Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

4K NAB WAR - SUMMARY

Hey, you forgot the cheap 4k JVC for $5k also (heard there is also a $2k camera coming). How I wish the $3k BMD had 4k, only 2.5k.
 
Hey, you forgot the cheap 4k JVC for $5k also (heard there is also a $2k camera coming). How I wish the $3k BMD had 4k, only 2.5k.


If it had 4k it wouldn't be $3k. If you want 4k you need $--K
 
Hopefully, but then again it can be much less. It is not useful basing things on perceived value, which happens often in professional discussions, as the manufacturer suppliers control perceotion and distort reality. A cheap nex5n is only a few dollers off of doing a similar trick, as far as uncompressed goes. Cheap consumer brands are one of the main places to get an real idea of what can be done in real value, because until you go really high end performance quality and features, the costs do not go up sharply for the extra given performance.

The reason you see greater than 4k resolution consumer stills that pull greater than a 4k frame off the sensor at video rates for interpolation and focusing internally etc, is that it is CHEAP. We allready know there is a great capacity there in a professionally cheap camera to compress it that is not used. If you want real value that costs a lot, you are talking 8k recording in (3d) from the last quote I gave somebody to design a system. If you look at 8k there is very limited caoacity to compress it in cheap cameras, and i personally do not know fir certain anynthat can pull a 8k frame off, but if you look at it as 4 4k compression capable $1000 cameras, you are talking about... in real value. Why do you think Red is working on a 28k sensor, if 4k is much technically, great Casino security camera, lots of money to be made, because it delivers very high quality for evidence and wide coverage.

Raw wavelet compression is the tricky one, until you have mass market parts that do it properly, it is a small production run custom job, costing, like Red is doing. But the reality is, that there may well be a gpu out there somewhere now days, that can burn through it at a cheap cost. People, that should know better, make assertions about the incapacity of certain GPU's to handle wavelet compression, but it is a lot to do with how you design your codec to work around the limitations, and then, picking a gpu design without much limitation. I'll spell it out, old trick from Forth community, transform your maths and how you use it to something more compatible with the hardware used. Parallel process the transfers too. So even wavelet bayer compression can be cracked one day, but I thought they already made some advance, at least in playback side, have not been tracking it. My alternative scheme (non wavelet) uses much simpler hardware, that you can then pack more units in. All this stuff around is primitive.
 
Revised that.
 
Which consumer camera reads out the full sensor (4K) at 30fps?
That's news to me.

The reason you see greater than 4k resolution consumer stills that pull greater than a 4k frame off the sensor at video rates for interpolation and focusing internally etc, is that it is CHEAP. We allready know there is a great capacity there in a professionally cheap camera to compress it that is not used.
.
 
When Red decided to move Scarlet to a higher price point (with corresponding higher capabilities), I was disappointed. There was something very seductive about 3k for 3K, or 3K for 7.5K even. But now that we can see how the market is evolving (1080p getting eviscerated by cheap DSLRs), I must acknowledge that Red took an unpopular but correct decision.

I have to say I broadly agree.

I loved the concept of Scarlet Fixed, and if that came back from the dead I'd *still* buy one tomorrow. (1)

Ditto Scarlet Cine; loved the idea of putting all that 16mm glass to work.

Various things were said when Scarlet-X was announced and the 2/3" was killed. At the time, I said 3K vs. 4K wasn't what mattered, what mattered to me was **RAW**. R3D was the real selling point, the must-have feature, and I stand by that.

Now, having bought Scarlet-X and more recently got my EPIC-X, I'll say Jim was mostly right and I was partially wrong. I can still use that old 16mm glass at 2K-3K, in fact I'm having a like-new, never-used, Angenieux 12-240mm remounted to PL even as I write. But Scarlet-X is a hell of a more capable and more useful camera than the 2/3" concepts would have been.

(1) and I don't accept that Scarlet Fixed, in some form, is irretrievable. The big problem was the lens; despite the best in modern computer design and modeling they somehow ended up with a lens that wasn't parfocal as intended. Well, you know what? Now or soon, with all-electronic lenses, they won't be made parfocal by design and optical craft, they will be made parfocal by software and electronics, making the needed constant fine adjustments to focus automatically, throughout the zoom range. In fact there could be a LUT for all focal length combinations that would make the lens perfectly parfocal and perfectly breathing-free, and customised to the inevitable tiny manufacturing variations apparent in each individual lens...

Mike
 
Till now this is the best summary of NAB 2012
4k camera availability, and is courtesy of nofilmschool.


$60,000 – Arri Alexa – 12-bit 3K RAW
$65,000 – Sony F65 – 16-bit 4K RAW

For all of the RAW cameras (except RED), you’re going to have to
add $5,000-$10,000 (or more) for RAW recording ability.

I'm sure I saw a much higher figure mentioned for Alexa; wasn't it something like $50k for the Codex RAW box?

What's the equivalent number for the F65? That base $65k price doesn't include the recorder?

Mike
 
Which consumer camera reads out the full sensor (4K) at 30fps?
That's news to me.

Nikon j1 sensor can, and it is supported by the ambarella chipsets that are used in those sub $200 cameras for use in non sensor debayering interpolation (apoarently many sensors do a terrible job of it) and useful for focusing. It is just we don't see it supported as full res video feed. The DMD could have used something like that, and when queried if they were going to do 4k, a Nikon executive indicated they would in future (presumably on a future model, not the J1, but I think people would go ape if a 4k 50/60 output was available for the J1, but the HDMI and flash interface probably dies not support it (I think the sensor might do 50 or 60 fps). The problem is that Panasonic has not done this, rumors are the next camera will be 1080p60, what the gh2, even the gh1.3 should have been. That would be excellent.

Reading all the stuff, you see that BMD tried to hunt down very good s35 sensors, but they could not get the price, but the Nex5n has one at $599, so they are out there, if you can get hold of them, and Sony is one of the biggest sensor suppliers.
 
The Nikon sensor, might not be the most ideal, but should be available cheap from Aptina, once any supply contract runs out. There is little to stop BMD or Red even, going that path for a low end no frills system, instead of the scarket fix, a real b grade 4k camera done right with raw. It is easy, I can say easy, to get something like this in a small module that can go in a pro camcorder case similar to Jvc's, or Eng case, or as a digital cinema camera format, for the user to swap around at will (but people would buy 1 or 2 modules to go with there cases). In matter of fact, with a low cost system like that you would just buy 2 or three modules, incase anything goes wrong on the job, no rental or waiting for repairs, swap and go (with a little replacable fuse moduke uncase it is the case or the power nuking the module). Is it really that hard to make some moderate imorivements that will vastly improve the industry, in that way?

I spoke to somebody the other day, chasing some obscure little technology, the guy was instantly switched on, also worked on the tech side, successful, easy to talk to. Turns out he is from up near where I live too, but what he said (without knowing I was of similar nature) was spot on, something along this intent (I forget), even if you come up with the most obvious, simple, great idea, you will have to literally bash the ordinary person over the head with it to get them to see it, I just chuckled to myself and said I knew. That is the reason I am getting ready to leave this group, too many people cannot see the obvious, they only see the negatives as to why it cant be done, but not how to solve things so that it can be done. I have been in the low end HD+ digital cinema camera industry from the beginning, and many things in the cameras are suggestions I directly seeded, and my guiding hand has been behind a few of the cameras (the only ones successful before Red). I feel I have wasted a number of years here, that could have been better spent elsewhere, even talking to professional naysayers that it is no wonder they have so much time to be around here on me instead of being busy doing high end large production run product engineering at a large company, with the reality of the economics, business and marketing end of things that really drive how products are engineered and delivered, that even engineers in those positions might be lucky to ever be privy too, for legal reasons. However, every top engineer I meet, is a meeting of minds, and with the advice of my dear departed friend, Harold Jung, I need to find a lot more intelligent people to hang around with, even the smart assed ones at Uni were barely cutting it.
 
correct mike,

the arriraw onboard recorder with a mag and transfer station would be roughly 50k on top of the 60k of the alexa.

the sony "package" is sub 80 k and comes with everything needed to do full f65 SR SQ raw. the PC-4 which is a vtr and data offloader, the rc-4 backpack SR media recorder, an additional RC-1 recorder for creating another deliverable from the dual SDI output, and 3 SR sticks 2 x 128GB and 1 x 512GB. unfortunately the 128GB sr sticks are too slow to record raw but the 512 is capable.



I'm sure I saw a much higher figure mentioned for Alexa; wasn't it something like $50k for the Codex RAW box?

What's the equivalent number for the F65? That base $65k price doesn't include the recorder?

Mike
 
Let's not forget that currently Arri doesn't offer a way to record ArriRaw on the Alexa.. you have to buy a separate third-party recorder and strap that on top of the Alexa... AND that recorder with recording media costs AS MUCH AS the Alexa itself. To do this list properly, you would need to factor in all of the accessories needed to make a recordable image at 4K or above specs...
 
I have to say I broadly agree.

I loved the concept of Scarlet Fixed, and if that came back from the dead I'd *still* buy one tomorrow. (1)

Ditto Scarlet Cine; loved the idea of putting all that 16mm glass to work.

Various things were said when Scarlet-X was announced and the 2/3" was killed. At the time, I said 3K vs. 4K wasn't what mattered, what mattered to me was **RAW**. R3D was the real selling point, the must-have feature, and I stand by that.

Now, having bought Scarlet-X and more recently got my EPIC-X, I'll say Jim was mostly right and I was partially wrong. I can still use that old 16mm glass at 2K-3K, in fact I'm having a like-new, never-used, Angenieux 12-240mm remounted to PL even as I write. But Scarlet-X is a hell of a more capable and more useful camera than the 2/3" concepts would have been.

(1) and I don't accept that Scarlet Fixed, in some form, is irretrievable. The big problem was the lens; despite the best in modern computer design and modeling they somehow ended up with a lens that wasn't parfocal as intended. Well, you know what? Now or soon, with all-electronic lenses, they won't be made parfocal by design and optical craft, they will be made parfocal by software and electronics, making the needed constant fine adjustments to focus automatically, throughout the zoom range. In fact there could be a LUT for all focal length combinations that would make the lens perfectly parfocal and perfectly breathing-free, and customised to the inevitable tiny manufacturing variations apparent in each individual lens...

Mike

He, he, Mike, forgot to comment on yours the other day. I questioned their decision to stick with 3k, pushing to go 4k or drop the fixed for the $5-7k (I forget which originally) s35, that got canned for epic s (scarlet x), as the fixed was too little resolution and predatory on the Epic line. I pushed for 2/3rds lens on s35 using cropped resolution. The reality is the fixed would have been the camera you owned when you wanted 4k in the future. It does not matter if not much could be done with 4k last year or this, people are still likely to have the fixed next year and two years after that, some for ten years. So, it could be possible you could have spent much of the camera life wishing to upgrade. 4k upfront, you don't care so much for the next 5-10 years. Plus, Red could have kicked everybody else but for a while.

The lens issue makes me wonder, what happened, who cares if it had to get redesigned and stuck out even three inches (with some Eng like lens control attachments), or stuck to the side. The first cheap high quality 4k camera under $3k could be de bomb (or so my young friends used to say) unless there are just so many of them it dies not matter or something else.

Manufacturing in America in house instead of foxcon, was going to cost. Guys with long production run experience can zip down assembly and prices. This is electronics, hardly easy compared to certain things.

What you say about computer correction is true, I realised in my own lens technology I could just computer correct much in post and design differently. But the reality is better on 8k down scaled to 4k then 3k (pixels count).

I'll give you an award for insight fullness Mike.
 
You know, years ago when we started recording HD and fullhd off camera heads to computers, it was not practical to record 4k, so Red did a custom job, ;) , happy they did. But in tge future a $500 computer may do the same trick in 4k, the tech should be there, at least in part, and arm processors past that. I had my eye on a 5kx5k s35 chip.

Thanks


Wayne.
 
Alexa etc, are so not the future to be concerned about, in a way.
 
Was Apertus at NAB, they have announced they are doing a low cost 4k digital cinema camera design, have a sponsor, good on them, will keep people on their toes.

http://www.apertus.org/en/apertus-camera-dp-team-up

Will take a while, and Red will be on to 6k or 8 by then. The dream lives.

NAB for camera release is overated, that is why everybody turns up ;)
 
If we are talking numbers the sensor bit depth of the AD is as important as camera output in my view.

Sony have increased the AD of broadcast video and the F65 from 14bit to 16bit. It all starts at the A/D.


Mike Brennan
 
Back
Top