how large will MF digital back sensors get?

[Ray]

[font color=\'#000000\']Who knows, eyefloater! It's like trying to predict the trends in the stock market. Engage in a bit of science fiction, then a billion pixels spread across an 8x10" sensor might eventually become a reality.[/font]

[howard smith]

[font color=\'#000000\']It is my understanding that at least some of the Marleboro Man bill boards were shot on 35mm film with a hand held camera Nikon.  (I wasn't there so this could be quite wrong.)  Would a MF have made a better bill board image?[/font]

[howard smith]

[font color=\'#000000\']20x30 is a large print.  Maybe the answer lies in how big film and photo paper got.  I don't think it was technology but demand that limited how big a sheet of photo film or paper Kodak could make.  As a first cut, I'd guess digital cameras will never get larger than 4x5.  Not much demand out there for even film cameras bigger than that.

Of course, it wasn't very long ago that people wondered what a digital camera was, not how big can Canon make one.  Maybe before they evolve to 4x5, there will be some new technology to take over.  Remember the head of the US Patent Office wanted to close it down in 18-something because everything had been invented.[/font]

[Gary Ferguson]

[font color=\'#000000\']In the UK the first "48 sheet" or billboard poster to be shot on 35mm was also for a cigarette brand, "You're never alone with a Strand", it was taken in the 1960's by Bert Hardy the chief photographer for Picture Post.

It's relevant only because obscure facts like these keep getting re-cycled as evidence that the small format can do big format work. It can't.

A poster is a special case that has almost no relation to the photography done by the majority of people on this forum. A poster is seen from a huge viewing distance, the first thing we all do when confronted with a photograph is scrutinise it from six inches away. And the sharper the photograph appears the more we're pulled into it in order to explore its contents. The problem with digital is that sharpness comes cheap, but resolution costs dear. Anyone can turn up the wick with unsharp masking and interpolate up, but it's raw pixels that finally deliver the detail.

Furthermore, there seems to be a photographic trend towards printing bigger. Maybe it's the influence of the Dusseldorf school, maybe it's the convenience and availability of wide carriage inkjets. But whatever the cause the all too frequent result is photographers wildly overstepping the limits of their equipment and their technique. I'm sure I'll be shot down in flames but IMO the sensible thing is to consider your RAW file as a 300dpi negative, and if you want to go bigger in the print then proceed with caution. Sometimes you'll get away with it, but oftentimes you won't. And just because someone once shot a billboard on 35mm doesn't mean you're guaranteed to get a satisfactory 40"x60" print from 6MP.[/font]

[dbarthel]

[font color=\'#000000\']The new full frame medium back sensors on something like a Contax 6x45 should really rock. When Michael can't stand it and buys one, I eagerly await the review, and his comment that he just sold the 4000 and got a 9600  [/font]

[Lin Evans]

[font color=\'#000000\']

I've done enough stitching to know what a headache it can be. There's also a huge disadvantage in effectively having to use a very long shutter speed. How long did it take Max Lyons to shoot his 196 frames?

It took 13 minutes for Max to shoot the 196 frames for the gigapixel stitch.

Here's a link for full information:

http://www.tawbaware.com/maxlyons/gigapixel.htm

Best regards,[/font]

[Ray]

[font color=\'#000000\']

I don't understand the 'long shutter speed' statement?  Would you please explain?

Certainly! Max Lyons took a high resolution photo consisting of 196 images taken over a period of 13 minutes. He chose his subject matter well - a rocky, desolate canyon. There's probably too much going on in the average scene for this technique to be viable. For example, there might be no wind where the camera man stands, but there might be a sudden breeze a hundred yards away causing a blurring of foliage in the stitching process. Clouds can move a lot in a 13 minute period, to say nothing of changing lighting conditions and movement of people and cars etc in the areas of overlap.

Some of these problems can be overcome, with a lot of patience, and some might go unnoticed - for example, the same car, person or train (whatever) featuring in adjacent frames. However, an array of precisely angled, relatively cheap cameras, or a supercamera built from the same components of a cheaper camera, say a dozen 300D sensors each with its own telephoto lens, all firing simultaneously, would avoid these problems.

Of course, the 13 minutes of Max Lyon's gigapixel image is not exactly analagous to a 13 minute exposure from a single very large format camera, but in general the larger the format, the slower the shutter speed for equivalent DoF and FoV. If you want to exceed the quality of 8x10 film by stitching a number of 35mm frames, you don't need to use a 300/f2.8L lens and stitch about 80 shots. That would restrict you to the same shutter speed, aperture and DoF as the 8x10 field camera and give you a resulting image of much, much higher resolution than the 8x10 piece of film with its single 300mm lens. A dozen or so shots from a cheap Sigma 100mm lens would be more than sufficient. (Or maybe 150mm lens - I haven't got my calculator out   )[/font]

[Edward]

[font color=\'#000000\']It does not take much breeze to make stitching difficult. I shot a multi-image pano in a graveyard on what I though was a still morning.  Not much moving, but in the prints the Spanish moss on trees was gently blowing.:-) Contributed a surreal note, but not one I wanted.[/font]

[eyefloater]

[font color=\'#000000\']i'm a recent photography graduate (sheridan college, just outsite of toronto) ... to start off my career i'll likely be doing most of my work with a dslr (decent output, inexpensive workflow).  in a "few" years, i hope to move up to shooting medium format + digital back.  

i've been thinking about physical sensor sizes ... what size will the industry produce them up to?  full frame for a 645?  how about full frame for a hasselblad, etc?  will anyone even bother going ahead with a back to fill 6x7s?  of course, there's always the argument that the possibility is there to pack the maximum number of megapixels into the largest sensor ... or just get the same amount into a smaller one.  maybe the industry will just get up to sensors big enough to fill a 645 and then let everyone else use masks over their ground glass ... or maybe they'll make specific models for the larger cameras.

so, thoughts?[/font]

[Bobtrips]

[font color=\'#000000\']I think it might be best to approach this question in terms of typical maximum print sizes.  What are people producing, size-wise - with MF film/backs that can't be produced with a 1Ds/14x type camera?  How many more pixels would it take to replace all MF film?

Past the point of "typical maximum print size" the air gets very thin indeed.  A camera that shoots larger might never be produced since the job could be done with a "gang" of high-rez digitals firing in unison.  

Just imagine a tripod frame that would hold 6 14xs, allowing them to move as one unit, focus and set exposure as one unit, fire as one unit....[/font]

[Bobtrips]

[font color=\'#000000\']Nikon posted a Times Square billboard shot with a two meg digital a few years back.  (Quite impressive from what I've heard.)  Billboards, as you know, do not push the limits of the medium.

I think the question is "How large do people print hi-rez shots?".  Images that will be viewed at somewhat close range.  

If 20"x30" is about as large as the market demands then there is not likely going to be a digital that will make high quality prints of a larger size.  The market just won't support the cost of development and tooling.  The occasional larger shot will be taken via stitching.[/font]

[Ray]

[font color=\'#000000\']Actually, Bob's idea of an array of smaller cameras working in unison would not only be more economical, but potentially would produce a much higher quality image than a single 8x10 or even 4x5 sensor. You would also get the benefit of greater DoF and faster shutter speeds. In fact I'm surprised this hasn't already been done.[/font]

[image66]

[font color=\'#000000\']I never did understand why the fixation with large-formats being necessary for billboards.  Given the viewing distance, 35mm (when shot properly) has done just fine.

Back in the "old days" tranny/neg size did matter some due to tonality loss with excessive enlargement. But with digital (scanned or DC), there is no discernable loss of tonalities during the enlargment process.

We've got to seperate the format from photographic technique. I've seen and shot many billboards that were done with 35mm which were as good as any medium/large format camera.  Typically it has come down to "control" during the photograph (movements/perspective control) or cropping options. During layout, there is a lot of cropping of the original that can occur and having a larger negative with greater margins helps. Since a typical product photograph will be used in many different layouts and for multiple purposes, it is desirable to shoot it in a larger format and give big margins.

The average billboard is viewed for 1/2 second, the exceptional one is something like two seconds. Given the "normal" viewing distance, that 40' billboard is much smaller than a 3R print held in your hand.  I can guarantee you that (when all else is equal) you cannot identify a billboard shot with any specific format at "normal" viewing distances.

To me, the biggest reason to shoot large-format for billboards is that it's easier to justify my Day Rate and per-shot charges.  It doesn't look good to charge $1000 per day, plus double all material costs, plus per shot, plus usage fees and then show up with a beater 35mm camera.

Ken N.[/font]

[akclimber]

[font color=\'#000000\']Actually, Bob's idea of an array of smaller cameras working in unison would not only be more economical, but potentially would produce a much higher quality image than a single 8x10 or even 4x5 sensor. You would also get the benefit of greater DoF and faster shutter speeds. In fact I'm surprised this hasn't already been done.[/font]

[font color=\'#000000\']Ray, some would argue that by using panotools and stitching together a number of images from a DSLR, this is defacto what you get   Max Lyon's has produced a 1 gig image this way by stitching 196, 6.3 meg images together.

The idea of using a sensor array is very appealing tho.

Cheers![/font]

[Ray]

[font color=\'#000000\']some would argue that by using panotools and stitching together a number of images from a DSLR, this is defacto what you get   Max Lyon's has produced a 1 gig image this way by stitching 196, 6.3 meg images together.

The idea of using a sensor array is very appealing tho.[/font]

[font color=\'#000000\']I've done enough stitching to know what a headache it can be. There's also a huge disadvantage in effectively having to use a very long shutter speed. How long did it take Max Lyons to shoot his 196 frames?

Supercomputers are built from lots of small CPU chips working in parallel. The same principle could be used to build a supercamera   .[/font]

[Bobtrips]

[font color=\'#000000\']I've done enough stitching to know what a headache it can be.

There's also a huge disadvantage in effectively having to use a very long shutter speed.

(Cropped and edited.)[/font]

[font color=\'#000000\']Are you aware of the tripods that move the camera through the 'stitching positions'?  One of those should make the job a breeze.

----

I don't understand the 'long shutter speed' statement?  Would you please explain?[/font]

[Bobtrips]

[font color=\'#000000\']OK, I follow you now.  Over time objects in a stitched photograph can certainly move.  That, obviously, limits the use of stitching.

Thing that threw me is that I misread "long shutter speed" as "long shutter speed".     [/font]