Just bought the Sony HC3

[Jerry Jones]

I've used the Sony 3 chip HDV cameras and personally think they are significant leap from DV camcorders.

- Mark

Have you compared the video from a 2/3" CCD DVCam camcorder such as this...

http://tinyurl.com/cgl8b

I realize the Sony camcorder referenced costs $18,000 and the Sony HDV camcorders only cost around $3,000.

But when I saw the difference... on a widescreen HD monitor... I was surprised by how good the DVCam picture looked in comparison to the HDV picture and it gave me a better appreciation for camera optics as opposed to raw video resolution.

For me, the H.264 doesn't necessarily mean *more* compression.

It means *more efficient* compression.

In other words, given *EQUAL* data rates, the H.264 will compress video more *EFFICIENTLY* than MPEG-2 and produce a better picture (all other factors being equal).

Moreover, it does seem that the HDV group is "tied to tape" as explained on the following Web page of the HDV licensing entity:

"HDV and HDV logo will be licensed only to the TAPE DRIVE products which can record and play video cassette compliant to the HDV Format (HDV Tapes)."

http://tinyurl.com/o2xcu

I guess I'm a little puzzled when I read your comment that H.264 is "more difficult to edit than MPEG-2."

I guess my question - then - would be... HOW IS THAT POSSIBLE?

Especially when you consider the fact that so many NLEs edit MPEG-2 by first transcoding to intermediate codecs.

If that turns out to be the case for H.264, then how is that "more difficult?"

It would seem to me that it would be "the same."

When it comes to native editing of H.264, then I would - again - pose the question.

How is it "more difficult?"

Is it really "more difficult" for a software engineer to write a program that can edit H.264?

Or is it simply "more difficult to edit H.264" - at present - because we don't yet have a native H.264 editor on the market?

Jerry Jones

http://www.jonesgroup.net

[Jerry Jones]

By the way, I realize many people are satisified with HDV for the reasons you stated.

I appreciate that and I offer no criticism of individuals who make that choice.

Of course, it boils down to individual choice.

But for me the HDV choice was really a premature decision.

To be honest, I'm not wealthy and can't splurge on new high definition plasma or LCD screens.

On the other hand, I'm in the same boat as a lot of people.

The software vendors and the camera manufacturers put a lot of pressure on those of us who enjoy digital video to rush into the decision to buy HDV camcorders when it didn't make any sense at all... at least not for some of us.

No doubt your three-chip Sony produced a better picture than my JVC JY-HD10, which was a single-chip HDV camcorder.

That said, I was pretty bored with the higher resolution images produced by my HDV camcorder after just a few days.

MiniDV - for me - is more *fun* than HDV.

Why?

For all of the reasons stated at the beginning, but I think the really big reason is that it doesn't CHOKE MY COMPUTER PROCESSOR to handle MiniDV clips.

I can edit them so EASILY without the hassle of waiting for my computer to "catch up" because it's not fast enough to handle HDV clips.

I don't have the fastest computer on the market, but "for crying out loud" it's a 2.4ghz 64-bit AMD Athlon chip... and it's not exactly "slow."

In addition, I don't have to down-convert my finished video to be able to share it with others -- all of whom don't have the ability to view high definition video on DVD.

No... for me... the HDV thing was really a bust.

Jerry Jones

http://www.jonesgroup.net

[Hulk]

Jerry,

As you recognize in your post, it's not really fair to compare an $18,000 camera to a $3000 one. And I agree with you that there is more to picture quality than counting pixels. Optics of course are extremely important, as is lighting. A great example of this is one of my favorite Sci-Fi series "Farscape." Check out one of those DVD's, the quality is amazing.

I haven't made a direct comparision of a large format SD cam with a three chip Sony unit using the same subject material. But I wonder how this would go if the scene was professionally lit and if the monitor was large enough, say 55" or more so that the increased resolution of the HDV format could be seen? It is my experience that superior optics allow for more "wiggle room" as far as working with non-optimal lighting than lower quality optics that's why I think the scene would have to be professionally lit to give the HDV camera a chance to show it's stuff.

The increased color space of the more expensive camera makes it's output look to be higher resolution than the numbers would indicate too.

Still, at anywhere near equal price points, when shot by people who know what they are doing, and of course viewed on good HD monitors I think HDV cams will deliver a more exciting presentation.

As for MPEG-2 vs. H.264, you are right to a point. At lower bitrates H.264 will look better than MPEG-2 at the same bitrate. But, and this is a big but, H.264 was NOT designed for high bitrates and MPEG-2 is (was). I don't remember all of the technical details, but this was the case. Things may be changing and H.264 might start being used for higher bitrates but I would be surprised if this happened as H.264 was designed from the start to excel at low bitrates.

The H.264 compression is more advanced and therefore more efficient than MPEG-2, that's why it does better than MPEG-2 at low bitrates. I believe there is an extra step in the compression chain, along with other enhancements. Again, I apoligize for not having this information at my fingertips. But the point is that the additional overheard will make editing more difficult (processor intensive) when editing natively.

In addition, since H.264 requires so much processing power for good real time recording the camera will have to have a seriously fast and efficient processor to compress to this format on the fly.

I would be surprised if H.264 cams recorded at bitrates equal to HDV cams.

If they did, and they could provide better video than MPEG-2 then using an intermediary would be a possibility.

But as I stated above I don't think H.264 cams will work out as a higher than HDV quality format for the following reasons:

- H.264 was not designed as a high bitrate (low compression) solution
- H.264 is more processor intensive to encode and quality realtime encoding could be problematic in mobile camera applications (of course faster processors and better routines could help here)
- Native editing of H.264 is more processor intensive than MPEG-2 editing

I have a feeling manufacturers are looking to H.264 HD cameras so that they can compress the video to LOWER bitrates than the standard HDV spec as approximately equal quality so that Flash type media are a viable option for these cameras.

If my HD camera must use temporal (interframe) compression I think I'd rather go with MPEG-2 over H.264.

Time will tell.

- Mark

[Jerry Jones]

Yes - the HDV might allow for a "more exciting" presentation... IF... you will be making the presentation to an audience that has access to a high definition DVD player and a high definition screen.

When I was a TV journalist putting together video presentations for the local news, the vast majority of people watching in the 80s and 90s didn't have access (and most still don't have access) to high definition screens.

So - for ENG - high definition isn't "there" yet.

When I put together video presentations for the Boise mayor and council (90s & 2000s), they didn't have access to high definition screens (and still don't).

So - for local government here in Idaho - high definition isn't "there" yet.

I now work as a Video Instruction Manager at the Boise campus of Idaho State University where we deliver graduate-level courses using a statewide network of videoconferencing classrooms equipped with Tandberg codecs. We're not at the point where HD can be utilized for videoconferencing.

So - for university-level education here in Idaho - high definition isn't "there" yet.

I would be more excited about HDV if there were far more ways of sharing it, but the reality is that HD still faces practical difficulties in this part of the country.

I'll do some reading up about your claim that H.264 wasn't designed for high bitrates.

I'm not so sure about that one.

But your suggestion is quite interesting.

Later,

Jerry Jones

http://www.jonesgroup.net

[Jerry Jones]

By the way, there's a great Panasonic professional camcorder Web page here:

http://tinyurl.com/9zre8

What stands out is the price list for the various camcorders.

If you look closely at that list, you begin to see that the HUGE difference is the CCD size.

As soon as you put a 2/3" CCD into a camcorder - even if it's an SD camcorder - the price jumps to about $18,000.

In other words, there are no consumer or "prosumer" camcorders - to my knowledge - that feature 2/3" CCDs.

At the risk of repeating myself, I've seen the SD material from these 2/3" CCD camcorders displayed (zoomed to fill the screen) on HD monitors and it is really quite surprising to see how well it scales up to compete with HDV material shot using a 1/3" CCD camcorder.

Give it a test, yourself, and you might be as surprised as I was.

Jerry Jones

http://www.jonesgroup.net

[Hulk]

Jerry,

I don't doubt how good really high quality SD can look with great cameras. A good camera can really make each picture count. I can take a picture at the lowest resolution with my Canon Digital Rebel and Tamron 28-75mm XRDi lens set at 3MP and it will look better than a 4MP point and shoot.

A note about H.264 being for low bitrate. Last year I did a two part series on frameserving from Sony Vegas to Nero Recode to obtain MPEG-4 output. In part two I did some informal tests of MPEG-2 vs. MPEG-4 vs. WMV encoding quality at various bitrates. One of the guys in the Sony Vegas forum that does a lot of day-to-day work compressing video told me that I was comparing apples to oranges when comparing MPEG-2 to MPEG-4. One is meant for high bitrate and the other for low bitrate. I forget who it was and can't find the e-mail exchanges. I do remember he was a very reputable source.

It was my finding that at low bitrates MPEG-4 was superior to MPEG-2, but as you can see from the link below the MPEG-2 video did have a "contrastier" look, closer to the original in that respect, just suffering from more macroblocking and other artifacts.

I also remember fooling around with high bitrates and I was not able to get the MPEG-4 encoded video to look as good as the MPEG-2 encoded video. Of course this was an informal test but I do remember spending quite a bit of time on this article.

- Mark

[Jerry Jones]

H.264 was NOT designed for high bitrates and MPEG-2 is (was). I don't remember all of the technical details, but this was the case.

As I said earlier, your suggestion above was interesting.

However, I found the following PDF scientific analysis of H.264:

http://tinyurl.com/fxpur

This particular scientific paper seems to contradict your suggestion, Mark.

"Unlike the MPEG-2 video coding standard, AVC/H.264 targets a wider range of video applications, ranging from video at mobile devices and bit rates as low as below 30Kbit/s to HDTV and bit rates of 20Mbit/s and above."

This paper also explains multiple tests, which concluded with the following statement:

"The tests showed a noticeable superiority of AVC/H.264 compared to state of the art MPEG-2 video encoders in almost all test cases."

"The AVC/H.264 encoders, which were at a very early stage of optimization, competed with MPEG-2 encoders that have been optimized for nearly a decade."

"Even at a very early stage of optimization AVC/H.264 has shown a remarkable superiority in coding efficiency compared to current MPEG-2 encoding technology."

But I think you're right about the need for even faster processors to edit H.264.

That's why I'm waiting.

I think APPLE will be the company that will offer FLAWLESS H.264 video editing because APPLE - unlike Ulead and other Windows-dependant software companies - can control the hardware.

So when APPLE makes the claim that their editor will edit AVCHD, then they can't use the excuse that "it was the computer" or "it was the processor" or it was "other software installed on the computer" or whatever.

The problem facing Windows software companies is they can't control the hardware.

This is a HUGE DISADVANTAGE, in my mind.

Jerry Jones

http://www.jonesgroup.net

[Hulk]

Jerry,

As I said the person who actually works with these encoders day-to-day told me the MPEG-2 bitrates could be in the neighborhood of 40 to 50MB/s, which is significantly higher than the 20MB/s mentioned in the article, and I wonder if "and above" means that far above. He is a highly regarded poster over in the Sony forums so I believed him. He could be wrong, but as yet I haven't seen any H.264 video with really high bitrates. We'll see what happens.

I have no doubt we'll be able to edit H.264 eventually. Probably sooner rather than later. In 1998 I was told we'd never edit MPEG-2 but a year later we were. The problem is MPEG-2 is still not as responsive as editing avi (non-temporally compressed video). I would just rather edit something that is less compressed rather than more compressed, especially if we're going to have hard drive based cameras where storage space and efficiency of encoding isn't as much of an issue.

I'm waiting on HD too. I'm waiting for the right camera, cost efficient distribution, and better HD playback devices. Nothing in my living room is HD at this time. As you said really good SD still looks pretty good.

- Mark

[Jerry Jones]

As I said the person who actually works with these encoders day-to-day told me the MPEG-2 bitrates could be in the neighborhood of 40 to 50MB/s, which is significantly higher than the 20MB/s mentioned in the article, and I wonder if "and above" means that far above.

Well, Mark, the math still doesn't seem to "add up" with that claim.

I guess my question would be:

How can a codec, which is proven to be more efficient at lower bit rates, suddenly and mysteriously lose its advantage at higher bit rates?

I've searched through many Web documents on this topic and I keep finding more and more documents that seem to contradict the theory that H.264 "isn't as good as MPEG-2" at high bit rates.

What I seem to be finding - over and over - is the claim that, to the contrary, H.264 is simply the "likely successor" to MPEG-2 for literally all uses... including high definition broadcasting.

I suspect the reason you see references to "20MB/s and above" for H.264 is due to the previous observations that the H.264 codec can provide the same quality at roughly HALF the bit rate of MPEG-2.

But I'll keep my eyes open and if I see a document that states H.264 suddenly loses the battle against MPEG-2 at the 40MB/s threshold, I'll post it here and I'll "eat my hat."

Jerry Jones

http://www.jonesgroup.net

[Hulk]

Jerry,

Okay. I am going to try and track down the person that told me about this. I don't find it that strange actually that a codec could be optimized over a certain range of bitrates.

Mark

[Jerry Jones]

This PDF document seems to confirm my own belief that there are, indeed, very high quality profiles established for H.264, which would seem to contradict the theory that it was designed only for relatively low bitrates:

http://tinyurl.com/mzo89

The report concludes:

"The new video standard known as H.264/AVC presents a rich collection of state-of-the-art video coding capabilities that can provide interoperable video broadcast or communication with degrees of capability that far surpass those of prior standards."

"With the new FRExt amendment, and especially the new High Profile, H.264/AVC further bolsters its position as the premier design for standardized video compression."

"We believe these technologies provide a hitherto unavailable set of cost/performance points that will have a powerful impact on both consumer and professional video applications in the years to come."

The FRExt amendment defines four new profiles:

♦ High (HP)
♦ High 10 (Hi10P)
♦ High 4:2:2 (Hi422P)
♦ High 4:4:4 (Hi444P)

All four of these profiles build further upon the design of the prior Main profile, and they all include three enhancements of coding efficiency performance:

♦ Adaptive macroblock-level switching between 8x8 and 4x4 transform block size
♦ Encoder-specified perceptual-based quantization scaling matrices
♦ Encoder-specified separate control of the quantization parameter for each chroma component

All of these profiles also support monochrome coded video sequences, in addition to typical 4:2:0 video.

The difference in capability among these profiles is primarily in terms of supported sample bit depths and chroma formats.

However, the High 4:4:4 profile additionally supports the residual color transform and predictive lossless coding features not found in any other profiles.

Jerry Jones

http://www.jonesgroup.net

[Jerry Jones]

Your acquaintance may not have been aware that the H.264 standard completed in 2003 was followed by *extensions* designed to support high-end professional tasks and high bit rate uses of H.264.

According to the PDF document I cited in the previous post:

"While having a broad range of applications, the initial H.264/AVC standard (as it was completed in May of 2003), was primarily focused on 'entertainment-quality' video, based on 8-bits/sample, and 4:2:0 chroma sampling."

"Given its time constraints, it did not include support for use in the most demanding professional environments, and the design had not been focused on the highest video resolutions."

"For applications such as content-contribution, content-distribution, and studio editing and post-processing, it may be necessary to:

♦ Use more than 8 bits per sample of source video accuracy
♦ Use higher resolution for color representation than what is typical in consumer applications (i.e., to use
4:2:2 or 4:4:4 sampling as opposed to 4:2:0 chroma sampling format)
♦ Perform source editing functions such as alpha blending (a process for blending of multiple video scenes,
best known for use in weather reporting where it is used to super-impose video of a newscaster over video
of a map or weather-radar scene)
♦ Use very high bit rates
♦ Use very high resolution
♦ Achieve very high fidelity – even representing some parts of the video losslessly
♦ Avoid color-space transformation rounding error
♦ Use RGB color representation."

"To address the needs of these most-demanding applications, a continuation of the joint project was launched to add new extensions to the capabilities of the original standard."

"This effort took about one year to complete – starting with a first draft in May of 2003, the final design decisions were completed in July of 2004, and the editing period will be completed in August or September of 2004."

"These extensions, originally known as the 'professional' extensions, were eventually renamed as the 'fidelity range extensions' (FRExt) to better indicate the spirit of the extensions."

Jerry Jones

http://www.jonesgroup.net

[Jerry Jones]

Another PDF paper that shows a comparison between "fidelity range extension" (FRExt) H.264 and JPEG 2000:

http://tinyurl.com/enmou

Jerry Jones

http://www.jonesgroup.net

[Jerry Jones]

"Average bit-rate savings for the H.264/MPEG4-AVC High profile of 10% relative to MP (both using CABAC) and nearly 60% relative to MPEG-2 have been observed for a set of HD sequences."

"When restricted to intraonly coding, H.264/MPEG4-AVC HP outperforms the state-of-the-art in still-image coding represented by JPEG2000 on a set of monochrome test images by 0.5 dB average PSNR."

(In the CONCLUSION of this PDF document.)

http://tinyurl.com/enmou

Jerry Jones

http://www.jonesgroup.net

[Jerry Jones]

And here:

"The Fidelity Range Extensions demonstrates even further coding efficiency against MPEG-2, potentially by as much as 3:1 for some key applications."

http://tinyurl.com/zh9lg

(third paragraph)

Jerry Jones

http://www.jonesgroup.net