Future Medium Format Camera

[Stefan.Steib]

OK - what are your dreams of a High resolution (MF?) camera in lets say 2-3 Years from now ?
Does it need to have an electronic viewfinder, Live View and deriving from that full video recording ?
Should it be integrated (like an Leica S2) or modular (like a Blad) or even a tech cam ?
Should there be wlan, remote control, internet access, GPS, more of these technical features or
would you center on water and  dust protection, enhanced battery life, better usability overall ?

What would be a price range that is realistically possible for those features ?
And could a RED with more still resolution than the Scarlett X fulfill these needs ?

Please don´t hold it back, may it sound phantastic or crazy, I think ideas need brainstorming and input, speak it out.
If the industry reads this, they may also have some more motivation to realize it. Speaking about it does make things happen,
think about the Startrek communicator and Motorola who brought this to life with the first mobile phones.

And don´t we all want to have a future for this market segment so we can use these highend cameras
in the future ? What should be done to ensure this market will thrive ?

Interested to hear your opinions

Greetings from Munich
Stefan

[EricWHiss]

Did you get inspired by the earlier thread about digital Mamiya 7II?  That might fit the bill....

Really, I'd like a lighter, smaller camera, with good live view and multi point auto focus, another couple stops in ISO.

[Stefan.Steib]

Eric

no actually this is a question that was already of interest for me for a while. And the Mamiya 7 could  only be a first start for a format, none of the existing lenses do have autofocus and especially the wideangle is too short (flange focal range).

Here a link with some food for further thoughts - by Leica:

http://www.mirrorlessrumors.com/leica-says-optical-finders-will-get-less-and-less-important/

greetings from Munich
Stefan

[BernardLanguillier]

My perfect MF camera would be:

Imager
- The sensor would be no larger than 48x36mm with a resolution of no more than 60MP and a strong focus on DR,
- Able to deal with 4 min exposure without unreasonable noise build up,
- Feature a weak AA filter for optimally controlled aliasing,
- Sensor based VR/IS,

Focus
- Have live view capability for focus accuracy with sensor based AF accross the full frame (a la Nikon 1 series technology),
- Have access through SDK to the absolute focus distance of its lens (Get and Put),

Exposure/ISO
- Its exposure system would have to be accurate and optimized for ETTR with raw histogram and zero artificial highlight headroom,

Form factor
- Be mirror less for size and noise,
- The body would not weight more than 1kg with APS DSLR like bulk,
- Be fully waterproof (body and lenses),
- Have 2 tripod mounts for horizontal and vertical usage (see Pentax 645D),

Openess
- Be natively wifi and 4G enabled + Thunderbolt/10 GBe network port for thethering,
- Have an open OS (Android?) enabling the installation of applications that would be scriptable and workflowable,
- Relay on DNG raw format

Cheers,
Bernard

[Kagetsu]

Personally I'd just like to see an advanced AF with multi-points on the bodies

[Sareesh Sudhakaran]

OK - what are your dreams of a High resolution (MF?) camera in lets say 2-3 Years from now ?
Does it need to have an electronic viewfinder, Live View and deriving from that full video recording ?

I'm more from a filmmaking background, so:

A 40MP camera (RAW DNG) with 2K (or 3K max) resolution video at 4:2:2 uncompressed via HDMI or HD-SDI (A cross between a 645D and an Arri Alexa). Frame rates from 0-120fps. Live view as on Canon DSLRs is good enough. Is an optical viewfinder possible? DR of 12 stops (or more if possible). In-built high-quality variable ND filter. Strong Mount for telephoto lenses - Nikon/Canon/PL - interchangeable.

Should it be integrated (like an Leica S2) or modular (like a Blad) or even a tech cam ?

I love the RB67 camera philosophy. Why not have a revolving back for verticals? Everything else as solid as a 1DX. Can a semi-modular camera be sealed and weather proof?

Should there be wlan, remote control, internet access, GPS, more of these technical features or
would you center on water and  dust protection, enhanced battery life, better usability overall ?

If one can pile it on, why not? But as along as it takes great pictures and outputs uncompressed video, it would be a dream camera. A battery should run for 4 hours in video mode.

What would be a price range that is realistically possible for those features ?
And could a RED with more still resolution than the Scarlett X fulfill these needs ?

The Red Scarlet is nowhere near. The Epic is, to a certain extent. The Red is not really a simple machine, and its post workflow is not for the fainthearted. What would really blow my mind is a camera that outputs image sequences in DPX/TIFF/EXR/DNGRAW instead of relying on crappy wrappers and codecs.

I'm not a big fan of adding XLR inputs, but if we can, then why not? Price for such a beast? $20K or less.

[Stefan.Steib]

Here some more links for inspiration:

http://www.readwriteweb.com/archives/future_of_the_camera.php
http://www.fastcodesign.com/1663674/wvil-a-glimpse-at-the-future-of-photography-after-cameras-die-video

http://gizmodo.com/5580300/canons-camera-of-the-future

http://www.geeky-gadgets.com/canon-to-use-intels-thunderbolt-in-future-cameras-11-03-2011/

http://www.geek.com/articles/gadgets/lytro-light-field-camera-refocuses-the-future-of-photography-20110622/

http://news.bbc.co.uk/2/hi/programmes/click_online/9585903.stm

http://www.idownloadblog.com/2011/06/02/iphone-infrared-camera/

There´s plenty more, the designers of smaller camera systems are exploding with creativity.
So what do you think will make sense for you personally ?

Greetings from Munich
Stefan

[MrSmith]

a slightly smaller sensor size so none of your lenses are quite wide enough.
at least 150 megapixels 200 if possible
good at 50iso and with speeds up to 3200 but nothing usable above 200
live view but with a really annoying delay (and on a small lo-res screen)
a Bentley edition with green Connolly leather and milled aluminium accents

*i'm working on the theory that MF don't listen to their customers and do the opposite of what they want.

[Craig Lamson]

+1, I would love to shoot square again....

[Gigi]

Rather than features, can one look at the form?

Why do all the electronics have to be on the back of the sensor - can they be located elsewhere? If so, then the form factor of the back could be modified and  get back to a narrower camera body, opening up the door to a Mamiya 6 kind of simpler size. Something simpler, smaller, more usable.

The Leica S2 is wonderful for its usability, but unfortunately with the 3:2 profile (and very pricey) - a 6x6 sensor would be ideal.

Either fixed lens (50mm) or open mounts for either Hassy, Rollei, or some other large group of lenses.

[bjanes]

Really, I'd like a lighter, smaller camera, with good live view and multi point auto focus, another couple stops in ISO.

A couple of f/stops increase in ISO would be difficult to obtain, since current CCD sensors have a quantum efficiency of 25 to 50% (see Kodak spec for KAF 40000 and KAF 50100). CMOS generally has lower QE because the electronics take up part of the pixel area, although improved micro-lenses can make up for some of this. Sensor ISO is determined by the exposure (measured in lux seconds) required to saturate the sensor. A higher ISO is obtained by a higher quantum efficiency so that, for a given luminous flux, less integration time will be needed to collect the required number of photons. Current front illuminated sensors can achieve a QE of 85%. Doubling of the QE would produce 1 f/stop higher ISO.

Interestingly, if the full well capacity of the sensor could be increased and photon SNR thus increased without altering the pixel size, the ISO would actually decrease, since more time would be needed to fill the deeper well. However, charge density places an upper limit on full well for a given pixel area.

Regards,

Bill

[Stefan.Steib]

Bill

it seems that in scientific imaging devices sCMOS chips have already surpassed CCD´s in about every aspect see

 http://www.andor.com/scientific_cameras/neo_scmos_camera/

they use several technics to improve signal to noise, readout speed and even give the user a choice of rolling or global readout.
I think this can also be done on larger resolutions, maybe not necessarily with 1 electrons RMS performance but for normal imaging
with significant cost reduction against such Hightech solutions.

Greetings from Munich
Stefan

[ErikKaffehr]

Hi,

Here is what I'd like to see:

- Modular solution, essentially an MFDB with a high resolution display
- CMOS based design with live view, magnification actual pixels or more
- Focus peaking
- Zebra markings/blinking highlights based on raw data
- Ethernet and WiFi built in
- Electronic connection between back, body and lens, so no cabling is needed and exposure data is recorded
- About 4 micron sensor pitch so aliasing is minimized

Camera body can be any of:

- MF SLR (like Hasselblad, Mamiya, Phase One)
- Technical camera body like Sinar Arctech, Alpa or Hartblei

Many areas of photography are well covered with DSLRs. I don't necessarily feel that an MF DSLR is really needed. The SLR concept may create more problems than it solves. The reason SLR viewing was created was to achieve "what you see is what you get" viewing, the very same can be achieved with live view, without moving parts.

Best regards
Erik

OK - what are your dreams of a High resolution (MF?) camera in lets say 2-3 Years from now ?
Does it need to have an electronic viewfinder, Live View and deriving from that full video recording ?
Should it be integrated (like an Leica S2) or modular (like a Blad) or even a tech cam ?
Should there be wlan, remote control, internet access, GPS, more of these technical features or
would you center on water and  dust protection, enhanced battery life, better usability overall ?

What would be a price range that is realistically possible for those features ?
And could a RED with more still resolution than the Scarlett X fulfill these needs ?

Please don´t hold it back, may it sound phantastic or crazy, I think ideas need brainstorming and input, speak it out.
If the industry reads this, they may also have some more motivation to realize it. Speaking about it does make things happen,
think about the Startrek communicator and Motorola who brought this to life with the first mobile phones.

And don´t we all want to have a future for this market segment so we can use these highend cameras
in the future ? What should be done to ensure this market will thrive ?

Interested to hear your opinions

Greetings from Munich
Stefan

[aknicholas]

My interest would be to see image quality available in entry-level systems. I think that ease-of-use issues are more important only for high-end buyers. Back when I shot 6x7 film I was willing to put up with the lack of a motor drive or autofocus. For some, it seems, bringing DSLR features to MF are important -- so maybe a hybrid category may emerge (although Mamiya ZD didn't seem to take off...)

Aaron

[Stefan.Steib]

Hi Eric

good point for the Zebra stripes. I also think there are a lot of features of the video cameras that should be taken to the future "stills/video" devices.
Whereas there is probably more coming from the video side as these devices should be real 4k video devices also to be of dual use.
I also think that the idea of Canon´s 300C with the downsampled 4k->2k color scheme is absolutely ingenious. Phase already does pixel binning
why not channel binning on choice for smaller resolutions or BW with much faster readout ? This combined with a CMOS seems to be the way to go.

Greetings from Munich
Stefan

[Erick Boileau]

Hasselblad  503 digital  FF 56x56mm with live view and 60 mp   AF  and  new AF lenses   , 25 to 800 ISO , exposure up to  6 minutes, a TS module
15000 euro TTC   for the body + 80 mm

[ondebanks]

A couple of f/stops increase in ISO would be difficult to obtain, since current CCD sensors have a quantum efficiency of 25 to 50% (see Kodak spec for KAF 40000 and KAF 50100). CMOS generally has lower QE because the electronics take up part of the pixel area, although improved micro-lenses can make up for some of this. Sensor ISO is determined by the exposure (measured in lux seconds) required to saturate the sensor. A higher ISO is obtained by a higher quantum efficiency so that, for a given luminous flux, less integration time will be needed to collect the required number of photons. Current front illuminated sensors can achieve a QE of 85%. Doubling of the QE would produce 1 f/stop higher ISO.

Interestingly, if the full well capacity of the sensor could be increased and photon SNR thus increased without altering the pixel size, the ISO would actually decrease, since more time would be needed to fill the deeper well. However, charge density places an upper limit on full well for a given pixel area.

Regards,

Bill

Bill, this is all correct, but you're only looking at the signal side of the S/N equation...and ISO performance these days is primarily determined by the noise side of the equation. When people ask for "A couple of f/stops increase in ISO" they don't mean base ISO; they mean they want to get better higher-than-base ISOs; they want the S/N of current ISO 400 but at ISO 1600, the S/N of current ISO 800 but at ISO 3200, etc.

The reason why CMOS sensors in current DSLRs reach such giddy heights of ISO range is that they have got the readout noise down to around 2 electrons. MF CCDs are still 6 times or more higher than this. So although both types of sensors take in about the same signal flux per pixel, the CCDs can't compete on S/N per pixel in the shadows.

Ray

[BJL]

Above all, I would like to see a state of the art CMOS sensor, which would help with:

- Live View, with a modular system allowing the LCD or EVF to be moved relative to the lens/sensor unit. Shutter release and other controls on the "mobile" VF unit, maybe?

- Fast, accurate sensor-based focus confirmation, with VF feedback indicating what parts of the image are currently in focus, since at the high modern DMF resolution levels, checking critical focus visually otherwise requires zooming,  and is impossible with an optical reflex VF.

- A rotating back, which becomes easier with a mirrorless, live view system: no need for a huge mirror as big in each direction as the long dimension of the frame.
The RB approach of Mamiya was the beginning of the end for square format, and this could finish the job (except for that tiny faction that likes square images.) Own an expensive DMF system and yet can't even make the most rudimentary compositional choice between vertical or horizontal composition, or have an equally indecisive or unpredictable AD? Then take two shots, one each way up: storage is cheap.

[ErikKaffehr]

Hi,

The electronics are just a couple of micrometers thick, much thinner than a piece of paper. With CMOS they must be co located with the pixels, that is causing some disadvantage for CMOS, because some silicon area is used for the electronics. The physical thickness of the sensor is not affected by electronics.

Best regards
Erik

Rather than features, can one look at the form?

Why do all the electronics have to be on the back of the sensor - can they be located elsewhere? If so, then the form factor of the back could be modified and  get back to a narrower camera body, opening up the door to a Mamiya 6 kind of simpler size. Something simpler, smaller, more usable.

The Leica S2 is wonderful for its usability, but unfortunately with the 3:2 profile (and very pricey) - a 6x6 sensor would be ideal.

Either fixed lens (50mm) or open mounts for either Hassy, Rollei, or some other large group of lenses.

[ErikKaffehr]

Hi Stefan,

I see a single problem with CMOS, and that is that there is no CMOS sensor available for MF. On the other hand I'd presume that there is no great obstacle to making larger sensors, except cost.

Regarding the value of pixel binning I'm a bit sceptical. It is quite true that DR can be increased by pixel binning. If we bin four pixels physically the well capacity will quadruple, but read noise will be constant. But if we disregard read noise, and only look at shot noise binning does not really change anything. When we resize the image for final use we do binning in software.

This is very neatly demostrated by DxO-mark. The first screen dump compares Nikon D3X and Phase One P65+ regarding DR. The definition of DR that DxO uses essentially means that it is limited by read noise. The binning shows up as a significant increase of DR at high ISO.

The second screen dump is similar but shows "tonal range", tonal range in DxO definition is similar to DR but uses a requirement of a good signal quality like SNR = 8 (I'm not sure what value they use). With this definition the advantage of binning disappears, as shown in the second screen dump.

Binning may make more sense on present day CCDs which have relatively high readout noise, but may be less needed with modern CMOS which has much lower readout noise.

Best regards
Erik

Hi Eric

good point for the Zebra stripes. I also think there are a lot of features of the video cameras that should be taken to the future "stills/video" devices.
Whereas there is probably more coming from the video side as these devices should be real 4k video devices also to be of dual use.
I also think that the idea of Canon´s 300C with the downsampled 4k->2k color scheme is absolutely ingenious. Phase already does pixel binning
why not channel binning on choice for smaller resolutions or BW with much faster readout ? This combined with a CMOS seems to be the way to go.

Greetings from Munich
Stefan