MPEG2, X264, Xvid HD 1080i - The Lowdown

[Pensive]

Good afternoon gents


I'm going to try and stop the confusion (my own too - im researching as i go here so feel free to correct me) once and for all regarding HD content:

http://en.wikipedia.org/wiki/High_definition_television

MPEG2 is the youngest of the 3 formats, it has no theoretical bitrate limitations, is completely editable on the timeline (in HDV format), and is the format used to film in, and to edit low grade HD content in the media.

MPEG2 is also used to encode DVD videos, but this is a different application of the codec.

You can find examples of HDV MPEG2 work on every music video, they are all done with HDV cams. Also Robert Rodriguez filmed and worked with MPEG2 to make Once upon a Time in Mexico, the third film in the El Mariachi series. He also edited the entire movie and did all the effects himself at his ranch, in his home studio.

(if you look carefully at Johnny Depps eyes you'll notice a few mistakes Robert made during his first foray into proper digital effects - but they're barely noticeable).

MPEG 4-AVC is a specification designed for highly compressing and playing back HD content from the new blu-ray and HD-dvd formats, as well as streaming content through various mediums. The format is equally flexible as MPEG2, however, its benefits over MPEG2 exist mainly in HD content playback, where it provides higher quality and resolution at the same bitrates. The exception to this is its Lossless mode -where it has the capability of superceding MPEG2 in editable

Xvid is MPEG4, in a very simple form, a low bitrate high compression format with a few benefits over MPEG2 pertaining to compression ratios, but exhibits colour saturation, and noise which is not present in higher bitrate MPEG2 for example. My experiments in raising bitrates have not seen a reduction in these detrimentary artefacts.


What are Xvid, H264 and x264?
________________________________________

Xvid is the open source version of MPEG4, the lowest quality part of the MPEG standard - its designed around a movie file being lower or equal resolution than a dvd, and lower size, but similiar quality. It can be used to provide HD content however, although it is not really designed to do this.

http://en.wikipedia.org/wiki/Xvid

One can quite clearly tell when Xvid has been used to display HD content instead of MPEG4-AVC, from the colouration, and speckling artefacts.

X264 is the open source implementation of the High bitrate HD res version for the standard (MPEG4-AVC). It exhibits good colour management, great compresion ratios, and generally provides extremely high quality results in an HD playing field.

http://en.wikipedia.org/wiki/X264

H264 is simply the part of the MPEG4 AVC standard that relates to the video encoding of HD content. Its the specification for encoding - which x264 is trying to implement.

http://en.wikipedia.org/wiki/H264

WMV-HD uses VC-1, which is microsofts own proprietary HD codec, built into the Windows Media Video spec, and is actually really good quality too. Double disc sets with a WMV-HD Normal DVD disc accompanying the MPEG2 DVD are on sale in America, but people are reportedly having enormous trouble playing them. DRM issues have been reportedly been a pain and detrimental to the cinematic experience. They have however been largely successful in maintaining region coding control. For now.





Why only x264? What about all the other H264 codecs?
______________________________

X264, like XviD is open source. Going from my past experience, while open source is a little flaky at first (like all software), it is almost certain to culminate in a higher quality codec at the end of the day, than anyone elses.

Examples: Lame MP3, Xvid MPEG 4

Why? There are no budgets - noone is going to say "its good enough - if we spend any more money on it - we wont make a profit". Its done for the love, and anyone can get involved. The only thing that might happen is someone will say "its good enough - any more work will not bear any fruit". Which means its good enough already, and cant really be improved on much. Top Banana!

I have ignored the multitude of alternative H264 codecs since open source is the high and mighty ruler and will eventually 0wn all

Also there are 20 or 30 of them so any test would be painful and tearful. Feel free to add any results of tests you have performed.


I have been experimenting with converting an interlaced source to various progressive HD-DVD (M$ H264 equivalent) resolutions and found adaptive deinterlace - while slow - to be the highest quality result from an interlaced source.

Remembering that H264 does not lend itself directly to interlaced video, I would suggest that with quality tests - motion adaptive deinterlacing should be used. The principles of motion adaptive deinterlacing can be seen here - explained clearly.

http://www.stanford.edu/class/ee392j/projects/marc/HybridEdgeMADeinterlacer.ppt

[anonymez]

hello pensive,

MPEG2 is the highest quality of the 3 formats, it goes up to 80mbit/s, and further ( lossy ), is completely editable on the timeline, and is the format used to film in, and to edit low grade HD content in the media.

so you claim x264 cannot match mpeg2 in quality? . x264 cli has no bitrate limit-- has quant and lossless modes.

MPEG2 is designed for working on HD video and editing the video.

H264 is designed for both high and low resolutions and bitrates

X264 is the open source implementation of the High bitrate HD res version for the standard (MPEG4-AVC). Its designed to be higher bitrate than a dvd, bigger than a dvd, but higher resolution and better quality.

x264 is primarily used for dvd and hdtv backup. usually encoded at less than half the bitrate of the source.

It claims to be 1080i but it is interlaced (guess thats what the i stands for) - so it is in fact 1080p, so 1080p is the target format. Its the only thing that makes sense.

depends. if it's true interlaced there will only ever 540 lines of info every 50th or 60th of a second. but if originally progressive and converted to 1080i properly then its real 1080p when weaved

System Requirements: To play back the clip, you will need a Pentium 4 3Ghz with Hyperthreading really - and even then itll only play well in Media Player classic, which is the greatest media player ever

mpeg2 1080i playback is possible on my p4 2GHz (without HT). all thats needed is dxva compatible graphics card and decoder. although mpc is easily my favourite player, almost any other player can match its speed as long as the correct decoder is used.

i know xvid inside out, but x264 is not my forte. What is the best software to use, which settings? And should I deinterlace, with which method?

the best method imho is avisynth + x264 cli. you'll want to use meGUI. too many settings, will post my usual later if you wish. as for deinterlacing i'd like avisynth mvbob(), but i guess a tweaked tdeint() is also acceptable.

Possible softwares i have in mind are virtualdubmod and FlaskMPEG, I can also try Premiere, I'm even going to give canopus procoder a run if i can make it play ball with me (although for an expensive transcoder its rather flaky).

if you use vfw to encode x264 i'm afraid your results shall be invalid due to its limitations. cli only. flaskmpeg is horribly outdated, slow and relatively featureless. vdubmod is great for vfw, if not also a little outdated in some areas, but not for this.

Tell me what YOU think, which frames of the video should be compared between the three files, all that sort of thing, give me your input.

high motion, fog, fire, dark, i frames

Deinterlace frame with motion adaptive blending

elaborate? :^)

[Pensive]

hello pensive,
so you claim x264 cannot match mpeg2 in quality? . x264 cli has no bitrate limit-- has quant and lossless modes.

Well I dont think it can match it in some ways, certainly.

I don't claim to fully understand the encoding techniques differences for MPEG2 and 4 formats, its just that mpeg2 effectively takes out less video data and achieves less compression ratio, so thats why I stand by my statement.

MPEG2 has no theoretical bitrate limitation either of course, but there are practical limits to a codecs useful operation.

Lossless is lossless. Have you ever worked at this resolution on lossless 60fps video? I think its important to keep an argument practical for its application.
You'd need a RAID array at level 5 of SATA disks alone just to get at the data quick enough, its just not practical. My suggestion is that MPEG4 was around at the same time they chose MPEG2 for an eiditing platform. Perhaps in the future more support for a customised version of mpeg4-AVC which is editable will arise. But there's really nothing wrong with MPEG2 and it already ahs a lot of software support.

I cant see any improvement on the quality through higher compression ratios at all, and those higher compression ratios would be lost if the MPEG4 was frame editable anyway. So I dont think there would be any benefit of going down this route, I believe its horses for courses here. perhaps im mistaken?

H264 is designed for both high and low resolutions and bitrates

true, but that doesnt mean it handles them as well as existing technologies.
"JAck of all trades" springs to mind, although i have no evidence to back that up _yet_.

x264 is primarily used for dvd and hdtv backup. usually encoded at less than half the bitrate of the source.

To the best of nmy knowledge only in the illegal torrent world, and home copyright infringement markets, and I for one have no interest in including them in my study.

H264 is primarily used in the legal world by Microsoft for WMV-HD DVDs, and by NTL and other cable companies to distribute tv channels highly efficiently. They dont use MPEG2 anymore as far as i know. When they upgrade your set top box software they can patch the codecs too.

Also it is used for HD-DVD/blu ray videos - although its very important to note that both standards also support MPEG2 fully as well, this may be to save people re-encoding movies, but there may also be other reasons. Both technologies are extremely high quality either way.

Do they use H264 for hardware TV recording equipment these days? I figured they'd use SD for that.

depends. if it's true interlaced there will only ever 540 lines of info every 50th or 60th of a second. but if originally progressive and converted to 1080i properly then its real 1080p when weaved

Hmmm 540 lines @60 fps = 1080 lines @ 30 fps when weaved. Which is what I said. ?? have i missed something here? The total amount of visual data is equivalent, in my eyes, and I'm pretty sure about that.

1080i and 1080p at simliar bitrates should give similiar quality, its jsut one is interlaced and one isnt. I called it 1080p because as far as i know H264 will perform better with non-interlaced sources.....perhaps that assumption is incorrect.

mpeg2 1080i playback is possible on my p4 2GHz (without HT). all thats needed is dxva compatible graphics card and decoder. although mpc is easily my favourite player, almost any other player can match its speed as long as the correct decoder is used.

Really? Even this clip? I wasnt talking in general - I was talking about the 80mb/s clip linked there - did it work for you with those specs? Which player did you use? is that CoreAVC working its wonders? I havent tried it yet.....

the best method imho is avisynth + x264 cli. you'll want to use meGUI. too many settings, will post my usual later if you wish. as for deinterlacing i'd like avisynth mvbob(), but i guess a tweaked tdeint() is also acceptable.

But bob deinterlacing would render out to 120fps, then back down to 60, and reinterlace - there is no 60fps progressive mode in H264.......!!!??? surely a pointless exercise? I belive better results come from transcoding more effciiently in this situation

if you use vfw to encode x264 i'm afraid your results shall be invalid due to its limitations. cli only. flaskmpeg is horribly outdated, slow and relatively featureless. vdubmod is great for vfw, if not also a little outdated in some areas, but not for this.

ahhh. Showing my age there...(27 btw)..thank u. I will use your suggested software for my x264 encode.

high motion, fog, fire, dark, i frames

Sounds good.

elaborate? :^)

Motion adaptive uses an effective field combine method for stable parts of the image, and provides a ghosting where motion (change between fields) is detected, which results in fractional blurring on high motion - which is actually what I was after. Anything stable stays pin sharp however.

I would like to see the result of BOB deinterlacing then rendering down to 30fps progressive, I wonder if it makes much difference?

the below clips are already encoded - let me know if you'd like to have a look - i can put them up on my website for you if you like.

WMV-HD encoding to 30fps non interlaced (progressive), QVBR at 75 quality.
1)With Adaptive Deinterlace
2)Without Adaptive Deinterlace

3) I also did MPEG2-HD (non editable, MP@HL) 1080p, and as expected it was 28mb where the MPEG4 was around 22Mb, and perceptibly slightly lower quality (not a lot, still looked awesome), which backs up both of our arguments - mine that mpeg2 performs better at HD+higher bitrates than lower ones, and yours that MPEG4 performs better at HD res/lower bitrates. I should note that it played REALLY nicely and smoothly though - where the other WMV-HDs did not play so reliably - they stuttered once every 10 seconds (only minutely). I actually really enjoyed watching the MPEG2 one. My Nvidia 256mb Geforce 7300 gs supports both mpeg2 and mpeg4 hardware accel @ HD, so it cant have been that (or shouldnt at least)

Finally I tried WMV-HD to 60fps interlaced, QVBR @ 75 Quality, with no deinterlacing, but was unable to play the result. If i can't play it on my new beasty, then whats the point? It loads but it doesnt play back smoothly at all..... and its twice the size. perhaps i should try AVC Core......

Ill set the clips i have encoded uploading to my site cos id bet you'd be intersted to see what I'm talking about.

Thanks for your input - sorry for the officially Long-Ass Reply(tm)....

[anonymez]

Have you ever worked at this resolution on lossless 60fps video? I think its important to keep an argument practical for its application.
You'd need a RAID array at level 5 of SATA disks alone just to get at the data quick enough, its just not practical.

yes i have tested it previously out of curiosity. played just fine

But there's really nothing wrong with MPEG2 and it already ahs a lot of software support

can't argue with that-- mpeg2 will be around for a while yet

true, but that doesnt mean it handles them as well as existing technologies.

x264 is still under heavy development, who knows what will come in the next 6 months ;^)

To the best of nmy knowledge only in the illegal torrent world, and home copyright infringement markets, and I for one have no interest in including them in my study.

i back up dvd and hdtv to x264 for htpc playback. i work with 1080i content every day-- i'm no pirate

Really? Even this clip? I wasnt talking in general - I was talking about the 80mb/s clip linked there - did it work for you with those specs? Which player did you use?

other than the bitrate, that clip is no different to what is being broadcast in australia. the clip is not 80mbps, more like 30. yes it worked with thos specs. use a dxva codec like cyberlink, intervideo, nvidia purevideo, etc. i prefer dscaler+ffdshow for quality though, but much much slower, forced into software mode.

is that AVC Core working its wonders?

you mean coreAVC? works wonders with avc, but doesn't decode mpeg2.

But bob deinterlacing would render out to 120fps, then back down to 60, and reinterlace - there is no 60fps progressive mode in H264.......!!!??? surely a pointless exercise? I belive better results come from transcoding more effciiently in this situation

the clip is 30fps. bobbing will create a 60fps file, dunno where you got 120fps from. 60fps is how the clip is meant to be viewed, i'm sure its being bobbed when you play it on your pc. at time of writing x264 doesn't do interlaced encoding (not practical at HD res anyway, even with coreAVC!)

Motion adaptive uses an effective field combine method for stable parts of the image, and provides a ghosting where motion (change between fields) is detected, which results in fractional blurring on high motion - which is actually what I was after. Anything stable stays pin sharp however.

lol i know what it is; i wanted to know what's doing the bobbing

imho the source isn't of high enough quality regardless of the bitrate-- its a little dark. up to you whether you want to use them, but something a litle more suitable: http://demod.dvico.com/hdtv/Australia_1.tp or http://demod.dvico.com/hdtv/Australia_2.tp

[Pensive]

goodness me you're right - dont know where that 60fps came from either.

or the 80mbit/s thing......must have been tired, jeez......

Im using procoder thus far but its never going to work with Xvid or x264 - already tried xvid - just froze.

I'm not convinced there's really much point doing this to be honest. Ive watched all the clips I encoded - they're all such good quality - the only issues i appear to have are playback which can most likely be solved with different players, I believe i have embarked on a pointless exercise.

WMV-HD colour is far richer than the first HD sample I saw from Apple, so I think my main gripe with it is unfounded.

Motion adaptive deinterlace looks better than no deinterlace (plain BOB), i believe. without the deinterlace it never really sharpens up the same. I have succeeded in doing HD quality tests with interlacing techniques, so its not been a waste for me though. Plus Ive found that procoder deals with HD really well - something i wasnt expecting.

Hey ho, we live and learn, onward to the future and bring on x264 - in editable form!

[anonymez]

Motion adaptive deinterlace looks better than no deinterlace (plain BOB), i believe. without the deinterlace it never really sharpens up the same.

exactly, but tdeint() and mvbob() do dumb & adaptive, the latter is known to be the best, though it'll bring any dvd res encode down to 1fps, 2 on a dual core

Hey ho, we live and learn, onward to the future and bring on x264 - in editable form!

glad you've seen the light ;^)

[popper]

with all due respect, your talking rubbish, lets have some real facts and try and correct this thread from now in case someone takes your inacurate and down right wrong text so far as true facts.

"What is MPEG-4?

MPEG-4 (ISO 14496) is a broad Open Standard developed by the Moving Picture Experts Group (MPEG), a working group of the International Organization for Standardization (ISO) which also did the well known MPEG-1 (MP3, VCD) and MPEG-2 (DVD, SVCD) Standards, standardizing all sorts of audio/video compression formats and much more
By its nature the MPEG-4 Standard doesnt aim at standardizing one potential product (eg something comparable to DVD) but covers a broad range of Sub-Standards, which Product Providers can choose from to follow, according to what they need for their product

The MPEG-4 Standard, as mentioned, is divided into many different sub-standards, where for us users on Doom9 the following parts might be of major interest:
- ISO 14496-1 (Systems), Animation/Interactivity (like DVD Menus)
- ISO 14496-2 (Video), e.g. Advanced Simple Profile (ASP), as followed by XviD, DivX5, 3ivx...
- ISO 14496-3 (Audio), Advanced Audio Coding (AAC)
- ISO 14496-10 (Video), Advanced Video Coding (AVC), also known as H.264
- ISO 14496-14 (Container), MP4 container format (uses the .mp4 extension)
- ISO 14496-17 (Subtitles), MPEG-4 Timed Text subtitle format

This information thread now aims at providing some usefull infos on most of these parts, with a focus on MPEG-4 ASP and AVC/H.264"

"
ISO 14496-10 (Video) - Advanced Video Coding (AVC)
<HR style="COLOR: #ffffff" SIZE=1><!-- / icon and title --><!-- message -->With AVC/H.264 the MPEG-4 Standard defines one of the newest and technically best available, state-of-the-art Video Coding Formats

The AVC/H.264 Video Coding Standard was together finalized and identically specified in 2003 by two Groups, the MPEG (Moving Pictures Experts Group) from ISO and the VCEG (Video Coding Experts Group) from ITU (International Telecommunication Union), a suborganisation of the United Nations (UNO), which also standardised the H.263 format (mainly used in video conference software now)
The AVC/H.264 Standard itself was developed by the Joint Video Team (JVT), which included experts from both MPEG and VCEG

Looking from the MPEG side the standard is called MPEG-4 Part 10 (ISO 14496-10), looking from the ITU side, it is called H.264 (the ITU document number), by which the format is widely known already
As "official" title for the new standard Advanced Video Coding (AVC) was chosen by MPEG - as video counterpart to the Advanced Audio Coding (AAC) audio format"

http://forum.doom9.org/showthread.php?t=96059

An overview of AVC/H.264 compared to other popular video coding formats:


available AVC/H.264 Codecs

AVC/H.264 implementations are available atm already from x264, Nero, Apple, Sorenson, Elecard, Moonlight, VSS, mpegable, Envivio, Hdot264 (binary), DSPR, JM (reference software) (binary), ffmpeg, Philips, FastVDO, Skal, Sony and many more "

[Pensive]

with all due respect, your talking rubbish, lets have some real facts and try and correct this thread from now in case someone takes your inacurate and down right wrong text so far as true facts.


An overview of AVC/H.264 compared to other popular video coding formats:

Thats a lovely table, and very useful - thanks for posting it up.

My point is to most people that table is totally useless, im trying to make english sense out of the formats, with practical application examples.

See my amendments I have made above, based on my interaction with anonymez over the last two posts.....plus im not bothering with frame based analysis since its entirely dependant on bitrate and quality requirements and codecs are becoming so flexible that these side by side tests are no longer sensible.


But what difference does Multiple Reference Frames make? your post does not address this. Thats what my thread is supposed to cover without dwelling on the teccy points too deeply.....

It means that with a trade off of dotting around anyone of 20 or 30 preceding frames, it can make even more efficient use of the disk space - which has no knock on effect on visual quality (in itself) apart from leaving more bitrate free for other information; At the same time however, it totally undermines the versatility of the stream (with regard to potential applications). It now needs to refer back to a number of other frames to render this frame - far more than MPEG2 ever had/has to.


now....read this:

http://en.wikipedia.org/wiki/HDV

Ive worked in the editing industry for sometime - so where my codec engineering knowledge is lacking i make up ofr with a feel for requirements.


QPEL, GMC and 8*8 Motion blocks, and Multiple Reference Frames will be all but useless in a frame editable application. Their effects on the admin overhead in the stream could be also highly inefficent. Discuss.

I'd like Popper to reply to this with his thoughts please.

It'd be a nice idea to build a simliar table comparing Xvid, X264, HDV and ProDV, and MPEG4-AVC Lossless to suit - a project I'm going to begin in earnest tomorrow.


Also - this is interesting - it indicates the the HD editable solution currently in place doesnt quite cut it - leaving room for MPEG4-AVC lossless in the future.....scrummy.

http://en.wikipedia.org/wiki/ProHD

[Pensive]

yes i have tested it previously out of curiosity. played just fine

But a standard requirement in the editing industry is at least 5 streams of realtime rendering all at once.

Canopus DV Storm cards can render 5 streams at once out to DV from a raid array.

Imagine the tranfer rates and HD arrays required for 5 realtime streamsof MPEG4-AVC lossless? in fact can you calculate that for me please?

I have no idea what kind of compression they will use - likely not Run Length Encoding a la bitmaps surely......but they cant use a Huffman technique because each frame will have to be autonomously encoded otherwise realtime performance will suffer enormously.

This does not count for pro HD mixing as that will be done with dedicated hardware, and a dedicated drive per stream, i mean for the semi-pro eidting market, who want 24fps HD progressive editing, without buying a roomful of hardware.

I also dont care to read the spec and find out - I'd rather ask you for a practical test result - what kind of bitrate did it require?

[Zodiaque]

1) M$ will use VC-1 Advanced Profil for HD-DVD and BR. VC-1 is not MPEG4 AVC codec but standard completely different.

2) Quick Time AVC codec is very crap H264 codec. Elecard, Ateme or x264 are really better and by far. Anyway even QT7 AVC done excellent result for 1080p at 8-10 Mbps for really complexe source like trailers. IMO good H264 codec will encode more than 1080 i/p 120 min on simple DVD5 ... MPEG2 can't make that.

3) Use 24 Mbps for H264 or VC-1 is completely useless simply because at 1080 i/p all codec done similar result and MPEG2 too. With same source quality by pixel will be equivalent between 5 Mbps 576p MPEG2 and 15 Mbps 1080p MPEG2.

4) MPEG4 AVC is simply actually the most efficient codec in the world and it's true particulary for "low bitrate". With this codec you can make that:

http://jfl1974.free.fr/upload/superman.mp4

http://jfl1974.free.fr/upload/XMenIII.mp4

And my simple Sempron 2.1 Ghz can play that (video 1920*1088*24 at 2 Mbps) very well with CoreAVC and CoreAAC filter without hardware acceleration ...

[Pensive]

1) M$ will use VC-1 Advanced Profil for HD-DVD and BR. VC-1 is not MPEG4 AVC codec but standard completely different.

typical microsoft - i better edit my text above......thanks for the info

IMO good H264 codec will encode more than 1080 i/p 120 min on simple DVD5 ... MPEG2 can't make that.

but you cant edit H264 Lossy. You cant really compare MPEG4-AVC and MPEG2-HDV like that - thats the whole point of this thread....

3) Use 24 Mbps for H264 or VC-1 is completely useless simply because at 1080 i/p all codec done similar result and MPEG2 too. With same source quality by pixel will be equivalent between 5 Mbps 576p MPEG2 and 15 Mbps 1080p MPEG2.

Rubbish - MPEG2-HDV is 25mbitps - thats 3.5Mb/s, it was chosen because that was the point where perceptibly the MPEG2 codec reached its practical limit. If there was no perceptible difference between 25mbps and 10mbps, dont you think they would be using 10 mbps? That of course is a rhetorical question.

noone has addressed my point about things being editable. Again you must address the application.

4) MPEG4 AVC is simply actually the most efficient codec in the world and it's true particulary for "low bitrate".

Yes this is true. This does not however make it a one stop shop, if you read the entirety of this thread, or even my previous two posts you will see why.

Efficiency has a price.

[Zodiaque]

but you cant edit H264 Lossy. You cant really compare MPEG4-AVC and MPEG2-HDV like that - thats the whole point of this thread....

Well editing is not codec problem but simply container problem. You can choose for H264 exactly the same GOP structure than actual MPEG2. You can place Iframe like you want and exactly like the other codec. Editing H264 is really not a problem ... ???

Rubbish - MPEG2-HDV is 25mbitps - thats 3.5Mb/s, it was chosen because that was the point where perceptibly the MPEG2 codec reached its practical limit. If there was no perceptible difference between 25mbps and 10mbps, dont you think they would be using 10 mbps? That of course is a rhetorical question.

Doesn't mean anything. Transparent encoding threshold is not bitrate problem but MPEG2 encoder efficiency and Quantizer threshold problem. I can find really complexe source with heavy perceptible difference at 25 Mbps for 1080p encoding. With source A and encoder X transparency can be at 5 Mbps and with source B and encoder Y transparency can be at 50 Mbps. Transparency is certainely not at 25 Mbps for MPEG2 at 1080i/p (with CBR mode I imagine).

[Pensive]

Well editing is not codec problem but simply container problem. You can choose for H264 exactly the same GOP structure than actual MPEG2. You can place Iframe like you want and exactly like the other codec. Editing H264 is really not a problem ... ???

Thats possibly true, but what about generation loss?

The more lossy a codec is (i.e. how efficient it is) the worse the generation loss will become - I think, thats an assumption, in a perfect world you would be right - i guess the only way to findout is a test

Take HQ 1080p or i source
transcode to x264 with same GOP arranegment as HDV MPEG2
transcode to lossless frames
transcode to x264 w/ HDVMPEG2 GOP structure
transcode to lossless
transcode to x264 w/ HDVMPEG2 GOP structure
transcode to lossless
transcode to x264 w/ HDVMPEG2 GOP structure

then

Take HQ 1080p or i source
transcode to MPEG2-HDV
transcode to x264 lossless
transcode to MPEG2-HDV
transcode to x264 lossless
transcode to MPEG2-HDV
transcode to x264 lossless
transcode to MPEG2-HDV

then compare results....anyone up for it?


maybe the extra bitrate helps with gen. loss so itd have to be x264 w/ MPEG2 HDV GOP structure @ CBR 25mbps

[Pensive]

Thought so.......

My only conclusion from this thread is -

MPEG4 has the potential to be awesome, and for encoding/playback it is - but may not be the right solution for an editing system , as it is fundamentally flawed by being too lossy, furthermore, too EFFICENTLY lossy, to process more than a couple of times. I cannot be bothered to investigate further, likke i said at the start of this thread - horses for courses. MPEG4 is the way forward, but lossless HD editing will be out of the playing field for most for a very long time. Until then MPEG2 will likely reign.

This may change in the future, but that is how things are right now.