Sony make a very reasonably priced DVD recorder that lets the consummer pop their MS into it and make BD playable disk on a conventional DVD. That's all they want. They're very happy, they haven't had this kind of convenience since VHS-C.

Bob.

The fact of the matter is that H.264, MPEG4, AVC whatever you want to call it, is a newer, more bit-efficient method of compressing video with distinct technical advantages over MPEG2. It does require much more horsepower to decode and encode than MPEG2. With identical, pristine HD source material, at encoding bitrates higher than, let's say, 25Mbps, there is going to be very little difference in image quality. It is only at lower bitrates and particularly broadband Internet bitrates that h.264 really shines. H.264 is certainly the future, but it's not quite there.

To answer the original question and what you're saying.
Professionals use whetever they have to to get the shot whilst taking into consideration budget and deliverables. They test, test, test. This is not trivial. The HDV from the Z1 looks GREAT. It's a real mess when printed to 35mm due to how the photochemical printing process lifts blacks. If you're not delivering 35mm prints, this doesn't matter.

Spot/DSE jumps out of planes with a AVCHD camera on his head.
One of his mates jumps out of planes with a 35mm camera on his head.
Both are professionals. Neither format is easy to post. You factor that in to your budget and workflow.

Bob.

Even today the highest quality mpeg codec is mpeg-1.
All development after mpeg-1 was done to fit an acceptable image into lower bitrates.

Bob.

"The fact of the matter is that H.264, MPEG4, AVC whatever you want to call it, is a newer, more bit-efficient method of compressing video with distince technical advantage over MPEG2. It does require much more horsepower to decode and encode than MPEG2. With identical, pristine HD source material, at encoding bitrates higher than, let's say, 25Mbps, there is going to be very little difference in image quality. It is only at lower bitrates and particularly broadband Internet bitrates that h.264 really shines. H.264 is certainly the future, but it's not quite there."

I would say that sums up my point of view perfectly

In the Spot v. Blink skirmish of words, I'm declaring Spot winner if only for using the word, "measurbate." Twice! Best word I've heard in a long time.

"measurbate." Had to due a double take! ;-)

A lot of people say that somehow, some day, there will be improvements in AVCHD compression that will make it better.

I don't believe this.

Years ago (1992) I dealt a lot with Compression Labs, a company that licensed various video compression technologies for videoconferencing and other similar applications. I actually hired several of their engineers. What I learned is that it is extremely difficult to discover improvements in compression quality over time. In fact, there is actually no guarantee that any improvement is possible because improvement requires that the initial implementations be terribly flawed. However, since most of these encoding standards have already been in the works for years, and since we're starting the third decade of inter-frame coding (where each frame depends on its predecessors), it is very unlikely we're going to see changes that are large enough to change the rules of the game.

Small improvements, yes. Big improvements, no.

As for HDV being a tape format, the fact that the camera Spot describes has an extra unit embedded which records directly to disk means that "being a tape format" only means that it was originally developed for tape, not that it is in any way dependent on the tape transport or the specific characteristics of a linear format. The existence of that unit means that an HDV camcorder can be built without any tape whatsoever, and without having to make it into an expensive kludge.

So I believe that the lack of HDV solid state or HD recording medium is entirely to protect the AVCHD format. Why is this so? I don't know. I do know that any manufacturer who builds and markets an HDV solid state or HD pro camcorder will get a lot of market share from people like those represented here in this forum.

You could probably turn the problem around to see why AVC is used in solid state cameras while HDV is not.

AVC in consumer cameras typically uses a lower bitrate than HDV. I'd bet that when it was in development Sony and Panasonic were looking for a format that'd be sure to work with the memory cards they expected to have available for consumers. They needed something with a lower bitrate than HDV was using.

As far as improvements that make AVC more viable go...I'd expect it more in NLEs than in the AVC files.

(I also want to point out that panasonic went to a 50Mbps AVC format for their tapeless DVCProHD cameras. The previous tape based format was 100Mbps but not AVC. I don't think you'll ever see Vegas support this but I'm sure FCP does)

Rob Mack

It is unlikely that there will be any significant improvements in AVCHD compression, because if there were, it wouldn't be AVCHD anymore. The AVCHD format specs are etched in stone, H.264 by itself is not. H.264 gains efficiency over MPEG2 in a number of ways which are discussed here:

http://en.wikipedia.org/wiki/H.264

I guess that technically HDV is actually a tape format. Once it gets recorded to another medium, it ceases being HDV and becomes an MPEG2 Transport Stream with HDV specific encoding parameters. By definition, HDV, whether recorded to tape, flash memory card or hard drive, is limited to MPEG2 at a resolution of 1440x1080 with color sampling of 4:2:0. Where AVCHD cameras have an advantage is that they can encode at resolutions of 1920x1080 with an extended color range, which can look better assuming that the camera and lens is up to the task. H.264 itself is currently capable of 4096x2304 4:4:4 video at up to 960Mbits/sec.

Of course, the Sony EX1 and EX3 camcorders record 1920x1080 using VBR MPEG2 with an average bitrate of 35Mbps. They look absolutely spectacular, but a lot of that is a function of the quality of the camera's imaging chips. You really have the best of both worlds; 1920x1080 and easy-to-edit, high-bitrate VBR MPEG2 encoding.

I admit I'm still the new guy around here, so I won't stick my neck out and CLAIM anything, but my eyes don't fool me. I'm sure many of you have seen this....

http://www.usa.canon.com/dlc/controller?act=GetArticleAct&articleID=2326

I understand the pitfall of the CMOS sensor in its current state with DSLRs and the jelly/wobble effect, but I don't think anyone here can argue with the amazing video that the 5D MKII can produce. H.264 is amazing if implemented properly. As John Cline stated, Mpeg 2 is clearly limited compared H.264. That's not to say that Mpeg 2 sucks (again the Sony EX's prove that) but I would put my money on H.264 leading the way to bigger and better things. Panasonic's AVCCAM AG-HMC150 is supposedly producing brilliant images using AVCHD @ 21Mbps with 3 CCD's. It may be my next purchase.

Professionally....AVCHD? Ehhh, maybe. H.264? Absolutely.

Maybe we need to redefine what "professionally" means here, because the dictionary says its either A- getting paid for the same practice that one once did as an amateur, or B- making a living off of said practice.

I've already made money in my short side career using AVCHD. Does that count?

Of course, the Sony EX1 and EX3 camcorders record 1920x1080 using VBR MPEG2 with an average bitrate of 35Mbps. They look absolutely spectacular, but a lot of that is a function of the quality of the camera's imaging chips.
=============
A point I have been trying to make for years. This is where the consumer cams fall short and they have gotten worse in this respect over the past several years instead of better. The industry substituted useless bells and whistle features in place of larger and better imagers.

John

"I understand the pitfall of the CMOS sensor in its current state with DSLRs and the jelly/wobble effect, but I don't think anyone here can argue with the amazing video that the 5D MKII can produce. H.264 is amazing if implemented properly. As John Cline stated, Mpeg 2 is clearly limited compared H.264. That's not to say that Mpeg 2 sucks (again the Sony EX's prove that) but I would put my money on H.264 leading the way to bigger and better things. Panasonic's AVCCAM AG-HMC150 is supposedly producing brilliant images using AVCHD @ 21Mbps with 3 CCD's. It may be my next purchase."

Well to be fair, the video being shot there probably has more to do with the lens than the format/codec being used. I'm not sure that mpeg2 couldn't do the same job here. But I don't think anybody here (including myself) can't or does not see the benefits that H.264 can offer over mpeg. In fact the problem with H.264 really has nothing to do with the H.264 itself. It's the proper tools to adequately work with it that becomes the problem..... they still aren't there yet. Now that's not to say that you CAN'T work with it. At least it's now to the point where you can muddle through a H.264 time line if you have to..... which is better than a year and a half ago where you couldn't even find a time line to accept it. But to work with it on any broad scale ..... sorry.... just not there yet.

John:

Your whole post is just not accurate. I wish people would be more objective and stick with what they have experience with.

I shot HDV since 2003. I shot AVCHD (6 edited AVCHD events so far) since October 2008. AVCHD codec is the clear winner.
___________________________________________________
Well, as everyone has said, HDV is easier to edit, by far. No one has said AVCHD footage looks better. In fact, it seems that the consensus in the posts above is that, if everything goes well, it can look as good.

Your whole post is just not accurate. Really???

From my post: "AVCHD is slow to edit"

So you find that AVCHD is faster to edit than HDV??


From my post: "performance improvements in basic computing power have maxed out for the last six years, and the only improvements we're seeing is by providing parallel threads, processes, and CPUs"

So, do you have a 10 GHz computer?


From my post: "difference in compression between HDV and AVCHD is, at most, two to one"

Am I wrong here? Does AVCHD have a bigger advantage? Yes, the implementation does allow for high compression rates if you don't care about quality and want to store a lot of video in a small space, but the exact same thing can be done with MPEG-2, as we all know, so HDV could easily be adapted to solid state storage and, once it is not tied to the 13 GB/hour hard-wired encoding rate used to achieve compatibility with the various DV tape formats, you could apply any compression rate you want.

I can go on, but my post was actually entirely accurate, and the one thing that I think really got under your skin -- my assertion that the quality wasn't as good as what I have achieved with HDV -- is echoed by many of the other people who have posted in this thread and, if you look again at the subject line of the initial post, is the basic premise/question of the original poster.

Well, as everyone has said, HDV is easier to edit, by far.

Im not sure who everyone is, but Im not one of them. They edit exactly the same, but AVCHD renders slower, but not enough to be a consideration on a Quad core machine.

No one has said AVCHD footage looks better. In fact, it seems that the consensus in the posts above is that, if everything goes well, it can look as good.

Not sure where this is coming from either. Here is a Panasonic marketing example of the codec differences.

http://www.panasonic.com/business/provideo/avccam-features.asp

I have only one experience -- a full-length ballet shot with an FX1 (HDV) and a Sony SR12 for the wide-shot (AVCHD). I did post that I found the SR12 footage lacking in basic sharpness and detail, something I was able to overcome enough to not call attention to itself by using Sharpen set to 0.00. I shot on the SR12 using the top quality 1440x1080 mode. Perhaps I should have used the 1920x1080 setting, and by not doing so, put the camera at a disadvantage.

Wow, what a lot of experience. And with two consumer level cameras.

So, in my case, the quality difference may have had to do with the SR12 itself, or the setting I used (although it was the same setting as what the FX1 produces). AVCHD may not have been the issue.

Do you even know there is a difference between the codec and the image quality recorded on it? You call this a comparison?

The render of this project took longer than any other project I have ever created, and by a LOT. Those that think faster computers are around the corner are smoking stuff: performance improvements in basic computing power have maxed out for the last six years, and the only improvements we're seeing is by providing parallel threads, processes, and CPUs. This is great for rendering, but won't help much during playback. The basic clock speed has been at 3 GHz for over six years now (I just bought a new computer and the maximum clock speed available to me is identical to what it was years ago).

Thanks for the in-depth technical discussion and explanation of render times.

So, since AVCHD is slow to edit, it is going to STAY slow to edit for a LONG time. No miracle is around the corner.

Again, thanks for the in-depth technical discussion of actual render times.

Why Sony and others don't provide solid state recording for HDV is a complete mystery to me. The difference in compression between HDV and AVCHD is, at most, two to one, and unlike CPU speeds, the change in the cost and size of solid state memory is changing dramatically. Thus, it won't be long before 32 GB of solid state storage is only $20 or so. That's enough for two hours of HDV. At that point, I really don't see the advantage of AVCHD, other than as a delivery format for BD.

You obviously have done no research at all.

I just purchased what is just about the fastest single-CPU stock computer one can get (3.2 GHz i7), and I still expect that AVCHD editing will be uncomfortable and slow, although I expect the render times will come down dramatically because that can be done on parallel threads and cores.

So your dramatically faster render times are uncomfortable and slow? Give us some examples with numbers.

blink3times: Please show me a NLE that can do full recompress renders at more than 24Mbps (most won't even render above 18).... one that we can afford anyway. I can use vegas to get to 40 with mpeg2

Please excuse my ignorance here, but I am not quite sure what you mean. Are you saying that if I import AVCHD footage (which is a capture format) that Vegas and other NLEs limit the output bitrate that I can re-encode to? Seriously? That sounds peculiar to me, but since I do not have any AVCHD footage readily available to me to re-encode, I can not test it.

AVCHD is just H.264 with a specific frame rates and image sizes, so I don't see why it can not be re-compressed like any other footage, using a standard H.264 encoder to any bitrate you might want. Is there, for example, a limitation in the MainConcept H.264 encoder that will prevent you from re-encoding AVCHD to 50Mb/s H.264? That sounds unlikely to me since there is no such limitation when encoding other footage. I am encoding to 50Mb/s AVC at this moment with no problem. Not that Vegas 8.0 can do it, but 8.0 is a buggy piece of junk when it comes to H.264, 8.1 works no problem though.

Just a little confused here. Would love if you could help me understand what you mean.

H.264 contains a lot of improvements over MPEG2, as such, one can fully expect it to produce better images with fewer artifacts when compared to MPEG2 at the SAME BITRATE. This is an undeniable fact.

The encoding methods and algorithms for MPEG2 have not changed since the specs were ratified some years back. What HAS changed is on the decoding side, many improvements have been made to MPEG2 decoding software in the form of speed, de-blocking and mosquito noise reduction. Encoding is easy, decoding is not. MPEG2 decoding has been improving all the time and it has gotten very, very good.

I am certain that the same thing will happen with H.264, the encoding specs are etched in stone, but there will be improvements in the decoders for years to come. It will only get better and faster.

My primary beef with H.264 has nothing to do with the encoding specs, they are some amazingly brilliant pieces of advanced mathematics. Most implementations of H.264 have been to reduce bitrate while maintaining quality. I want it to work the other way around, improved image quality at the same bitrates. MPEG2 looks pretty good at 25Mbps, H.264 at 25Mbps can and does look even better assuming that the source material is high quality to begin with.

The ONLY way to compare the relative merits of H.264 and MPEG2 is by using identical source material. Comparing the images of, lets say, a Z1 HDV camcorder with those produced by a single-chip AVCHD camcorder has almost everything to do with the cameras themselves and not the underlying compression technology and any conclusions are simply not at all valid.

"Please excuse my ignorance here, but I am not quite sure what you mean. Are you saying that if I import AVCHD footage (which is a capture format) that Vegas and other NLEs limit the output bitrate that I can re-encode to? Seriously? That sounds peculiar to me, but since I do not have any AVCHD footage readily available to me to re-encode, I can not test it."

In a manner of speaking... yes. It of course has nothing to do with a limitation on H.264 itself. It can render out to much higher bitrates... but I'm sure you know that. It seems however to be a limitation of the present day nle's. There isn't one (at the consumer/prosumer level anyway) that will render avchd much above 18M. There are a few burner programs that will PASS THROUGH higher bit rates, but rendering.... no.

So if you get one of those flashy new Canon cams that record at 24M and do any editing (something that requires rendering) your max output will be around 18M.

Vegas is one of the few editors that actually SHOWS 20M as an option.... but I have yet to see a render come out of it at that rate (crashes). The ulead products.... if you try and set anything higher than 18M I think it was... it reverts back, and even with a 18M setting the PS3 shows nothing peaking at more than 17 (hovers more around 15)

I should add that the only nle/burner program that I don't know anything about is Encore, but I have heard similar complaints with that one too (third hand info so it could well be wrong though)

This all BTW brings up a question....
I know that there are 2 main types of avchd main profile used mostly by Sony and high profile used by Canon and Panasonic. The main profile has a bitrate limit of some kind on the order of 17 or 18Mbps. Just guessing but this suggests that the present day nle's are outputting main profile avchd regardless of which type is being inputted. Does that make any sense.... or am I off on the wrong track?

I'm not sure if it was Chris Hurd or myself that created this word...I can remember when I first tongue-slipped it, and then heard Chris use it, but he said he'd used it before he met me...so it's a Hurd-Spot word. it's in the Urban Dictionary...has been for a while now.
Fun word, for people that prefer self gratification via the nasty numbers vs actually doing the real thing....
;-)
A lot of that happens here.

Your analysis Jeff leaves me speechless.

blink3timesM: So if you get one of those flashy new Canon cams that record at 24M and do any editing (something that requires rendering) your max output will be around 18M.

OK, now I am utterly confused. I just downloaded a few of the clips that are out there from the new Canon 5D MKII and I dropped them on my time line. I didn't check the bitrate before doing so, but rendering them to 50Mb/s posed no problem whatsoever. I played it on the PS3 and it confirms the clip is 50Mb/s. What am I missing here?