New BetterLight back

[Greg Lockrey]

To Betterlight! (I like this guy. I talk to him at lenght every year at the VC conference)

Being realistic, this is what I need:

1)200MB scan

2)1 second or less scan times

3)ability to shoot untethered to a memory card or something

4)under $5,000

Give me that and it would be a no brainer.

Hell at $22K for the 416 mb I'm almost tempted to sell my wife's electric wheel chair and put her back into a manual one.

[Frank Petronio]

Whether it requires a 1 second exposure or a 100 second exposure hardly matters, you're going to need a rock solid rig to exploit the potential quality. For $20K I can get an entry level medium format digital back and a new Fujiblad kit, or a Canon Mk3 set for half of that. So don't look for any radical price drops -- it's a smaller market than you think.

[adrian tyler]

my lab bills run in at about 800$ per month, if this thing had the specs that kirk layed out i could justify 22k to the bank and half that monthly repaymet to 400$...

[adrian tyler]

problem with the medium format backs for me is using really wide angles and getting to use any decent amount of movement, the phase one p45 i've been playing with went haywire with my swc 38mm (see color variation centre outward) and i had to stop it down to f22 to get anything resembling sharpness in the corners. amazing in all other respects though...

[Henry Ambrose]

I'm pretty sure that its not getting the info off the chip, it is getting the movement of the array "fast and accurate". These things work kinda like a little bitty flatbed scanner in that something has to move the array over the image plane - little motors and leadscrews. The need for --extremely-- high precision should be obvious.

A couple of things about scanning backs that I am almost sure of:
They will likely always require tethered shooting.
They will never become "instant capture".

[Kirk Gittings]

Whether it requires a 1 second exposure or a 100 second exposure hardly matters

It matters enormously to me on architecture. I can make most daylight exteriors work at 1 second. I can make few of them work at 22 seconds.

[Gordon Moat]

my lab bills run in at about 800$ per month, if this thing had the specs that kirk layed out i could justify 22k to the bank and half that monthly repaymet to 400$...

Don't you bill out lab and film expenses to your clients?

[Kirk Gittings]

Gordon,

If Adrian follows my thinking. You have your clients expect to pay X amount in general when shooting film. Some or all of that film and processing costs become "capture fees" and helps cover some of the routine computer time and/or the costs of digital.

[Emmanuel BIGLER]

Does anyone here know what the specific technical challenges are in bringing down the scan times?

At least we can set a minimum scanning time corresponding to the time required to properly expose each pixel.
The numbers specified in the datasheet suggest a pixel pitch of about 9 to 10 microns, about 13,000 pixels on 120mm long and 10,000 on 96 mm wide. Most probably the actual aperture is smaller, say, 7 microns if there are no micro-lenses. But let's take 9-10 microns pitch.
If the sensitivity is set to ISO 400, and if we use the sunny-16 rule, each point of the sensor @f/16 should see the light for about 1/400 s. So this is the minimum time required for the stage to travel across the width of one pixel if we consider it like a focal plane shutter. In fact consider a very thin bright line that you want to properly detect on a dark background, but my feekling is that the model is like for a focal plane shutter. So for 13000 pixels we need 13000/400 ~= 32 seconds, minimum, not taking the data transfer time into account.

So except if the equivalent ISO sensitivity is increased by a factor 10, I do not see how we could scan really much faster than about 30 seconds.

And about the FTM specs, we should not forget that a pixel pitch of 9 microns is not capable to record any fine structure smaller than 18 microns as a period, we need 2 pixels per period, i.e. about 53 cycles per millimetre without aliasing effects. Except if we can fine-shift the array laterally by fractions of the pixel pitch.
And the figure of 60% contrast at 70 cycles per mm is not more impressive that the actual FTM of the blurring effect of an analog image by averaging the illumination across a square aperture of 7 microns width. If I cannot record beyond 53 cycles per millimiter due to aliasing effects, I wonder why I should be interested by what happens @ 70 cycles/mm ! Well if I'm lucky, there will be no high-contrast fine grid in the image beyond 50 cycles /mm and Nyquist & Shannon will be satisfied...

[Doug Dolde]

The new Seitz digital back has everything and more but at a price the equivalent (or hogher than) a top of the line Phase One back. They are even saying that HDR will be build into the camera.

Betterlight is old technology no matter how high they go on resolution.

http://www.roundshot.ch/xml_1/intern.../d925/f934.cfm