Jump to content


Photo

Bayer Filter'd 1 sensor vs. 3-Sensor Camera


  • Please log in to reply
34 replies to this topic

#21 Jim Jannard

Jim Jannard
  • Basic Members
  • PipPipPipPip
  • 361 posts
  • Cinematographer

Posted 14 December 2007 - 12:27 AM

Easy way to shut him up is just set up a test with him... then when he's proven wrong he can say he's sorry...


I invited Phil to come to RED and do anything he wanted to test the camera... he hasn't shown up. The invitation is always open.

Jim
  • 0

#22 Gary McClurg

Gary McClurg
  • Basic Members
  • PipPipPipPip
  • 304 posts
  • Producer

Posted 14 December 2007 - 12:41 AM

I invited Phil to come to RED and do anything he wanted to test the camera... he hasn't shown up. The invitation is always open.

Jim


Then maybe fly him out... who knows maybe you guys will become best buds... now I don't know if he's your most vocal critic... but when you get them on your side... they might do more for you... just a crazy thought for a late cold night...
  • 0

#23 Stephen Williams

Stephen Williams
  • Sustaining Members
  • 4708 posts
  • Cinematographer
  • Europe

Posted 14 December 2007 - 03:01 AM

Then maybe fly him out... who knows maybe you guys will become best buds... now I don't know if he's your most vocal critic... but when you get them on your side... they might do more for you... just a crazy thought for a late cold night...


Hi Gary,

AFAIK Phil was in LA a few weeks ago. He did see footage at IBC 2 years ago, maybe he has seen other demos too.

Stephen
  • 0

#24 Jim Jannard

Jim Jannard
  • Basic Members
  • PipPipPipPip
  • 361 posts
  • Cinematographer

Posted 14 December 2007 - 03:18 AM

Hi Gary,

AFAIK Phil was in LA a few weeks ago. He did see footage at IBC 2 years ago, maybe he has seen other demos too.

Stephen


If Phil saw RED footage and said it had low resolution, he is the only person in the world that thought so... if Phil was in LA, why didn't he stop by and shoot some tests? We would have welcomed him (we still would). But I am pretty tired of the resolution trashing and it is very easy for Phil to resolve instead of the ongoing banter. But I don't feel any need to spoon feed him. No one else has a problem with our 4K resolution and it is not my job to "fly him out".

As for Graeme... one could only hope to be the contributor that Graeme is to RED. We have filed many patents based on Graeme's inventions and it is pretty hard to listen to someone who has no inside knowledge of what is going on talk down about Graeme. Not only is Graeme brilliant, he is a wonderful and delightful person. Someone many of us could learn from.

Jim

Edited by Jim Jannard, 14 December 2007 - 03:22 AM.

  • 0

#25 Max Jacoby

Max Jacoby
  • Sustaining Members
  • 2955 posts
  • Other

Posted 14 December 2007 - 06:36 AM

I don't recall Phil saying anywhere that the camera is not high enough resolution.
  • 0

#26 Keith Walters

Keith Walters
  • Sustaining Members
  • 2252 posts
  • Other
  • Sydney Australia

Posted 14 December 2007 - 07:30 AM

http://www.high-tech.../Camera_t_1.htm

This analysis claims 65% of the full horizontal resolution and 80% of the full vertical resolution when using Bayer filter

Kodak, which invented the Bayer filter has similar claims.


Other sites and test claim 1/2 resolution for Bayer filter:

http://www.scss.com....olution/#camres

http://lagemaat.blog...ensors-you.html


How does Red do it? How are they able to achieve full 4K resolution from a 4K single CMOS sensor with Bayer filter when neither Sony, nor Panasonic are able to do it? I do understand that Red is using some innovative algorithm that is superior to Kodak, Sony, Panasonic and everyone else. Could someone shed some light on it? Is Red using licenced technology? Is it their invention? Is it patented? Could someone direct me to their patent number?

I have hard time believing that Red can achieve 4K from a 4K sensor, but everything i possible. The Red concept is quite innovating.


I started this thread about three months ago on this very subject. That rapidly degenerated into an online punch-up and was eventually closed.

I had already written another post for it "off-line" so I then started this thread, which rapidly degenerated into mind-numbingly tedious debate over the decisions made by TV engineers nearly 70 years ago!

As for the pronouncements about Bayer masks made by Kodak engineers 30 years ago, that assumed the decoding would be done with extremely inflexible analog processors, glass delay lines and so on. That's very much a 20th century "Feed Forward" paradigm: The raw signal is fed in at a fixed clock rate and RGB (or component) video comes out at the same clock rate. There is no "intelligence". The processor is very much like a car engine; there are hundreds of intricate parts all working together, but they all do the same thing over and over again, and there is no way of making them do anything different.

The first all-digital Bayer decoders were simply digital devices that emulated analog decoders in software, because that was all the technology was capable of, speed-wise.

However in recent years microprocessor technology has advanced tremendously, to the point where modern Bayer decoders bear very little resemblance to their analog ancestors. A lot of the "resolution" is still technically detail correction, but it's detail correction several orders of magnitude more sophisticated than the original "one-size-fits-all" thick black lines of 40 years ago. So while Bayer masks have been around for 30 years, it's only been in the last few years that the support technology has become available to extract all the information they capture.


Some of the arguments expressed here about Film vs "Digital" resolution have been pretty daft. We seem to get the same scenario repeated over and over again: The assumption that all film-originated cinema releases are perhaps going to be processed in some dodgy back alley film processor in the slums of Jakarta, while all digitally captured images are going to be shown on brand-new Sony 4K projectors. I have little doubt that one day digital projectors will become the norm, but I rather think that a lot of what they are projecting will still have been originated on film!

The other thing that annoys me is the sometimes ludicrously low resolution figures quoted for the "average" cinema. "750 lines" was the best one I've heard...
  • 0

#27 Mike Miller

Mike Miller
  • Basic Members
  • PipPip
  • 25 posts
  • Digital Image Technician

Posted 14 December 2007 - 08:55 AM

Yes, a 4K Bayer is equal (with Graeme magic) to about 3K-3.2K.


No one else has a problem with our 4K resolution and it is not my job to "fly him out".


Can we then agree that this camera has 3K resolution and Jim will quit calling it 4K and Phil less than 3K until true tests are carried out.

Could Jim pay Phil's ticket to come witness a resolution test or could he arrange for such a test to be carried in England with Phil's presence? I trust Phil. Jim is too much of a salesman.

It is obvious this is a 3. not a 4K camera as Jim is promoting it everywhere. This misrepresentation is what pushes Phil's and everyone else's button.
  • 0

#28 Phil Rhodes

Phil Rhodes
  • Sustaining Members
  • 12239 posts
  • Other

Posted 14 December 2007 - 09:23 AM

> I invited Phil to come to RED and do anything he wanted to test the camera... he hasn't shown up.

You did?

When?

Amazingly I have no memory of this.

Phil
  • 0

#29 Matthew Rogers

Matthew Rogers
  • Basic Members
  • PipPipPip
  • 175 posts
  • Cinematographer
  • Knoxville, TN

Posted 14 December 2007 - 10:30 AM

Can we then agree that this camera has 3K resolution and Jim will quit calling it 4K and Phil less than 3K until true tests are carried out.

It is obvious this is a 3. not a 4K camera as Jim is promoting it everywhere. This misrepresentation is what pushes Phil's and everyone else's button.



The camera is 4K. It records 4k pixel size. It's that simple. True, it may not RESOLVE 4K, but it records that natively off the sensor. But if we are going to get picky then we need to ask Sony, Panasonic, JVC and many others to label their products with what they truly will resolve. I'm sure that my Panasonic DVX-100A won't resolve as much as a Panasonic SDC-615, but both won't resolve what they claim...

If you want Jim to be honest with you, then first ask the other manufactures who have lied to you for years to be honest with you.

Matthew
  • 0

#30 Mike Miller

Mike Miller
  • Basic Members
  • PipPip
  • 25 posts
  • Digital Image Technician

Posted 14 December 2007 - 10:37 AM

F900 delivers 1440x1080 to tape and 1920x1080 to HDSDI. Red delivers 3K. If it had (3) 4K CMOS sensors or one 6K sensor, it could deliver 4K. Why then call it 4K just because it carries 3K over 4K. Would you call HVX200 1080p camera when it delivers 720p over 1080p?

As (1) 4K CMOS can't deliver 4K but only 3K resolution, the camera is not 4K

Sony 4K projector is 4K; this camera is 3K.
  • 0

#31 Mitch Gross

Mitch Gross
  • Basic Members
  • PipPipPipPip
  • 2873 posts
  • Cinematographer

Posted 14 December 2007 - 10:52 AM

F900 delivers 1440x1080 to tape and 1920x1080 to HDSDI. Red delivers 3K. If it had (3) 4K CMOS sensors or one 6K sensor, it could deliver 4K.

Wrong.

That is the format spec. It doesn't mean that this is the true performance out of the camera. Format spec on RED is 4K. Stop mixing this up.

Stop quoting specs. Take an F900 or any other camera and measure the true resolving power and then tell me what it delivers. Otherwise do not make this claim.
  • 0

#32 Mike Miller

Mike Miller
  • Basic Members
  • PipPip
  • 25 posts
  • Digital Image Technician

Posted 14 December 2007 - 10:57 AM

You are wrong! You can get true 1080p out of three 1080p chips but not from one. That makes F950 a 1080p camera, Red 3K, and HVX200 720p performance camera.
  • 0

#33 Mitch Gross

Mitch Gross
  • Basic Members
  • PipPipPipPip
  • 2873 posts
  • Cinematographer

Posted 14 December 2007 - 11:08 AM

Mike, I don't know where you are located since you have not listed your location in your signature (as required by this forum). If you were in the New York area I would invite you down to the shop to show you how you are incorrect. In THEORY you can get 1080p out of three 1080P chips, but in practice it just doesn't happen. The 950 is old news; the new F23 has an improved prism block to help increase the resolution (yes, because passing through the prism block reduces resolution before light even reaches the sensors). The little HVX200 may have a format spec of 720p, but the chips are 590 with pixel shift & uprezzing. I can put the HVX200 next to a Varicam, which both record 720p, and you will clearly see a vast difference in resolution. I can put two 3-chip 1920x1080 cameras next to each other and show how one can vastly out-resolve the other, and how neigther of them actually reach the full resolution of the format. How about the cameras that can get full res. in some color frequencies but not others due to the coatings on the prism block and the color masking on the chips?

Sorry to be rough on this, but your conclusions are not based on solid technical data. At least not the complete picture.
  • 0

#34 Mike Miller

Mike Miller
  • Basic Members
  • PipPip
  • 25 posts
  • Digital Image Technician

Posted 14 December 2007 - 11:12 AM

Of course there is some loss in the optical path. I agree with that.
  • 0

#35 Matthew Rogers

Matthew Rogers
  • Basic Members
  • PipPipPip
  • 175 posts
  • Cinematographer
  • Knoxville, TN

Posted 14 December 2007 - 01:48 PM

That is the format spec. It doesn't mean that this is the true performance out of the camera. Format spec on RED is 4K. Stop mixing this up.

Stop quoting specs. Take an F900 or any other camera and measure the true resolving power and then tell me what it delivers. Otherwise do not make this claim.


This is exactly what I was getting at. I don't think any camera actually resolves the resolution that it claims to. Think about this...do any of the cars that claim certain gas mileage actually get that? I don't think so.

The RED is 4520x2540. It's as simple as that. While it may only resolve around 3500 pixles, that's still impressive...

Matthew
  • 0


Paralinx LLC

Gamma Ray Digital Inc

rebotnix Technologies

Media Blackout - Custom Cables and AKS

Technodolly

Willys Widgets

Visual Products

Metropolis Post

Rig Wheels Passport

FJS International, LLC

CineTape

New Pro Video - New and Used Equipment

Broadcast Solutions Inc

Wooden Camera

Abel Cine

Ritter Battery

Aerial Filmworks

CineLab

The Slider

Glidecam

Tai Audio

Broadcast Solutions Inc

Rig Wheels Passport

Aerial Filmworks

Wooden Camera

FJS International, LLC

Technodolly

Willys Widgets

Ritter Battery

CineTape

Abel Cine

Media Blackout - Custom Cables and AKS

Visual Products

Glidecam

Tai Audio

Gamma Ray Digital Inc

CineLab

rebotnix Technologies

The Slider

New Pro Video - New and Used Equipment

Paralinx LLC

Metropolis Post