will be quick as we're just ripping it from the AVI file.

Next, is the audio:

 MP4Box -aviraw audio "source_movie_file.avi" -out aac.raw  

Here we're using MP4Box to rip the audio from the AVI to a file called "aac_audio.raw" as MP4Box will rename the output.

Rename it so that MP4Box can manage it as an AAC input:

 mv aac_audio.raw aac_audio.aac  

The next step is to pick up the framerate (FPS) of the original movie source for correct interpretation into MP4.

 mplayer "source_movie_file.avi" -noconfig all -loop 1 -identify -nosound -vo null -nocache -frames 1 2>/dev/null | grep '^ID_VIDEO_FPS' | tail -n 1 | awk -F= '{print $2}' > mp4fps.fps  

This outputs the FPS of the original converted movie (AVI file) to a file called mp4fps.fps.

The final step muxes the video and audio into an MP4 container compatible with the many MP4 hardware devices and media centres. This may not work with an iPod as it uses some advanced parameters in the encoding. I'll try and post up the iPod parameters another time...

 MP4Box -fps $(cat "mp4fps.fps") -add "h264_video.h264" -add "aac_audio.aac"#audio:name="LC-AAC Stereo" -itags name="source_movie_file.avi":comment="Tagged by Psyk on $(date)" -mpeg4 -new "source_movie_file.avi".mp4  

The final output will be something like "Source_movie_file.avi.mp4". You can obviously change these output files and edit the tags.

Lastly, the cleanup:

 rm h264_video.h264  
 rm aac_audio.aac  
 rm mp4fps.fps  

To make things easier I put all of these into a very simple shell-script.

 mencoder -profile H264ENC "$1" -o "$1".avi  
 sleep 2  
 mencoder "$1".avi -nosound -ovc copy -of rawvideo -o h264_video.h264  
 sleep 2  
 MP4Box -aviraw audio "$1".avi -out aac.raw  
 sleep 2  
 mv aac_audio.raw aac_audio.aac  
 mplayer "$1".avi -noconfig all -loop 1 -identify -nosound -vo null -nocache -frames 1 2>/dev/null | grep '^ID_VIDEO_FPS' | tail -n 1 | awk -F= '{print $2}' > mp4fps  
 .fps  
 MP4Box -fps $(cat "mp4fps.fps") -add "h264_video.h264" -add "aac_audio.aac"#audio:name="LC-AAC Stereo" -itags name="$1":comment="Tagged by lazza on $(date)" -mpeg4  
  -new "$1".mp4  
 sleep 2  
 rm h264_video.h264  
 rm aac_audio.aac  
 rm mp4fps.fps  

One of the more challenging items is adjusting the aspect ratio of your final video. It requires some understanding of Pixel Aspect Ratio which is different to your standard aspect ratio. I won't go into detail here, but basically to change your video format from 4:3 to 16:9 you'll need to feed the PAR parameters.

Here are the rough PAR specs on standard DVD's.
NTSC 4:3 - storage resolution 720 x 480,PAR 8/9 - display resolution 640 x 480 (= 4:3 aspect ratio)
NTSC 16:9 - storage resolution 720 x 480,PAR 32/27 - display resolution 853 x 480 (= 16:9 aspect ratio)
PAL 4:3 - storage resolution 720 x 576,PAR 16/15 - display resolution 768 x 576 (= 4:3 aspect ratio)
PAL 16:9 - storage resolution 720 x 576,PAR 64/45 - display resolution 1024 x 576 (= 16:9 aspect ratio)

Under HandBrake I've been using the preset settings for doing the encode and they're working well. Here's a list of the preset settings as per 0.94 of Handbrake:

 Apple:  
 Universal: -e x264 -q 20.0 -a 1,1 -E faac,ac3 -B 160,160 -6 dpl2,auto -R 48,Auto -D 0.0,0.0 -f mp4 -X 720 --loose-anamorphic -m -x cabac=0:ref=2:me=umh:bframes=0:8x8dct=0:trellis=0:subme=6  
 iPod: -e x264 -b 700 -a 1 -E faac -B 160 -6 dpl2 -R 48 -D 0.0 -f mp4 -I -X 320 -m -x level=30:bframes=0:cabac=0:ref=1:vbv-maxrate=768:vbv-bufsize=2000:analyse=all:me=umh:no-fast-pskip=1:subme=6:8x8dct=0:trellis=0  
 iPhone & iPod Touch: -e x264 -q 20.0 -a 1 -E faac -B 128 -6 dpl2 -R 48 -D 0.0 -f mp4 -X 480 -m -x cabac=0:ref=2:me=umh:bframes=0:subme=6:8x8dct=0:trellis=0  
  AppleTV: -e x264 -q 20.0 -a 1,1 -E faac,ac3 -B 160,160 -6 dpl2,auto -R 48,Auto -D 0.0,0.0 -f mp4 -4 -X 960 --loose-anamorphic -m -x cabac=0:ref=2:me=umh:b-adapt=2:weightb=0:trellis=0:weightp=0  
 ----  
 Regular  
 Normal: -e x264 -q 20.0 -a 1 -E faac -B 160 -6 dpl2 -R 48 -D 0.0 -f mp4 --strict-anamorphic -m -x ref=2:bframes=2:subme=6:mixed-refs=0:weightb=0:8x8dct=0:trellis=0  
 High Profile: -e x264 -q 20.0 -a 1,1 -E faac,ac3 -B 160,160 -6 dpl2,auto -R 48,Auto -D 0.0,0.0 -f mp4 --detelecine --decomb --loose-anamorphic -m -x b-adapt=2:rc-lookahead=50  
 ----  
 Legacy  
 Classic: -b 1000 -a 1 -E faac -B 160 -6 dpl2 -R 48 -D 0.0 -f mp4  
 AppleTV Legacy: -e x264 -b 2500 -a 1,1 -E faac,ac3 -B 160,160 -6 dpl2,auto -R 48,Auto -D 0.0,0.0 -f mp4 -4 --strict-anamorphic -m -x ref=1:subme=5:me=umh:no-fast-pskip=1:cabac=0:weightb=0:8x8dct=0:trellis=0  
 iPhone Legacy: -e x264 -b 960 -a 1 -E faac -B 128 -6 dpl2 -R 48 -D 0.0 -f mp4 -I -X 480 -m -x level=30:cabac=0:ref=1:analyse=all:me=umh:no-fast-pskip=1:psy-rd=0,0:bframes=0:subme=6:8x8dct=0:trellis=0  
 iPod Legacy: -e x264 -b 1500 -a 1 -E faac -B 160 -6 dpl2 -R 48 -D 0.0 -f mp4 -I -X 640 -m -x level=30:bframes=0:cabac=0:ref=1:vbv-maxrate=1500:vbv-bufsize=2000:analyse=all:me=umh:no-fast-pskip=1:psy-rd=0,0:subme=6:8x8dct=0:trellis=0  

The Normal and High-Profile settings should work just find on a PS3.

As an example, here are the details on a source DVD I'm encoding for the kids:

• + title 1:
• + vts 0, ttn 0, cells 0->0 (0 blocks)
• + angle(s) 0
• + duration: 00:39:46
• + size: 720x576, pixel aspect: 16/15, display aspect: 1.33, 25.000 fps
• + autocrop: 0/0/12/6
• + chapters:
• + 1: cells 0->0, 0 blocks, duration 00:39:46
• + audio tracks:
• + 1, Unknown (AC3) (2.0 ch) (iso639-2: und), 48000Hz, 224000bps
• + subtitle tracks:
• + combing detected, may be interlaced or telecined

The aspect is 4:3 with a pixel aspect of 16/15.

Another way to pick up the aspect and Pixel Aspect Ratio (PAR) parameters is to use FFmpeg.
For example, you could do something like ffmpeg -i movie.vob :

$ffmpeg -i movie.vob  
 Seems stream 0 codec frame rate differs from container frame rate: 59.94 (60000/1001) -> 29.97 (30000/1001) 
 Input #0, mpeg, from 'mvoie.vob': 
  Duration: 00:03:08.18, start: 0.060000, bitrate: 6939 kb/s 
   Stream #0.0[0x1e0]: Video: mpeg2video, yuv420p, 720x480 [PAR 8:9 DAR 4:3], 8000 kb/s, 29.97 tbr, 90k tbn, 59.94 tbc 
   Stream #0.1[0xa0]: Audio: pcm_s16be, 48000 Hz, stereo, s16, 1536 kb/s 
 At least one output file must be specified 

You'll note the PAR is 8:9 and DAR is 4:3 which matches an NTSC DVD as per above.  Applying a
--pixel-aspect 32:27 in Handbrake would change this to a Widescreen format.

By using HandBrake I was able to re-encode to a compliant MP4 with the correct wide-screen aspect display. The size of the file was slightly larger and the resolution increased slightly from 720x576 (or 704 with cropping) to 1001x576 (with cropping) in Widescreen format. Very handy tool this HandBrake :)

Here's the command I used:

 HandBrakeCLI -v -i source.avi -o target.mp4 --deinterlace --size 700 -e x264 -a 1 -E faac -B 160 -6 dpl2 -R 48 -D 0.0 -f mp4 --custom-anamorphic --pixel-aspect 64:45 -m -x ref=2:bframes=2:subme=6:mixed-refs=0:weightb=0:8x8dct=0:trellis=0   

Here's a quick fire way to get that sound clip, soundtrack or whatever it is into an audio format from a video file. For example, on youtube, there might be a good video clip with a great soundtrack, and the video is nothing more than just a fixed image and/or you want to stick on your iPod/iPhone etc. Or you might want to get an audio sample for use in a ring tone etc.

Below is a script that I use to rip audio files out of pretty much any video.

 #!/bin/bash  
 echo "Ripping audio......."   
 mplayer -nocorrect-pts -vo null -vc null -ao pcm:fast:file=audiodump.wav "$1"   
 normalize -v audiodump.wav   
 faac -b 256 -c 48000 --mpeg-vers 4 -o audiodump.aac audiodump.wav  

The first line uses mplayer to dump the audio to a standard wav file called "audiodump.wav". The "-nocorrect-pts" corrects errors in some certain video files.

I then normalize the audio file via the "normalize" command (make sure you have the normalize package installed).

It's then encoded into AAC format using faac. You could easily do an mp3 here by using lame. For example, change the line starting with faac to something like:

lame --preset extreme audiodump.wav audiodump.mp3