Is Betterlight Still the Standard?

[8x10 user]

Also it should be noted that when Sinar talks about megapixels they mean the total # of pixels in the final image. Betterlight uses the total amount of pixels multiplied by 3 for the Red Green and Blue channels. This is to differentiate between their pure RGB pixels and the ones created with single shot sensors that interpolate data from a bayer grid.

So a 384 "megapixel" betterlight back actually captures 128 megapixels. This is in "enhanced mode" where the scan is over sampled in one direction; scanning the space in between pixels. In normal mode the Super 8k HS is actually an 85 megapixel scan back.

Sinar 16-shot mode also oversamples only in two directions. That is how a 50 megapixel sensor is able to produce a 200 megapixel image with pure RGB data for each pixel site.

[Daniel Stone]

If only there was a Foveon-style digital sensor in either FF 645, 6x7 or 4x5 sizes, imagine the possibilities!

TBH, I'd love to see a slip-in(like a traditional film holder) digital "magazine" that could be used like a traditional film holder. It would have shallow pixel wells(better for movements w/o "needing" the required LCC calibration shots like current MFD backs do when using off-axis movements like rise and/or shift, especially with wide angle lenses. There's a fellow that's designing/perfecting a single-shot digital back for 8x10(then 4x5 later, both color and b/w IIRC) cameras:
http://www.largesense.com/

I'm not aware of his pricing schedule, or other details like RAW file support, technical support, software, etc., but it looks promising to say the least!
I enjoy using film for MY photography, and generally work with exposures in the 1/125-30s range. So rather broad, and I'd need a capable "brush" to be able to do said photography unencumbered with the needs of a computer/tablet with all of the cables, batteries, etc necessary for its operation. A wireless tethering solution to a lightweight tablet via a proprietary wi-fi connection(ala Phase One's tethering solution, where RAW files are written to a CF card, jpeg files are transmitted via the wireless connection to the tablet/computer). To me, this would be a great solution, and if it could have similar dynamic range and grain structure to current color neg and b/w emulsions(Portra, Ektar and Tmax are my personal favorites), I really do think it could be an instant hit with photographers of all price ranges. Of course different resolutions and expanded capabilities could vary with different models, but being able to access hi-fidelity single-shot capture, like is capable with film(vs scanning backs, or multi-shot/stitching workflows) would really open up new territories that current tech cannot replicate unless more time is dedicated to such.

Do I see this actually happening? Maybe, maybe not. I don't know the hardware, but I do know, and fully understand that if there's enough money laid out on the table and there is a manufacturer capable(let alone willing) to design, fabricate and support/develop a chip(or series of sensors) capable of developing such tech, albeit economically, I think there are enough people willing to try. Money talks, bullsh** walks as they say. For those of us who like FLEXIBLE tools in our arsenal, I'd like to see this happen, as I'm sure many others would.

A digital 4x5 full-frame chip, 50-80mp of single-shot capture, 16bit information... Man, I'd be on cloud 9. If it was affordable that is...

-Dan

[8x10 user]

Well there are certainly issues of diminishing returns with medium format digital. The higher resolution backs require a larger aperture -due to diffraction, which limits depth of field. Many MFD that I have spoken to have downgraded too lower resolution / more affordable backs because of the need of more depth of field. We already have plenty of lenses that are diffraction limited at F/16-F/22 we just need the sensors to utilize a larger image circle.

I use a 22mp MFD back for my digital needs. When doing close up shots I often shoot at F/16 due to DOF issues. A higher resolution back or a "sharper" "HR" lens would be no use to me in a practical sense. Newerbacks have better high ISO performance but at iso 25-50 my back is still "king".

[Adamphotoman]

If I want to stitch an art work I can stitch 20 Super 6K HS 216 megabyte files together with 30% overlap to create a file to print massive prints. i am talking 20-30 foot long prints.
If I use a pano adapter I can make a 180 degree pano in under a minute. Shooting with MF lenses I shoot at 1/400 sec line time and I can capture 100 sailboats in motion. Try that stitching. Boats JUST DON'T stay still!
This creates a 1 1/2 gigabyte file without stitching. I can print 20 feet. A wide angle phase back or a multishot back would be at a disadvantage here. Good luck with that.

In the end it is still all about light.

One should be able to see the glow and shimmer from metallics and at the same time peer into the depths of dark shinny oils. One should have the sense of smelling the paint and have the feeling that the Fine Art Repro will get under your nails if you touch it. This can only come from the operator.
& Not the Equipment.

[8x10 user]

Well of course technique comes into play when reproducing artwork. Even the big automated machines from Cruse and Metis take this into account; adjusting the angle/proportion of the light to bring out physical details or adding polarizers to remove glare. Having everything well aligned and centered is also important.

Stitching can be done with many capture methods. There are ways to do stitching in the image circle (quad stitch) with medium format backs, and of course stitching can be done by pivoting the camera around its nodal point, or by moving the subject or the camera around (keeping the distances the same, a better method). I actually did a multi gigapixel reproduction for a family member a few years back by using a Nikon D2x with a VR macro lens and a foba camera stand to move the camera around the subject while keeping the distance the same. The final image quality was excellent.

But we were talking about the integrity of the capture method. A multishot system that captures each pixel in red green and blue and has the same or perhaps higher integrity per pixel then a trilinear sensor based scan back depending on the iso. Again the CCD technology in a 8 year old Sinar multishot back is identical to the CCD technology used in the betterlight system... Both use the same era Kodak CCD's. The betterlight does not have a peltier cooler and is limited to isos of 100 or 125 and up. Per pixel the 25 iso sensor is going to have less noise and more integrity. The newer Dalsa based multishot backs will have lower noise at higher iso's. There are other factors including image size. I would think that a 200mp Sinar back would require a sharper lens to achieve its resolution capabilities. The newer lenses should be able to keep up but will not cover the betterlight format.

A betterlight super 6k HS has an actual native resolution of 48 megapixels, the sinar exact can produce 4 times as many pixels in less then a minute.

Sinars software is not as good for everyday purposes as Phase one's. Phase makes an excellent RAW convertor and they it be the best at interpolating data from a bayer grid. For repo purposes Sinar's software is very good. The back can be profiled in the software with a 200 + patch target. Each sensor is factory calibrated and adjusted. The user is then able to perform three types of calibrations in the software. One is done with a white translucent sheet that is placed in front of the lens. This calibrates the sensor, any movements that are applied in the camera, and the lens to the exact aperture/focus location that is being used in the shot. A second form of calibration includes scene calibration which levels the entire scene automatically removing any hot spots in the lighting. A third form of calibration is to make sure the piezos are fully calibrated. The last time I calibrated my back for 16-shot use, the software told me the results were accurate to ~ 200 nano meters which is smaller than the wave length of most visible light. I have not used a betterlight back but I would think their software is also very good for reproduction purposes.

Now that Sony is making MFD CMOS sensors, perhaps Mike can have them make a large CMOS tri linear sensor to produce a new line of scan backs with lower scan times and less noise. The CCD technology in the betterlight back is 20 years old. Not to say that it is low quality or lacks integrity but there are other high integrity methods out there.

[8x10 user]

Another notable approach is the one taken by Mega Vision. They use a black and white sensor that creates a color image by using a color wheel or colored lighting. It can capture work with more colors then RGB. Of course we can only see in RBG so multi spectrum imaging is not necessary for reproducing artwork for the human eye.

[Adamphotoman]

All I know is that I make a very good living using the BL.
Most of my old and all of my new clients will not accept anything but a BL scan.

My Phase experience and the Multishot all sit in a closet. [I should sell them] It is the real world after all.

Maybe in another 10 years the single shots will be good enough.

BTW
The Japanese accepted Mega Pixel Count does not make a Super 6 K a 48 Mega Pixel Instrument. Maybe in Monochrome.

If you can interpolate a pixel shift in a Multishot to add up to more Megapixel,
then you should be able to add all the trilinear pixels up.

[8x10 user]

Subsampling / over sampling is not the same as interpolation. I learned about the benefits of oversampling first hand when operating drum scanners.

Betterlight uses over/sub sampling in one direction and then interpolates in the other to create their enhanced resolution option.

Multishot systems capture each pixel location in red green and blue, its pure data without interpolating just like the Betterlight. So by Betterlight MP ratings the Sinar exact would be 576 megapixels.

I dont know much about it but Sinar has a new system which combines their multishot technology with color filters for an even greater level of color accuracy.
http://www.sinar.ch/en/category/products/ctm/

[Adamphotoman]

The Sinar back is impressive especially for the smaller 36X48 sensor size, however this means a smaller micron size for each receptor. This also means one needs lenses with higher resolution. The smaller well also impacts colour accuracy. The native sensor travel of the Betterlight is 72X96mm or 4X larger capture area. The Super 6K has a 12 micron photo site while the higher res Super 8K has a 9 micron photo site. Pros and Cons. The Sinar Exact sensor has even smaller photo sites. Meaning diffraction limitations which impacts depth of field.
Each system has strengths and weaknesses. The Sinar Exact is $30K versus $15K

I would say the Betterlight is still my first choice with Fine Art Repro. For other types of photography such as installation product and moving subjects the MF solution is the way to go.

Most of the time I don't need more megapixels just better ones. When I do need more megapixels I can stitch using a moving wall or I can move the art using the pano adapter.

In the near future the MF will not only catch up to Betterlight's image quality but will surpass it. The Betterlight still has some years left.

[Adamphotoman]

Back to the original post:
"I want to scan artwork in the most professional way possible. After some research it seems that this old Betterlight Scanback and a 4x5 camera is still 'state of the art'.
Is this true? Are there alternatives? Also what about macro lenses on a Canon 5d? Are the results much worse?

The main question...
If you had no equipment and wanted to start scanning artwork what equipment would you purchase?"

What is your budget?

Betterlights can be purchased for about 1/2 to 2/3rds the orig price...

Still the best for now but what do you really want to do with he images?