AG-DVX200 4K Large-Sensor Handheld Camcorder

[Mike Harvey]

Putting P2 (and AVC-Intra) into the AF100 may have increased cost of ownership by virtue of much higher memory costs - but TECHNICALLY there is no reason why it should have made the camera itself any dearer, that's the huge, fundamental change between tape/disc based recording and solid state cards. At the end of it we're talking about an H264 encoder chip, and a different card connector, the latter costing pence. What form of H264 - bitrate, long-GOP or I-frame only etc - the eventual signal takes depends on what the chip is instructed to do. There are very sound economy of scale reasons for manufacturers to have a single encoding chip for many purposes, and enable/disable features according to application.

Proof of that concept came out with the F5 hack to enable internal 4k recording to XAVC. It was proven that the F5 had exactly the same encoding chip as the (much more expensive) F55 - it was simply that some of it's abilities had been software disabled. But (as the hack proved) it was possible to enable such - albeit at the likely voiding of any warranty!

Point is that a high end codec doesn't necessarily mean a far more expensive chip. It's likely for many manufacturers (as with Sony) that the same encoding chip gets used across a variety of product with differing features enabled. Higher bitrates will mean more expensive memory, yes, but that's as far as the cost implications should go from a technical and manufacturing point of view.

And look at the Sony Z100. I don't think much of the front end because of the small sensor, but it's a similar price to the DVX200, and incorporates XAVC as standard - and the I-only modes at up to 600Mbs at that, not just the long-GOP versions. You could say they've put the (codec) guts of a F55 in a $4,000 camera, so why can't Panasonic do the same with the Varicam encoding engine in the DVX200? Or, at least, just the long-GOP version of it so it could still use SDXC? Think about it. They market it as "B-camera to a Varicam", after all!!



Marketing!!! Simple as that. Similar to the F5 and F55, though there is a lot of differences between those two other than internal XAVC recording. Same applies to many, many other products and not just in the camera field. I'm not knocking it - it's sheer design sense to use a single chip across a whole range of product. The cost of a chip is largely in it's design, not production.



But they shouldn't need to design any chips from scratch - just use the ones from the Varicam. Same as Sony have done with F5/55 coding chips in the Z100. What's the difference? Maybe limiting it to long-GOP only modes?

But if it's so simple as tweeking the recording processors, why hasn't somebody done it yet?

Think about it... as far as I know, there hasn't been a hack for the GH3 or GH4. There hasn't been a hack for the HC-X1000. There hasn't been a hack for the HMC150 (given the whole CCD/CMOS debate at the time, if ever there was a Panasonic camera screaming to be hacked...). There hasn't been a hack for the AC130 or AC160. No one ever hacked the DVCPro or early AVC-Intra cameras to record past 100mbs.

And you're assuming that because someone hacked one high end XAVC camera to record XAVC in 4k like the bigger sister version of that camera, that it should be simple to tweak the existing internals of the internally completely different DVX200 to record 4:2:2. My question is, did anyone ever do that with the GH2 hacks? They managed to get all sorts of improved bitrates out of it, but did they ever manage to get 4:2:2?

And if it's just marketing, then why is the PX270 nearly $6,000? It is capable or recording 10bit 4:2:2 to an SD card, so you don't need P2. It is a standalone camera (no lower or higher tier versions of it to protect). It's both a replacement for the HPX250 and AC160 all at once. It would have sold tons more if they lowered the price by $1,000... so why didn't they?

You're right, Sony took what was essentially some of the F5/F55 guts and put it in the Z100. BUT, nearly two years after it's release, it costs over $1,000 more than the DVX200... and it uses a just slightly larger than 1/3" sensor. Meaning, it has a much cheaper lens than a camera that is currently $1,300 less. At release it was, what, $6,000? Adjusted for inflation, the Z100 was almost $2,000 more expensive, using a much cheaper lens. For comparison, the HC-X1000, which appears to share a lot of the same internals (but without some of the external goodies like HD-SDI) as the DVX200... and had the same sized sensor but with a longer lens as the Z100... was $3,500 at release, and is going for $2,900 currently. Is that simply marketing? Clearly, you're not putting the Varicams (or the PX270's) guts into this camera at this price point. Would you be willing to pay $2,000-$3,000 more for those things? I know a lot of folks would, and I've said before Panasonic should release just such a camera.

But doing what Sony did with the Z100, but with a bigger sensor and more expensive lens for $2,000 less?

 

[Greg Smith]

And if it's just marketing, then why is the PX270 nearly $6,000? It is capable or recording 10bit 4:2:2 to an SD card, so you don't need P2. It is a standalone camera (no lower or higher tier versions of it to protect). It's both a replacement for the HPX250 and AC160 all at once. It would have sold tons more if they lowered the price by $1,000... so why didn't they?

Actually, as a practical matter, they did drop the PX270 price by $1000, in fact by $1500. The list price recently dropped from $6500-ish to $5750. And there's been a $750 rebate deal running for about six months now that brings the final purchase price down to $4399 from B&H. That's darn close to the DVX200. And there is a higher end model to protect, the new shoulder-mount PX380, albeit at close to three times the cost.

- Greg

 

[DLD]

Exactly what I also speculated (120%!!) a month ago - http://www.dvxuser.com/V6/showthrea...corder&p=1986563847&viewfull=1#post1986563847
...

Well, a 10% discount is something.

The AX1 has been discontinued so a replacement is likely...

[Johnny Carson] I did not know that [Johnny Carson].

In any case, if Sony doubles the price of RX10 MKII, which would bring them into a ~ $2,500 territory of XC-10. Canon's model has a nice codec and a recording bitrate but is deemed short of a few other things, so that is a possibility. FS100/700 replacement is likely to be one camera and, because it can't compete with FS7, it'd have to be in the $4,000-$4,500 range. (One move would be to bundle the FS7 attachment and/or lens into combo deal at ~ $9,000, once again, giving it a model price spacing of a double off $4,500). And who knows what's going on with the VG900 follow-up. IMO, the era of fixed lens camcorders above $5,000 is largely over and the ILC's with lens bundles will be the next step on the retail level a la G7 but we'll see.

But if it's so simple as tweeking the recording processors, why hasn't somebody done it yet? ...

But doing what Sony did with the Z100, but with a bigger sensor and more expensive lens for $2,000 less?

Look at FZ 300. Almost the same specs as Z100 and retailing for $600 only.

When counting the processors, it's the Moore Law. Today's $300 chip sells for $200 six months from now and for $150 in a year and, at the end of those 18 months, it's $50 in the discount bin. The smartphone processors double in power every couple of years too. That's the digital age in a nutshell.

 

[Mike Harvey]

Actually, as s practical matter, they did drop the PX270 price by $1000, in fact by $1500. The list price recently dropped from $6500-ish to $5750. And there's been a $750 rebate deal running for about six months now that brings the final purchase price down to $4399 from B&H. That's darn close to the DVX200. And there is a higher end model to protect, the new shoulder-mount PX380, albeit at close to three times the cost.

- Greg

True, but at release, the camera had a street price of ~$6,000. I have no doubt that a year after it's release, the DVX200 will be able to be picked up for ~$3,200.

But, seriously, $4,400 currently for the PX270? That's a smokin' deal :shocked:

 

[Mike Harvey]

Look at FZ 300. Almost the same specs as Z100 and retailing for $600 only.

When counting the processors, it's the Moore Law. Today's $300 chip sells for $200 six months from now and for $150 in a year and, at the end of those 18 months, it's $50 in the discount bin. The smartphone processors double in power every couple of years too. That's the digital age in a nutshell.

That may be true for Intel CPU's, but that's not necessarily what we're talking about here. Is Panasonic using off the shelf commodity processors, or brand new ones developed for their specific cameras? If anything, it helps underscore price savings of using the 4:2:0 codecs... if this is the same chips in the X1000 and GH4, then they've sold about a bazillion of them already, so the unit costs at this point is relatively cheap. But if they have to develop new chips from scratch, then there isn't really any cost savings. Even if they used the same or similar chips as the Varicams, they probably sell more GH4's in one month than they have the total number of Varicams. And the Varicam chips were most likely more expensive to develop to begin with, so per unit costs are going to be higher.

 

[harddrive]

But if it's so simple as tweeking the recording processors, why hasn't somebody done it yet?

Think about it... as far as I know, there hasn't been a hack for the GH3 or GH4. There hasn't been a hack for the HC-X1000. There hasn't been a hack for........And you're assuming that because someone hacked one high end XAVC camera .......it should be simple to tweak the existing internals of the internally completely different DVX200 to record 4:2:2.

No, that's reading far more into what I wrote than was intended. There may be many other reasons why any hack cannot be done to any given camera to give such a result - nothing to do with the coding chip. Or it may be hardware disabled. Or an early version of such an encoder. If I've given the impression of saying "this can be hacked, therefore that must be able to be" then I'm sorry. It's not what was intended.

I'm just trying to get across the principle that the cost of any encoding chip is not necessarily going to be related to it's performance in any given unit. The cost of a chip is largely in design, so the unit cost will be lower if such can be used in multiple products. Design for the top end, then disable functions as required for lower end product. Of course, a manufacturer is hardly likely to tell you that...... It *MAY* mean that hacking of higher functions is possible, but not necessarily. (A higher function may generate extra heat - no problem in product A, but may lead to rapid overheating in product B, for example.)

And if it's just marketing, then why is the PX270 nearly $6,000? It is capable or recording 10bit 4:2:2 to an SD card, so you don't need P2...... It would have sold tons more if they lowered the price by $1,000... so why didn't they?

I suspect it's because the P2 is geared towards broadcasters with deep pockets, and who are looking for card and codec compatability with other cameras they may have - and are prepared to pay for it. Maybe they would have sold more with a $1,000 price drop - but if those sales are at the expense of AC160 sales, would that be a good move? Somehow, I doubt it.

BUT, nearly two years after it's release, it {the Z100} costs over $1,000 more than the DVX200... and it uses a just slightly larger than 1/3" sensor. Meaning, it has a much cheaper lens than a camera that is currently $1,300 less........

But doing what Sony did with the Z100, but with a bigger sensor and more expensive lens for $2,000 less?

I'm not sure where you got those prices, but looking at UK pricing for the DVX200 ( http://cvp.com/index.php?t=product/panasonic_ag-dvx200 ) and the Z100 ( http://cvp.com/index.php?t=product/sony_pxw_z100 ) they're virtually identical for pricing over here. And I'm not for a moment saying I like the Z100 front end, but I think it's wrong to draw any conclusion that small sensor must mean much cheaper lens. Small sensor may make some design aspects easier - but on the other hand it's f1.6 max, and 20x zoom.

But DVX200/Z100 isn't by any means a like for like comparison, and that was never the intention. The Z100 was only mentioned to show how a full high end codec can be in a camera costing the same as the DVX200. Let's just see if a Z100/RX10 hybrid appears, which a few of us think highly likely. That would be an interesting comparison to the DVX200.

 

[Barry_Green]

There's a lot of speculation going on, and people presenting their opinions as if they were fact. The only fact we have here is that the folks making those speculations are NOT camera manufacturers. The truth is you DON'T KNOW what it would cost, you have no idea what it would cost, and presenting your opinion as fact is not only misleading, it's annoying.

Here are the facts: Panasonic is making a product to sell. They have decided within their team as to what features it should offer and at what price. You, as an individual buyer, can decide whether you think the product they produce is good value or not.

That's it. Any speculation on hardware chips or whether or not they could easily and simply add 4:2:2 and whatever, are speculation and with all due respect, many folks making those speculations simply do not know what they're talking about. There may very well be hardware limitations in the available chips that you don't know about. There may be cost issues. Is it possible to design a 4K processing engine that processes and records 10-bit 4:2:2? Of course -- but at what cost? You don't know -- so please stop declaring it to be a trivial task. Folks who are doing that are just making themselves look silly.

If there is an available, off-the-shelf component, then yes, sure, there's reason to speculate (but it would still only be speculation). But what if there is no other 4K/60p/10-bit/4:2:2 chip available? What if it simply doesn't exist? What if it's a proprietary product to one manufacturer, and that manufacturer refuses to sell it to another manufacturer? In that case, it might as well not exist. And so, if they have to build it from scratch, you can **** sure bet your *** that it would increase the cost of the camera more than a couple of pennies, right?

How about we let the camera manufacturers make the products, and we the customers do the buying (or the waiting, as appropriate?)

Putting P2 (and AVC-Intra) into the AF100 may have increased cost of ownership by virtue of much higher memory costs - but TECHNICALLY there is no reason why it should have made the camera itself any dearer

This is an example. With all due respect, you do not and cannot know that. You simply cannot. You are not privileged into the manufacturer's accounting, you don't know what their chip costs are, you don't know what their quantities are, you don't know what their economies of scale are, you don't know what product lines share what chips, you don't know what the regional differences may amount to. All you are doing is assuming, and it's not helping anything.

 

[CraigM]

One of our local dealers had a Panasonic rep in yesterday with the DVX200 doing demos with a pre-production/engineering model. Had a chance to hold it a bit and poke around. For someone who comes from an era when you only had manual focus the programable auto focus with the touch and pressure sensitive screen is pretty sweet. They had a small table top scene with colorful kids toys set up with the camera feeding a 4K monitor via HDMI. I really liked what I saw, decent enough shallow depth of field and nice color rendition. Turned off the lights and cranked it to 2000 ISO and it sill looked pretty good. Felt about as heavy and handled like a HPX250 to me.

Have owned and used a lot of cameras in 30 years of being in business in NW Florida. Like many of you my work involves many levels of production. Would I use this for a high-end, high production value commercial job? Maybe, maybe not, will have to wait to see what the final model is like. But what I do need for many other jobs is a camera I can pull out of a bag and put on the sticks quickly. One with built-in ND's, balanced audio, viewfinder, a decent LCD/LED and 4K that looks nice scaled to HD. 10-bit out of the HDMI is a nice bonus for external recorders as is the decent sized sensor.


Up until this demo I was leaning towards a C100 for this kind of work as we have lenses for it, but you know how many times I am running-and gunning with manual zoom lenses and screw on or matte box ND's and I wish I had a nice power zoom and ND's built in? A lot. (I know the c100 has built in ND's) This 4K camera really has potential to fill a need I and my customers have. If it ends up being even more useful for higher end commercial work I'm all for it.

 

[Josh Bass]

cool! lets pray they get all the issues we've been bitching about (skin tones etc.) worked out by ship time.

 

[harddrive]

This is an example. With all due respect, you do not and cannot know that. You simply cannot. You are not privileged into the manufacturer's accounting, you don't know what their chip costs are, you don't know what their quantities are, you don't know what their economies of scale are, you don't know what product lines share what chips, you don't know what the regional differences may amount to.

I don't disagree with the detail of that - which is why I stressed and capitalised "TECHNICALLY there is no reason".

TECHNICALLY there is no inherent reason why within a given codec family (such as AVC/H264) a coding chip should cost any more if it's encoding high bitrate I frame only, or low bitrate long-GOP. That is not speculation, and I'd quote the Sony XAVC "hobbling" of a common to multiple products chip as just one bit of evidence in support. The basic principle is common (and, IMO, sensible) manufacturing practice throughout industry. It's no different in principle to buying a car at the bottom of a model range and finding blanked off holes where extra switches may be fitted in models higher in the range. There are economies of scale for the car manufacturer in having a single design of dashboard, and fitting blanks or switches according to model.

The factors you quote - such as manufacturer's accounting, economies of scale etc - are indeed a different matter, you're quite correct. But they are not TECHNICAL reasons why a camera with true pro codec must *of necessity* cost far more than a consumer one.

In tape days, a DVCProHD tape deck was vastly more complicated than such for DV, or even DVCPro for that matter. (Let alone economies of scale.) The former cost far more to manufacture than the latter - so was inevitably far more expensive. But chipsets aren't like that.

 

[Barry_Green]

I don't disagree with the detail of that - which is why I stressed and capitalised "TECHNICALLY there is no reason".

But again -- you do not KNOW that. Look, I was a professional computer programmer for nearly two decades. I know that there can be limitations where the layman doesn't think there should be. It doesn't matter what the layman thinks, what matters is what the chip can do or not do.

TECHNICALLY there is no inherent reason why within a given codec family (such as AVC/H264) a coding chip should cost any more if it's encoding high bitrate I frame only, or low bitrate long-GOP. That is not speculation

Yes it is. You don't know what any given chip can do -- you just don't. Just because one chip can do it, doesn't mean another can. The Intel 80386 runs all the instructions that the Intel 8086 does. But the 80386 could do a hell of a lot more things than the 8086 could, and there was no way to "hack" the 8086 to do everything the 80386 could.

, and I'd quote the Sony XAVC "hobbling" of a common to multiple products chip as just one bit of evidence in support.

All that proves is that the chip used in the F55 was capable of being hobbled. It does not prove that any and every h.264 chip is capable of doing more. I'd point to the GH2 "hack" as evidence of that -- sure, they could crank up the bitrate, but then the camera became unreliable. For professional reliability, you operate chips within their design parameters. If the F55 chip was later used in a less-expensive camera, that was Sony's choice, and that's a fine choice -- IF you have the chip that was capable of the higher performance in the first place.

The basic principle is common (and, IMO, sensible) manufacturing practice throughout industry.

Again -- your opinion is your opinion -- it is not documentable fact. You don't know what the chips can actually do, you're assuming that just because one chip did it, that they all can. And that's just not so.

It's no different in principle to buying a car at the bottom of a model range and finding blanked off holes where extra switches may be fitted in models higher in the range. There are economies of scale for the car manufacturer in having a single design of dashboard, and fitting blanks or switches according to model.

Wildly appropriate and inappropriate analogy. But instead of looking at switch holes, let's look at, oh, say, engines -- where there's actually a direct correlation. I do not know of a single manufacturer of cars anywhere in the industry who regularly builds an 8-cylinder engine, and then just disables 4 of the cylinders in their economy models. Do you? Sure, they may gain great economy of scale, but -- they don't do that. It would involve excess weight and poor mileage to do that, so they instead use a differently-designed engine for the economy models, because it makes overall sense.

Have you heard of anyone hacking their 4-cylinder car and turning it into an 8-cylinder? I sure haven't.

But they are not TECHNICAL reasons why a camera with true pro codec must *of necessity* cost far more than a consumer one.

It is absolutely "of necessity" if, for example, the chips that you are talking about are NOT AVAILABLE to the manufacturer in question. That's just one glaring example that points out the folly of asserting that there is "no reason".

If you're trying to say "If I was king of the world, I'd make sure all chips were designed so that they could be re-used and make great economies of scale". Fine -- say that -- and we'll see if you get hired as the head of R&D at a major camera company, maybe they're all incompetent and they're just waiting for someone to drop this pearl of wisdom on them.

I, on the other hand, having been in professional software development, having worked on many different platforms, and having watched camera development closely for years, and having toured their factory, will stick with a simpler explanation: they know what they're doing, and we don't. They know their business and we don't. There may very well be reasons and limitations that we don't know about -- and that's okay, because there's such a thing as trade secrets and proprietary information and competitive advantages and all that, which mean it isn't any of our business anyway.

The former cost far more to manufacture than the latter - so was inevitably far more expensive. But chipsets aren't like that.

And you know this because you're a professional chipset manufacturer?

With all due respect, you simply cannot assert the things you're saying. You have no idea what's involved between the companies, their own factories, their partnerships where they buy chips from other companies, their licensing agreements, their patent and royalty obligations, the quantities they're manufacturing in, the heat and power requirements to run a more-powerful chip in a smaller-body camera, and on and on and on. I'm trying to be polite here, I'm not trying to be insulting -- and I'm saying, you do not know that you're asserting is true when it comes to THIS manufacturer, making THIS product. You have no idea what the complicating factors are. None of us know what they are.

What I don't understand is -- why do people automatically assume that the manufacturer is just stupid? What is this notion that someone sitting behind a keyboard somewhere, who works in a totally unrelated field, is somehow smarter and wiser than the people who've spent their entire careers making products like this? Why can't people just say "Hmmm, I don't understand the reasons why, but I'm sure there must be some reasons."

Anyway, this is entirely sidetracked and isn't talking about the purpose of this thread anymore, so hopefully we can just put this to bed with the generally accepted wisdom that it's never wise to assume, and when you don't actually know for sure what you're talking about it's best to not talk about it because it contributes nothing of value to the conversation.

 

[DLD]

That may be true for Intel CPU's, but that's not necessarily what we're talking about here. Is Panasonic using off the shelf commodity processors, or brand new ones developed for their specific cameras? ...

There are three different processor types that can be in play - the ARM based processor for the operation of the camera; the graphic chips; and the logic board/s (either ASIC or FPGA). The former two are off-the-shelf simply because that's how it is. These chips are built in millions, if not billions. Beyond the sensors, the ASIC/FPGA boards is what usually separates an X-camera from a Y-camera. The boards are individually programmed for each product - naturally, there are overlaps - but the components of these boards are off the shelf as well. None of it is a secret. People have been breaking down designs of a lot of cameras and smartphones to find out your average cost of its internal components. There are actually whole sites devoted to it.

http://www.techinsights.com/teardown.com/SamsungGalaxyS6/

Naturally, this doesn't mean that the engineers just slap sh!t together and then send it out to a retailer. It's more like building a car, where you know your choice of engines and gear boxes and what not but you still have to put together the whole package and that means huge engineering departments working for years to get a particular model ready for use.

 

[Darren Levine]

ya know what would be great? if people who didn't actually know, stopped pretending to

 

[R Hudson]

ya know what would be great? if people who didn't actually know, stopped pretending to

BEST post so far!

 

[demoboy]

So the current industry camera with the most potential and bang for the buck is the X70 for only $2K and providing XAVC Long-GOP codec with 4:2:2, 10-bit color sampling at 50 Mb/second. IF, by the holy ghost of Sony allows 8-bit 4:2:2 for its 4K upgrade - because some smart intern whispers that the Canon XC10 has 8-bit 4:2:2 - then as Tony Montana says, ''nowee talkin baby!''

 

[Alpha Beta Omega]

But the IQ from the X70 sucks in my opinion despite the 10-bit 4:2:2 recording. Wasn't 4k from the X70 also an embarrissing 60mb/s? While I think the XC10 looks great with 8-bit 4:2:2.

If a camera has poor implemented image quality overall 10-bit or 8-bit doesn't matter.

 

[demoboy]

If a camera has poor implemented image quality overall 10-bit or 8-bit doesn't matter.

poor image quality? the 1:1 image of 4K of the X70 looks great, and let's not forget much of the contextual criticisms regarding the type of camera choice is regarding the optical zoom.

 

[philiplipetz]

But the IQ from the X70 sucks in my opinion despite the 10-bit 4:2:2 recording. Wasn't 4k from the X70 also an embarrissing 60mb/s? While I think the XC10 looks great with 8-bit 4:2:2.

If a camera has poor implemented image quality overall 10-bit or 8-bit doesn't matter.

Yes, the color looks great on the XC10. And on the XF series of 4:2:2 8 bit. both better than the more expensive C100 8 bit 4:2:0 cameras our nonprofit owned(not that those colors were anything but good). I think of those colors whenever anyone tells me that 4:2:2 does not matter. It does matter with a proper implementation. With a poor implementation it does not matter. And I think that is the cause of a lot of confusion - looking at poor 4:2:2 such as the X70. I know the theory and I know what my eyes tell me.

 

[jc01108]

I find it interesting that Panasonic is 'crowdsourcing' the Beta testing of the camera and its features prior to launch. This is actually a pretty smart move on their part. They get feedback on the camera, its feature set and the software prior to the hard launch of the product. They also create a buzz about it, not all positive by any measure, but a buzz, none the less. I have learned a lot about the camera here and elsewhere and will seriously consider it once I see footage and read about the launch feature set and software; it may be a nice compromise/upgrade from the EX1r and AF100 and other cameras I now use for various projects. I can wait until October to find out what the camera will do for me, in fact, I don't have any choice as Panasonic won't be sending me one to test. In the meantime, I will monitoring this topic to learn as much as I can.

 

[Mike Harvey]

I've been editing all day, and in the back of my mind all I can still think about is "Holy crap, the PX270 is $4,400!" :shocked:

If it wasn't for the DVX200, I'd be all over that right now.