Skip navigation

Leibson's Law

Steve LeibsonLeibson's Law: It takes 10 years for any disruptive technology to become pervasive in the design community. This blog is about the disruptive technologies that either have or will win over electronic engineers, some that won't, and why. Written by Steve Leibson, Tensilica's Technology Evangelist. See my history site at www.hp9825.com. You can email me by taking the first letter of my first name, appending that to my last name, then the magic email symbol, followed by the name of the company I work for, and then a dot followed by com.

View Steve Leibson's profile on LinkedIn

Profile

RSS Feed

  • Add this blog to your RSS newsreader!

Recent Posts

Recent Comments

Most Commented On

Archives

By Category

Blog

Thursday, July 10, 2008

Want that Kodak 50Mpixel Sensor in a Camera? Hasselblad’s H3DII-50: $40K+

Jul 10 2008 5:13PM | Permalink | Email this | Comments (6) |
Blog This! using:  Blogger.com | LiveJournal |
Digg This | Slashdot This | add to Del.icio.us


Yesterday, I wrote about Kodak’s introduction of a 50Mpixel CCD image sensor for medium-format cameras. Today, I want to show you the first camera to use this sensor. It’s Hasselblad’s H3DII-50. One report gives the price for this machine at $40K. Another says that’s the price for the 39Mpixel H3DII-39 and that the -50 will run higher. In either case, that price is for the body. Hasselblad’s lenses by Carl Zeiss are extra. A beautiful camera. Not coming to my house any time soon.

 

 

The data sheet for this beauty says that a RAW image file consumes 65 Mbytes and that the maximum capture rate is 33 images/minute (1.1 seconds/image capture). Think about the system-design issues surrounding the need to move that much data from a sensor through the image-processing engine and out to a CompactFlash memory card in a battery-powered product.


Related entries in: Imaging | Multimedia | 

Reader Comments


at 7/11/2008 2:24:34 PM, stiggle said:
I wonder what ISO rating the sensor has. The larger format sensors need a lot of glass to to gather light and have a flat field of focus. Otherwise the high F-stop (small aperture size) reduces the amount of light and increases the exposure times. A low ISO rating further increases exposure times and the amount of light needed...

at 7/11/2008 5:13:27 PM, Steve Leibson said:
stiggle, I don't think ISO has much meaning until you hang an analog front end and A/D on the sensor. The Hassy H3DII-50 gets ISO 50-400 from the sensor. Not fast, but usable by pros who can afford lighting (and lenses) to go with a $40K camera body.

at 7/13/2008 11:38:56 PM, Dave J said:
I would have thought the sensor is the analog front end. The A/D and everything that comes after is the back end. :) There are some pretty good treatises on sensitivity and noise in image sensors on the net. It's worth a google for those interested, but I agree with Steve that ISO is not a way to describe a sensor. By the way, there is nothing astounding about the pixel density of this sensor. If it were a Canon APS-C sensor (22.2mm x 14.8mm) it would correspond to 9.5MP -- well under current norm for digital imaging. This new sensor is not a technological leap so much as it is evidence of a willingness of certain studio photogs to pay for an awful lot of scrapped dice.

at 7/14/2008 12:52:54 AM, Steve Leibson said:
DaveJ, the CCD pixel buckets dump into on-chip amps, but I'm pretty sure more amps follow off chip, if only to add a variable amount of gain. As for "nothing astounding" about this sensor, I can't agree. It's size is indeed astounding and probably sets a new bar for a big-sized die with non-zero yield. I think that you mean that the per-pixel dimensions aren't as small as the ones seen in commercial dSLR sensors, and there I agree with you. But the s/n ratio should therefore be better, which is another figure of merit that pro shooters are aware of and willing to pay for while the typical dSLR buyer looks for megapixels, megapixels, and....little more beyond brand, in my opinion.

at 7/14/2008 11:17:41 AM, Dave J said:
My point was that the per-pixel dimensions are about the same as what you're seeing in consumer-grade SLRs right now, and so the per-pixel S/N will also be about the same. However, as you point out, the S/N of the image overall will be higher with the MF sensor. A lens capturing the same angle of view in a MF camera compared to an APS-C camera is projecting the image over a much larger imaging sensor (and more pixels), so yes, the image S/N will be better.

at 7/14/2008 11:25:31 AM, Dave J said:
By the way, I'm perhaps less impressed by you about the size of the sensor. People have been making ridiculously large, low-yielding chips forever. There is always someone willing to pay. Also, yield isn't often as bad as you'd think. It was not uncommon when I was at Intel (0.6um to 0.35um era) to get 100% good dice from a wafer, especially if you had robust design rules and a decent speed range of part numbers to bin into. Also, imaging sensors don't need to be 100% perfect as do most chips. My Canon 400D has hot and dead pixels. Not a lot, but a few. The camera maps them out with data from surrounding pixels. I'm not sure what rate of dead pixels per area is deemed acceptable, but clearly Canon has some number in mind. Defects scale with area, so presumably, there is little reason for yield to decrease very much with area -- at least not from *individual*, scattered defects. What probably *does* decrease yield for large sensors is overall process variation over such large areas and large clusters/blobs of defects that knock out many nearby pixels at once.

Post a comment


Display Name

Before submitting this form, please type the characters displayed above:

ADVERTISEMENT

About Us   |   Contact Us   |   Free Email Newsletters   |   Free Magazine Subscription  |   Advertising   |   RSSRSS   

©1997-2008 Reed Business Information, a division of Reed Elsevier Inc. All rights reserved.
Use of this Web site is subject to its Terms of Use | Privacy Policy

Please visit these other Reed Business sites