Why not include this feature??

[Juan P. Pertierra]

I've been reading the forums here for several weeks now, and it seems that most prosumer/pro camcorders are flowing towards the 16:9, 24p/30p/60i (and maybe HD)formula more or less...but even though everyone has been very vocal about asking for these features in future camcorders, there is something I haven't seen mentioned...

So far, most of us have been stuck with miniDV, 5:1 compression 4:1:1 sampling...but all of these cameras have firewire capability...is it really so difficult to have the camera sample at, say 4:2:2 and have an option to output a uncompressed(or maybe 2:1) stream on the firewire port(s)? To record to tape, all you have to do is throw away a couple samples and compress right?

It seems like most of the CCD's used in these machines can output much more data than is recorded to the miniDV tape, and the same is true for the firewire interface, so I can't see any significant additional cost related to adding this feature...hell, i'd pay at least $500 extra for any camcorder that could output 4:2:2 uncompressed, even if it is still SD. When on the move, use the tape...but any other time i could hook it up to a harddrive and get the extra quality that is ~already~ there thanks to the CCD's and the bandwith of firewire.

Any comments?

[Mike Rehmus]

The problem is that you are no longer miniDV compatible and the readily available and inexpensive DV computer systems will no longer be adequate. In other words, what you want raises the price tag and greatly narrows the choice of available hardware that can handle the data rates.

I know, pure DV users could just ignore the options but the options cost money in terms of camera electronics, a new 'standard' for the video, etc.

I think you will see an upgrade of DV25 to DV50 in the not too distant future. JVC and Panasonic already have that capability in DVCPro50 and D-9. Sony is the only one that has not joined the party. But it will be a more expensive party and perhaps one that will not occur at the prosumer level for a few years.

Truth is we greatly benefit from the use of DV by the consumer marketplace. Consumers won't see the need for more capability until they get HD. And then 4:1:1 HD will probably be good enough.

Interesting issue.

[Juan P. Pertierra]

Thanks for the reply!

I kinda see your point...but I guess I might be underestimating the difficulty in making a system that can capture 4:2:2 using standard 4:1:1 DV hardware...sounds like plain clocking and software differences to me. I'm an EE so I might be looking at this in too simple a way, but where would the extra cost come in, as far as taking some extra color samples when sampling the images?

The camera can still record standard 4:1:1 DV to the cassette and even output the same standard DV stream via firewire by default, but maybe an option that would allow outputing a (preferrably prexisting standard)4:2:2 format via firewire somehow. The compatibility would still be mantained with the standard recording to the casette and default 4:1:1 firewire settings, but it would allow to better use the CCD hardware that is already there. I also don't see any complications in down-converting the 4:2:2 to 4:1:1, since afaik is only a matter of throwing away the extra color samples that where captured, and compressing to DV.

I've always felt that the CCD's in miniDV cameras are heavily under-exploited due to the limitations of the casette format, but the relatively high-bandwith interface like firewire would allow more efficient usage of the hardware, even if you can't record it to casette, and even if most consumers would use the FW in the standard 4:1:1 standard. One thing for sure, I bet that every other person out there(prosumers, professionals) will be using the high quality output, even if it means hawling a laptop or harddrive around. I know I would.

Indeed, most software probably would not support this theoretical 4:2:2 stream(specially if it turns out to be propietary), but within a month or so the support would be there. Look at the DVX100...all the large program manufacturers added support for it rather quickly.

I'm probably missing something, but in either case, there's always the reason that companies don't want to make their low/mid end products rival their high-end costy products even if it ~were~ cost efficient. For example, if the DVX100 did what i'm talking about here, you could just hookup a harddrive to it, and it might perform close to the 25,000 SDX900, minus the differences in CCD size, etc.

Cheers! :)
Juan

[Jeff Donald]

The manufactures are not going to give away those features for a mere $500. Take a look at the SD models that have 4:2:2 sampling and lower compression. I can't find any for under $5,000. I think what you'll find is Digi Beta models costing $10,000 plus dollars. I think everybody here would gladly pay an extra $500 for that kind of quality. But it isn't going to happen real soon.

[Mike Rehmus]

My DV editing system converts the video to 4:2:2 as soon as it gets on the timeline. And doesn't down-convert it until it goes out the firewire port or into an AVI file. But the hardware/software didn't cost $500. It costs, in its current form about $1200.

The dual mode cameras have to have dual embedded firmware, perhaps a faster DSP and buffers (certainly larger) and so on. Remember Sony, for example, has spent a lot of $ driving the power requirements of their prosumer cameras way down.

I don't think it is as easy as you might think. Underused doesn't mean that these electromechanical marvels are brim full of potential that can be tapped. I wouldn't be surprised to find that the CCDs really cannot support 4:2:2 sampling at their present price-point. I've been there with Japanese manufacturers of electromechanical devices (scanners) and they cut every possible corner.

As for hauling more hardware around. I have to haul too much as it is now. I don't think my customers would pay $1 more if I shot their wedding on a 4:2:2 camera.

[Juan P. Pertierra]

Converting video that was recorded as 4:1:1 to 4:2:2 doesn't do anything for quality, and it's not what i'm talking about. This is equivalent to taking a 640x480 image and stretching it to 800x600, the samples just aren't there.

CCD devices by themselves, AFAIK only react to light impinging on them, and don't do any sampling. It is up to the (8-bit for DV) sampling hardware to take a certain number of samples. Only reasons that I can think of why the current CCD's won't 'work' for this is that either they don't react fast enough such that faster sampling is of any benefit(unlikely), or the manufacturer is using a packaged solution in which the 3 CCD's and sampling hardware is hard-wired to 4:1:1. This is possible for low end cameras, but might not be true of the higher end prosumer/pro models.

I think Jeff hit the nail in the head...i think this is possible, but manufacturers choose to couple anything better than standard 5:1 compressed 4:1:1 video with their larger CCD's professional models. And of course, it makes sense...specially since better formats like DVCPRO50 can support it.

This whole thing came to me while reading about the thomson viper camera...it records 4:4:4 HD, but since no tape format out there can handle that data rate, the internal tape only records 4:2:2...but still, you can hookup a hard disk to it and get the full 4:4:4 which is what pretty much everyone seems to be doing.

[Ken Tanaka]

" I don't think my customers would pay $1 more if I shot their wedding on a 4:2:2 camera."
I think Mike captured part of the essence of the underlying demand chain issue. Consumers of low/mid-end cameras and consumers of most video services understandably don't give a damn about such ultra techie matters. Consumers want convenience, reasonable performance and durability at a low cost. Service clients want good productions at reasonable prices. Since approximately 30% of all humans have some type of difficulty discerning color and contrast anyway, selling an "expanded color space" camera would probably not be a viable development/marketing strategy in the lower environs.

It isn't until you reach the hallowed inner cloisters of video technicians that such matters loom large. I would make an argument that even in that "professional" environment it's a technical nit within the context of the overall value of a production (i.e. the story, the acting, the subject, etc.).

And that comes from someone who's considering buying a DV50 camera next year! <g>

[Juan P. Pertierra]

>>it's a technical nit within the context of the overall value of a >>production

If this is true, then why doesn't everyone just use a cheap $200 camcorder to do their feature film? Do ~you~ use an Emerson $100 VHS-C camcorder? If not, does that mean you're just crazy about the technical aspects, which only you will notice and no-one else?

I don't think so. This is a visual medium. Appearance counts, and the more lattitude the medium has, the more you can express yourself. The cripsper the images, the better the color, the better the control over frame rate, the more you can control these parameters and create your own unique view. Film gives you so much control to do all of this(albeit very expensively), that I think this is why every digital camera is flowing towards film-like features. Emulating the flexibility and quality of the medium is the goal.

I have to admit, I am coming from the point of view of digital cinematography, and the fact that most of us can't afford film or a thomson viper, and are trying to get the most out of the cameras that we can afford. But once again, to those who might say that this is just a technical issue, are you using the cheapest camcorder you could buy? There are standards...what i'm talking about here(like I mentioned in the first message) is what could be in the ~next~ standard camcorder...just like 24p was not in yesterday's camcorder but will certainly be in many of the high end models in the future. All it takes is for one manufacturer, like say Canon, to exploit this in their XL2 for example, and all the others will follow. A new standard.

And if you don't care about a better color, and better images with less(or none) compression, then you don't have to change anything, or carry anything extra. But just imagine, for people who do a lot of chroma keying for example, how awesome it would be to reduce or eliminate all the compression artifacts that are in DV....and the few technical weirdos who would take uncompressed 4:2:2 over standard DV anyday... :)

BTW, thanks for all the replies..if you can't tell already, i love a good discussion. :) Now if I could only afford a DVCPRO50....

[Jeff Donald]

Sony, Panasonic, Canon, JVC, etc. are not going to give away the good stuff. If you want it, it's available, but your going to pay dearly for it. That's the market place. Wish all you want, but none of the major manufactures are going to produce cameras that output Digital SD 4:2:2 with 2:1 compression for the same price as DV. Nor are they going to do it for a $500 premium. It's not worth their R & D dollars to give away their high end products. That's the market place.

Your basically preaching to the choir here. Almost every member would want what you're talking about. But there's little any of us can do about convincing the major manufactures to see it our way. The technology will trickle down as new technologies in R & D right now come out. Nothing in this industry is static for long.

[Juan P. Pertierra]

Indeed, that is the bottom line.

Q:Why not include this feature??
A:What jeff said.

Cheers!
Juan :)

[Ken Tanaka]

Juan,
I love a good discussion, too!

Don't misinterpret my remarks, above, as saying that expanded color space and lower compression don't produce better results. Certainly, they can.

My point was not whether or not such a feature could technically be introduced into prosumer cameras, but whether or not it would enhance product marketability. That's ultimately the deciding factor. Jeff's remarks are also on-target; why discount features you can easily sell for full-price in a different market segment?

But who knows? How many of us would have predicted 3-4 years ago that we would have the DVX100's or HD10's features in such inexpensive packages today?

[Mike Rehmus]

Originally posted by Juan P. Pertierra :

"Converting video that was recorded as 4:1:1 to 4:2:2 doesn't do anything for quality, and it's not what i'm talking about."

Actually that is not exactly true.

Good editing systems convert to 4:2:2 and perform all of their edits at 4:2:2. Makes a big difference to Chromakeys, the quality of color changes, etc.

Even better systems also convert to more than 8-bit video so that the effects and transitions, when converted back to 8-bit for output, look better.

"If this is true, then why doesn't everyone just use a cheap $200 camcorder to do their feature film? Do ~you~ use an Emerson $100 VHS-C camcorder? If not, does that mean you're just crazy about the technical aspects, which only you will notice and no-one else?"

No, but then those cameras really do deliver inferior image and audio. They also do not have the manual controls that are necessary for use in the pro and prosumer applications. Every one of those controls costs a lot of money when you consider the design, test, and manufacturing of the product. A dime saved can be $100,000 in profits and that's important to any product's success. In many respects, saving cost is better than selling more.

You are right, it is a visual medium. Strange thing is that the current cameras give a very sharp, crisp, clear image and then some opt to make it look like blurry, strobing film. Go figure.

BTW, do you know that there are only a few manufacturers of image blocks? Sony and Canon are the big names.

The group of camera manufacturers in Japan are like a club. They all meet several times per month, they all agree what and how they are going to accomplish their tasks, and they are all driven by the bottom line. CCD block features are designed by a committee of many different manufacturers. If you take a optical block apart on any of the consumer or prosumer cameras, you'd know how cheap they really are. Elegant but cheap.

[Juan P. Pertierra]

>Good editing systems convert to 4:2:2 and perform all of their
....
>for output, look better.

Of course! if you take an 640x480 image and blowit up to 2000x1000 and render effects at that resolution, the EFFECTS will look better when down sampled back to 640x480! But that doesn't change the fact that the original color and resolution data was not there. if that where the case, all we would need is a tiny CCD, and up convert it to whatever resolution we wanted like you're stating.

>You are right, it is a visual medium. Strange thing is that the >current cameras give a very sharp, crisp, clear image and then >some opt to make it look like blurry, strobing film. Go figure.

This is the first ever mention i've ever heard of film(as in 35mm) being blurry. I'm sorry, but for something that has an effective resolution of nearly 2000 lines per frame, this doesn't qualify as blurry to me.

Further more, if DV is so crisp as you say, then all those hollywood directors must be total morons to be paying thousands of dollars an hour to use a panavision, when all they need is the crisp image of a DV camera.

I don't think so.

[Mike Rehmus]

"This is the first ever mention i've ever heard of film(as in 35mm) being blurry. I'm sorry, but for something that has an effective resolution of nearly 2000 lines per frame, this doesn't qualify as blurry to me."

Whenever there is movement, the image blurs. Has nothing to do with resolution. A slow shutter speed does the smearing

"Further more, if DV is so crisp as you say, then all those hollywood directors must be total morons to be paying thousands of dollars an hour to use a panavision, when all they need is the crisp image of a DV camera."

Well, yea, maybe. Just because everybody does it does not have to make sense.

It has to do with issues that were posted earlier in this thread. The folks who really want 24 p, for example, tend to be the guys who want to be 'just like the movies.' From teenagers down, the kids are used to crisp images. And not even continuous tone images. Hence the popularity of Japanese anime.

Ken's first post hit it right on the head:

"It isn't until you reach the hallowed inner cloisters of video technicians that such matters loom large. I would make an argument that even in that "professional" environment it's a technical nit within the context of the overall value of a production (i.e. the story, the acting, the subject, etc.)."

And I claim that applies to the 24 P camp too.

I do agree that HD is great. Unfortunately most of the public have no way of playing HD.

[Juan P. Pertierra]

>Whenever there is movement, the image blurs. Has nothing to >do with resolution. A slow shutter speed does the smearing

????
Last time I checked, every professional film DP has the ability to change the shutter angle and frame rate(or switch camera) to suit whatever movement depiction is needed, given the right budget. If i'm not mistaken, some slow-motion shots from the matrix(not talking about 'bullet-time') where shot at over 100fps, and last time i checked, there isn't a drop of blurriness in them. Try that with DV. In addition, just because 24fps is the standard for projection, this does not eliminate the techincal flexibility of the medium that I am talking about. If you really wanted to, you could film and project at 30, 60 or whatever fps you wanted to. In the worst case, you might need to come up with new hardware for very hi-speed projection, but the flexibility of the medium is already there.

What you're suggesting is that anyone who shoots 35mm is limited to a slow shutter speed and thus will be blurry. That's like saying that anyone who drives a ferrari has to drive at 15mph and thus should just trade it in for a Yugo. You >can< drive 15mph in a ferrari...but it's good to know there's more there if you want.

>I do agree that HD is great. Unfortunately most of the public >have no way of playing HD.

Right. But HD(and film), crunched down to SD still holds more detail and color than what your DV camera can do at the same resolution. That is a fact. Current miniDV cameras simply capture less detail than HD or 35mm film cameras do. Which brings us back to what I'm saying....it would be >nice<(not gonna happen soon, we agreed that) to be able to get more out of the camera, even if it means having to bypass the tape format.

To me, the new wave of people asking for progressive modes, 24fps 30fps on their DV cameras, are for the most part, people not just blindly after the film look...they are looking, just like most of us, for >flexibility<. Like I said, the more latitude, the more control you have over the image, the more you can express yourself. And what gives you more control, is simply more data to work with.

Just like there are people asking for 24p, others are jumping up and down about the 60p mode of the new JVC camera because they can now do decent slow motion on consumer DV for the first time. They got more freedom. I'm sure people aren't shouting for higher frame rates out of miniDV not because they don't want them, but because they are aware of the format limitations, and it's just too much to ask for right now.