Flash Player to include AVC and AAC and MP4 and TTXT support

[bond]

a big step for the mpeg-4 standard

Adobe announced that it will support the following parts of the mpeg-4 standard (and more) in an upcoming update to their flash player:
- AVC: A software based H.264 codec with the ability to decode Base, Mainline and High profiles...Extended, High 4:2:2 and High 4:4:4 profiles are not officially supported at this time. They might or might not work depending on what features are used. We have no artificial lower limit on B-frames or any problems with B-pyramids like other players do. We also decode field coded streams...the H.264 decoder is...provided to us by MainConcept. It weights in at less than 100KB of compressed code
- AAC: An AAC decoder supporting AAC Main, AAC LC and SBR (also known as HE-AAC). The corresponding ISO specification is ISO 14496-3.
- MP4: An file format parser implementing parts of ISO 14496-12. In terms you might understand this means a very limited sub set of MPEG-4, 3GP and QuickTime movie support.
- TTXT: Support for the 3GPP timed text specification 3GPP TS 26.245. Essentially this is a standardized subtitle format within 3GP files...Any number of text tracks are supported and all the information, including esoteric stuff like karaoke meta data...Language information in the individual tracks is also reported. That means you can have sub titles in several languages.for the details read more here:
http://www.kaourantin.net/2007/08/what-just-happened-to-video-on-web_20.html (thanks to Drachir for pointing me to it) and here:
http://forum.doom9.org/showthread.php?t=129118

a beta version is available here:
http://labs.adobe.com/technologies/flashplayer9/

everything needs to have a downside:
1) mpeg-4 asp (divx, xvid) will not be supported
2) mpeg-4 systems advanced content (basically an open standard alternative to advanced flash content) is not supported
2) the flash player will only support mp4+avc+aac files streamed via adobe servers (but afaik this is the case with all flash content)


As being open standard you should be able to create supported streams with the following free tools:
AVC: x264 (http://www.videolan.org/developers/x264.html) and all tools using x264
AAC: faac (http://sourceforge.net/projects/faac/) and all tools using faac
MP4: mp4box (http://gpac.sourceforge.net/), mp4creator (http://mpeg4ip.sourceforge.net/), ffmpeg (http://ffmpeg.mplayerhq.hu/)
TTXT: mp4box (http://gpac.sourceforge.net/)

[Dark Shikari]

Sweet.

[Schrade]

They have a link to test out the flashplayer once you install it. Here's the link: http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/

It has a streaming video that is 1280x720 in it and it's an MP4 container. You can switch between that or the On2 VP6 encode.

Unsurprisingly, it chows up all my CPU when playing the h.264 encode. I have an AMD Athlon XP 2400+. They need to license CoreAVC or something to make this work right.

[Kurtnoise]

What about the GPAX (http://gpac.sourceforge.net/doc_embed.php) ?


btw, it's definitely the end of story for the Ogg/Theora.

[Dark Shikari]

What about the GPAX (http://gpac.sourceforge.net/doc_embed.php) ?


btw, it's definitely the end of story for the Ogg/Theora.Definitely not, Ogg/Theora serve a particular purpose; they have no royalties. Both H.264 and AAC have royalties, even if most people don't pay them.

Wikipedia, for example, one of the top 10 most popular websites in the world, uses Ogg and Theora exclusively, and as the web moves towards more rich media, they will stick with that simply for legal reasons.

[Schrade]

Does GPAC's plugin have any controls at all yet? Last time I checked it was similar to VLC's plugin I think.. just a video playing, no volume control, no seeking/stop/play/pause controls, etc.

[iwod]

I am now wondering if the Decoder will support Hardware acceleration. But 100K H.264 decoder is pretty amazing

[Kurtnoise]

Definitely not, Ogg/Theora serve a particular purpose; they have no royalties. Both H.264 and AAC have royalties, even if most people don't pay them.
In fact, nothing have been proved that Xiph tools are patents free. Anyway, I tend to agree with you on this point. I missed that.

Wikipedia, for example, one of the top 10 most popular websites in the world, uses Ogg and Theora exclusively, and as the web moves towards more rich media, they will stick with that simply for legal reasons.
Well, what about the nine others ? wikipedia is very special for the contents hosted and proposed on it. It's definitely not a commercial website.

[GmorG McRoth]

They have a link to test out the flashplayer once you install it. Here's the link: http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/


If anyone wants to check videos without flash player:

http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/backcountry_bombshells_4min_HD_1500_96.flv&OBT_fname=backcountry_bombshells_4min_HD_1500_96.flv
http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/backcountry_bombshells_4min_HD_H264.mp4&OBT_fname=backcountry_bombshells_4min_HD_H264.mp4

Is there a way to know what encoder was used to create h.264 file?

[CruNcher]

the thing is ON2 just compares it with VP6 so in practicaly they still claim that VP7 is more advanced then ;)

[Mc Onyx]

Tried the flash link and it looks great, finally some good looking streaming. But the h.264 stream does look better than vp6 and it takes about 25-40% of my cpu time C2D@3.15Ghz.

[Kurtnoise]

Is there a way to know what encoder was used to create h.264 file?
Probably the Mainconcept's (http://www.mainconcept.com/site/news-9/current-news-129/adobe.html) one...

[GmorG McRoth]

Probably the Mainconcept's (http://www.mainconcept.com/site/news-9/current-news-129/adobe.html) one...

Indeed, this is plausible.

[clsid]

If anyone wants to check videos without flash player:

http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/backcountry_bombshells_4min_HD_1500_96.flv
http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/backcountry_bombshells_4min_HD_H264.mp4
This MP4 file plays great with Haali's splitter. I am glad Adobe didn't use a new proprietary container.

The FLV file on the other hand doesn't play properly with Gabest FLV splitter. The video track is not recognized :(

I hope someone can fix this before the new Flash version hits the mass public. Perhaps it only requires a small patch, since the file does seem to contain a normal FLV4 (VP6) video stream, at least that is what GSpot tells me.

[smok3]

wow and wow.

and yes, just tested: things encoded using x.264 + nero aac inside mp4 are working!

short sample (had to rename the mp4 file to flv):
http://somestuff.org/flashAVC/flvplayer.php
(something weird happening on cuts thought, i guess thats why this is beta)

[agilpwc]

This pretty much eliminates the need for QuickTime.

From reading the press release it looks like it will support more encode options, unlike QuickTime.

[Shapierian]

I find the inclusion of the rarely used AAC Main profile odd.

...
- AAC: An AAC decoder supporting AAC Main, AAC LC and SBR (also known as HE-AAC). The corresponding ISO specification is ISO 14496-3.
...

[Sergey A. Sablin]

Probably the Mainconcept's (http://www.mainconcept.com/site/news-9/current-news-129/adobe.html) one...

level 5.1 for 720p and 2 refs, no aspect ratio parameters, no hrd parameters and many smaller details makes me think something different...

[bond]

vp6 didnt handle to beat xvid (mpeg-4 asp), so i heavily doubt it has any chance against a serious avc encoder at all

http://labs.adobe.com/technologies/flashplayer9/fullscreendemo/backcountry_bombshells_4min_HD_H264.mp4&OBT_fname=backcountry_bombshells_4min_HD_H264.mp4

Is there a way to know what encoder was used to create h.264 file?the mp4 mentions apple media video and audio handler, so i assume it has been encoded with apple's avc and aac encoders

[Sergey A. Sablin]

vp6 didnt handle to beat xvid (mpeg-4 asp), so i heavily doubt it has any chance against a serious avc encoder at all

the mp4 mentions apple media video and audio handler, so i assume it has been encoded with apple's avc and aac encoders

Is it you bond? Something strange happens to you...
Have a try to open stream in hex editor. You should know key word.

[Blue_MiSfit]

@Smok3

What x264 settings did you use?

Are there any indication as to which profile specific features are supported - ie 8x8 DCT, CABAC, B-Frames etc?

~MiSfit

[slavickas]

to those lazy to check themself in hex editor/viewer
x264 - core 54 - H.264/MPEG-4 AVC codec - Copyleft 2005 - http://www.videolan.org/x264.html - options: cabac=0 ref=1 deblock=1:4:0 analyse=0x1:0x111 me=hex subme=5 brdo=0 mixed_ref=0 me_range=16 chroma_me=1 trellis=0 8x8dct=0 cqm=0 deadzone=21,11 chroma_qp_offset=0 slices=4 nr=0 decimate=1 mbaff=0 bframes=2 b_pyramid=0 b_adapt=0 b_bias=0 direct=1 wpredb=0 bime=0 keyint=150 keyint_min=25 scenecut=40 rc=2pass bitrate=1600 ratetol=1.0 rceq='blurCplx^(1-qComp)' qcomp=0.60 qpmin=10 qpmax=51 qpstep=4 cplxblur=20.0 qblur=0.5 ip_ratio=1.40 pb_ratio=1.30

[Placio74]

to those lazy to check themself in hex editor/viewer
Or Avinaptic.
[ About file ]

Name: opr0Q3FN.mp4
Date: 22/08/2007 08:33:04
Size: 50,965,091 bytes (48.604 MB)

[ Generic info ]

Play duration: 00:03:59 (239.28 s)
Container type: MP4/MOV
Major brand: ISO Base Media version 1
Creation time: 17/08/2007 00:44:39 UTC
Modification time: 17/08/2007 00:44:43 UTC
Number of streams: 2
Type of stream nr. 1: audio (mp4a) {pple Sound Media Handler}
Type of stream nr. 2: video (avc1) {pple Video Media Handler}
Audio streams: 1

[ About quality ]

Resolution: VERY HIGH (1280 x 720)
Width: multiple of 16 (GOOD)
Height: multiple of 16 (GOOD)
Average DRF quality: LOW (28.389053)
Standard deviation quality: HIGH (2.737534)
Std. dev. weighted mean: HIGH (2.558997)

[ Video track ]

Codec: avc1
Resolution: 1280 x 720 (*)
Frame aspect ratio: 16:9 = 1.777777
Pixel aspect ratio: 1:1 = 1
Display aspect ratio: 16:9 = 1.777777
Framerate: 23.976 fps
Number of frames: 5737
Bitrate: 1603.275462 kbps

[ Audio track ]

Codec: mp4a
Bitrate: 96.149661 kbps
Channels: 2
Sample rate: 44100 Hz

[ About H.264 encoding ]

User data: x264
User data: core 54
User data: H.264/MPEG-4 AVC codec
User data: Copyleft 2005
User data: http://www.videolan.org/x264.html
User data: cabac=0
User data: ref=1
User data: deblock=1:4:0
User data: analyse=0x1:0x111
User data: me=hex
User data: subme=5
User data: brdo=0
User data: mixed_ref=0
User data: me_range=16
User data: chroma_me=1
User data: trellis=0
User data: 8x8dct=0
User data: cqm=0
User data: deadzone=21,11
User data: chroma_qp_offset=0
User data: slices=4
User data: nr=0
User data: decimate=1
User data: mbaff=0
User data: bframes=2
User data: b_pyramid=0
User data: b_adapt=0
User data: b_bias=0
User data: direct=1
User data: wpredb=0
User data: bime=0
User data: keyint=150
User data: keyint_min=25
User data: scenecut=40
User data: rc=2pass
User data: bitrate=1600
User data: ratetol=1.0
User data: rceq='blurCplx^(1-qComp)'
User data: qcomp=0.60
User data: qpmin=10
User data: qpmax=51
User data: qpstep=4
User data: cplxblur=20.0
User data: qblur=0.5
User data: ip_ratio=1.40
User data: pb_ratio=1.30
SPS id: 0
Num ref frames: 2
PPS id: 0
Entropy coding type: CAVLC
Weighted prediction: No
Number of frames: 5737
Drop/delay frames: 0
Corrupted frames: 0

P-slices: 2053 ( 35.785 %) #########
B-slices: 3622 ( 63.134 %) ################
I-slices: 62 ( 1.081 %)
SP-slices: 0 ( 0.000 %)
SI-slices: 0 ( 0.000 %)

[ DRF analysis ]

Average DRF: 28.389053
Standard deviation: 2.737534
Max DRF: 35

DRF<20: 0 ( 0.000 %)
DRF=20: 7 ( 0.122 %)
DRF=21: 16 ( 0.279 %)
DRF=22: 19 ( 0.331 %)
DRF=23: 133 ( 2.318 %) #
DRF=24: 221 ( 3.852 %) #
DRF=25: 453 ( 7.896 %) ##
DRF=26: 738 ( 12.864 %) ###
DRF=27: 637 ( 11.103 %) ###
DRF=28: 885 ( 15.426 %) ####
DRF=29: 608 ( 10.598 %) ###
DRF=30: 555 ( 9.674 %) ##
DRF=31: 569 ( 9.918 %) ##
DRF=32: 546 ( 9.517 %) ##
DRF=33: 198 ( 3.451 %) #
DRF=34: 128 ( 2.231 %) #
DRF=35: 24 ( 0.418 %)
DRF>35: 0 ( 0.000 %)

P-slices average DRF: 27.172430
P-slices std. deviation: 2.635232
P-slices max DRF: 34

B-slices average DRF: 29.121479
B-slices std. deviation: 2.503479
B-slices max DRF: 35

I-slices average DRF: 25.887096
I-slices std. deviation: 3.278005
I-slices max DRF: 33

This report was created by AVInaptic (10-04-2007) on 22 ago 2007, h 08:57:01

[smok3]

Blue_MiSfit: lowb and highb are quicktime compatible, the ipodlike is the easiest one:

namely, qt:
x264.exe --pass 1 --bitrate %bit% --stats "%inpfullnoext%.stats" --bframes 1 --subme 1 --analyse none --me dia --progress --no-psnr --filter %fil% --output NUL

and ipod:
x264.exe --pass 1 --bitrate %bit% --stats "%inpfullnoext%.stats" --level 1.3 --no-cabac --subme 6 --analyse p8x8,b8x8 --qpmin 16 --vbv-maxrate 768 --vbv-bufsize 244 --me umh --merange 12 --threads auto --thread-input --progress --no-psnr --no-ssim --output NUL

[bond]

Is it you bond? Something strange happens to you...
Have a try to open stream in hex editor. You should know key word.man :eek:

shouldnt be that fast with my conclusions ;)
guess they remuxed the stream with quicktime?

[Corben9]

For the laziest of us a profile for MeGui vould be nice :devil:

[Corben9]

I made a demo in a skate edit i recently made under

http://www.studioc9.com/CevapiSession

under the HD version. Now it all works nicelly locally but form the server for some odd reason it does not ... Now as i understood it the file should start playing progressivey .. since i encoded the entier thing in 720p 4Mbit it is a rather high bitrate but nothing fast net shouldnt be able to handle... now the file is 20 min long so i was counting on the progressive downlaod and playback.... but it aint working.... I tried a few other test files and it waried form working to not working :mad:

id like some oppinions ... id really like to keep the h264 version in prestine quality ,:):):)

[Hellworm]

If you use progressive download instead of FMS make sure that the moov atom (which is the index information in MPEG-4 files) is at the beginning of the file. Otherwise you have to wait until the file is completely downloaded before it is played back.

http://www.kaourantin.net/2007/08/what-just-happened-to-video-on-web_20.html

[Corben9]

Jup I think thats the problem right now since i muxed the file in MeGUI .. i made another mux with MP4Box coz its supposed to put the moov at the start even by default ... Am I correct?

[Schrade]

Yeah, in YAMB 2.0.0.7 it's checked by default I believe.

http://i12.tinypic.com/5y123c0.png

[bond]

Jup I think thats the problem right now since i muxed the file in MeGUI .. i made another mux with MP4Box coz its supposed to put the moov at the start even by default ... Am I correct?megui is "just" the gui. important is what the tools used in the background do:
1) x264 is a video encoder writing the video stream info directly into the mp4. doing so the header will be written at the end of the mp4, as the header is only known at the end of encoding
2) for muxing the video with audio mp4box is used in megui. mp4box writes the header at the beginning of the file with the default settings

therefore your mp4 should already have the header at the start. you can check this by running the mp4dump tool from mpeg4ip over the file. if the moov atom (the headers) is placed before the mdat atom (the plain video data) your header is at the start

edit: are you trying to stream this file:
http://www.studioc9.com/CevapiSession/CevapiSession.mp4

if yes, this seems to be borked. how exactly did you produce this?

[Corben9]

hej ! At the time u wrote the post i was just upoading the version I muxed via comandline on mp4box.

I produced it as follows>

Adobe Premiere for editing cuting ect exported as m2t hdv compliant transport stream so i can back up my master on tape.

then I unmuxed the m2ts with DGindex

Then using a avisnyth script and the Tdent and Spline36resize comand I had it frameserved to megui where i encoded at 25 fps High level x264 in mp4 at 4 Mbit bitrate .. no extra settings applied just standard high profile with 4 Mbit 3 pass encoding...

I also transcoded the audio in megeui with Winamp AAC ... then i muxed everything with the megeui mux tool and now again manually with MP4Box...

I also thought that it might be player related so i set manual buffer parameters but that didnt help eitehr...

I also tried to play it directly form the web with vlc it also transfers data but the screen stays black :confused::confused::confused:

can u please check the file and see if its ok im doing another upload coz the file might have gotten corrupted while upoading via FTP *gonna be up in like an hour*... it still doesnt work from the server but locally it plays fine over flash and VLC but it doesnt seem to work in Quicktime .. i still think that its a moov atom related issue but honestly I just dont know anymore...:confused::confused::confused::confused:

If its borked and i have to encode again what settings should i use to make sure its compliant ... ?

[bond]

edit: are you trying to stream this file:
http://www.studioc9.com/CevapiSession/CevapiSession.mp4

if yes, this seems to be borked. how exactly did you produce this?are we talking about this file? this one is borked

try encoding the aac with faac or nero or quicktime and mux the resulting audio and your video to a new mp4 (so do not mux one of the streams to an already existing mp4)

[Schrade]

Is there some way to detect if a file has its media information indexed at the beginning or not?

[bond]

therefore your mp4 should already have the header at the start. you can check this by running the mp4dump tool from mpeg4ip over the file. if the moov atom (the headers) is placed before the mdat atom (the plain video data) your header is at the starthere it goes

[Schrade]

Woops! Missed that. Thanks, bond. I was up really late last night when I first read your post where you said that. Thanks for pointing it out to me.

Is there some benefit for putting the moov atom after the mdat one? Why not just have all files with the moov first?

I tried it with QuickTime Atom Viewer and it works! Thanks for the tip on what to look for.

Here's standard Avidemux output, won't stream:
http://i18.tinypic.com/62nb6kp.png

And here's an MP4 that will stream:
http://i1.tinypic.com/4qoe5gm.png

[Corben9]

are we talking about this file? this one is borked

try encoding the aac with faac or nero or quicktime and mux the resulting audio and your video to a new mp4 (so do not mux one of the streams to an already existing mp4)

Sorry to bug again and for been such a noob ... i have a version up now that should have the moov atom at the start and it works localy and over my home network if is stream it to anther comp... but it wont work form the server .. can u please check it .. its at the same place as before

http://www.studioc9.com/CevapiSession/CevapiSession.mp4

And another question regarding reencoding... when I encode the video to x264 what to you mean exactly with dont mux the aac to an already existing mp4?

Here is a screenshot from the begining of the file:
http://www.studioc9.com/CevapiSession/File.jpg

It looks rather different form the one Shrade posted that does stream... question remains how to edit the header to make it work ?

[Schrade]

You need to remux it into an MP4 again using YAMB 2.0.0.7. Make sure the option I showed in my post with the YAMB picture is checked.

I just checked and your moov atom is below the mdat one, it needs to be before it.

[Corben9]

I remuxed it again with all the options and checked, and the moov atom is still at the complete end of the file ... i really don't know why ... anyone know a way to manually move it or select the order they should be applied with .???

I find it odd since the options are turned on and it doesn't move it to the begining .. id really like a way to control these tags :confused:

[Hellworm]

On this (http://www.kaourantin.net/2007/08/what-just-happened-to-video-on-web_20.html) page that was posted several times in this thread is a link to a program (http://svn.mplayerhq.hu/ffmpeg/trunk/tools/qt-faststart.c?view=markup) that does only move the atom to the beginning.

[Corben9]

I did apply QT index swapper at the file begining and the moov atom is also there ... i checked.. yet for some odd reason it doesnt start playback, it does play ok in vlc... could it be something regarding timeouts in flash itself ????

any idea if there is a way to buffer the playback to get it working .. its supposed to work as a progressive downlaod now yet it doesnt start playing at all .. i have a 8 Mbit conection and the file is 4 Mbit a sec yet it downlaods at max but doestn play progressivelly ...

im really runnin out of ideas..

[bond]

you seem to use rather big files, so it might be that mp4box switches to 64bit times and atoms. it might be that the flash player has problem with those

try some small bitrate/resolution stream created exactly the same way as your normal samples (same settings, same tools...) and report if they cause problems in flash

[Corben9]

It might be the reather big bitrate... I just started another encode form the orignal at a bitrate avorage og 1750 at the same res the adobe demo was 960x540 ...

Since my current file works locally it might be that the player times out before it reads the entiere atom data on large files ... that vould explain why it works over my wifi and not over net .. but i do have a 10 Mbit pipe and my server also has 100Mbit but still ...

ill let u know .. this encoding is gonna eat up another day :mad::mad::mad:..

===========

I also discoverd something strange right now .. seems that my server for some reason doesnt give out the filesize of the files .. if anyone knows, could that be the reason why it even if the moov atom is at the begining the file doesnt stream ?.. the FLV version does stream progressivly eventouh if u try to save it the filesize is unknown...

[Schrade]

On this (http://www.kaourantin.net/2007/08/what-just-happened-to-video-on-web_20.html) page that was posted several times in this thread is a link to a program (http://svn.mplayerhq.hu/ffmpeg/trunk/tools/qt-faststart.c?view=markup) that does only move the atom to the beginning.

I just compiled this and tested it. It doesn't work on MP4s with the PSP atoms as output via the PSP container in Avidemux. I tested with a regular MP4 and it worked.

[Corben9]

I just upoaded the smaler HD version i encoded at 960x540 moov atom at the beginign YET IT DOESNT STREAM ... im runnin out of ideas really... Can someone check the file to see what im doing wrong :mad::mad::mad:

[bond]

did you add hint tracks?

[Corben9]

No. What are hint tracks... are they like markers and if they are important for the streaming how do i add them correctly ?

Update1>

I hinted my file with mp4box -hint file.mp4 and uploaded it again .. yet still id doesnt work :(

[giandrea]

No. What are hint tracks... are they like markers and if they are important for the streaming how do i add them correctly ?

Update1>

I hinted my file with mp4box -hint file.mp4 and uploaded it again .. yet still id doesnt work :(

Perhaps your file is too long. Try to split it to like 15 minutes with MP4box and see if it "streams" (well, progressive downloads).
Use:
MP4Box -split-chunk 0:900 your_file.mp4

[iwod]

Any news on Hardware Acceleration Support?

[Pacoup]

Ok. Not very impressive AVC content. Quite deceiving in fact.
And almost identic VP6 content.
They should just go and implement VP7 instead if they can't implement h.264 correctly.

btw. Quicktime is so bad at implementing h.264 that the same bitrate turns out as much as 2 times worse then Theora video, via Quicktime encode.

In my personal tests for codecs VP7 and h.264 are largely the best video codecs, really close to one another and difficult to decide which is the best (VP7 from On2's free edition and x264 through mEncoder)
although, VP7 is way faster to encode

btw, sorry for the newb, baseless post with all sorts of comments lol



Oh and, I got full VC-1 and AVC hardware acceleration on my ATI Radeon HD 2600 XT.
Here's the tests:
http://www.hardware.fr/articles/690-4/dossier-quelle-carte-graphique-video-hd.html
http://www.hardware.fr/articles/693-4/dossier-amd-radeon-hd-3800.html