I guess we are going to see more MF CMOS introductions…

[Sareesh Sudhakaran]

From what I have seen from stills, my 24 MP full frame camera is a good match for my Pentax 67 using Velvia, so I feel 24x36 would do, but larger is mostly better.

It's not the same with motion. Two reasons:

• Motion blur and camera movements reduces resolution. The lower the frame rate the worse it gets. 60p would be just about okay.
• Interlacing reduces resolution.
• Chroma subsampling reduces resolution.
• Compression destroys resolution. Contrast that to fine art printing, where the goal is to preserve and even 'upgrade' resolution as much as possible. Video is the other way around.
• Poor display technology and viewing conditions reduces perceived (and real) resolution.

For these reasons and more, I imagine, to match 6x7, as I get with my Mamiya RB67, I'd need 40 MP of video, or 8K.

Same rules don't apply.

So you think 8K is around the corner?

Best regards
Erik

There is no corner anymore, Erik. NHK has already successfully broadcasted the first 8K video - which means they have the camera, the display and the compression technology to make it happen. By 2020, there will be 8K cameras covering the Olympics, in the hands of both professionals and fans. You may find this interesting: http://wolfcrow.com/blog/camera-trends-over-the-last-decade-what-does-the-future-hold/

The data rate for UHDTV2 or 8K 12-bit RAW is about 1.2 GB/s - easily achievable today with consumer drives in RAID 0. Thunderbolt technology is already ready for more than twice the data rate. All of this may sound too much, but it is worth it, IMO. Anyone who has seen a true IMAX documentary in a dome theater will know why. It's exhilarating as only motion can be - it makes 30' fine art prints look like post cards - with all due respect.

[ErikKaffehr]

Hi,

I know little about motion. I have seen some IMAX much more than cinema, actually.

It's nice to hear about 8K, but right now I see 39 MP from my P45+ on a 2MP projector. Si I am waiting for 4K projectors to be affordable. Well I ca buy 3 of them for an IQ 250, everything is relative.

I hope 4K goes down in price, I am sure it does.

Best regards
Erik

Ps. Thanks for the, I have not read it yet, but I will! Your writing is always interesting!

Pps. Case of beer posted ;-)

For these reasons and more, I imagine, to match 6x7, as I get with my Mamiya RB67, I'd need 40 MP of video, or 8K.

Same rules don't apply.

There is no corner anymore, Erik. NHK has already successfully broadcasted the first 8K video - which means they have the camera, the display and the compression technology to make it happen. By 2020, there will be 8K cameras covering the Olympics, in the hands of both professionals and fans. You may find this interesting: http://wolfcrow.com/blog/camera-trends-over-the-last-decade-what-does-the-future-hold/

The data rate for UHDTV2 or 8K 12-bit RAW is about 1.2 GB/s - easily achievable today with consumer drives in RAID 0. Thunderbolt technology is already ready for more than twice the data rate. All of this may sound too much, but it is worth it, IMO. Anyone who has seen a true IMAX documentary in a dome theater will know why. It's exhilarating as only motion can be - it makes 30' fine art prints look like post cards - with all due respect.

[eronald]

Doug,

 Dealers can do all you say. But some countries have dealers who have all the interactivity of an electric utility company. Yes they can make things happen for you. Well, they could.
 The problem with the dealer model is that a lot of it is territory, and once they have a territorial monopoly with the institutional contracts that go with it, they have zero incentive to provide more service than ... your local electric utility.

Edmund

Dealers also lower support and marketing costs at the manufacturer's end. Someone has to be paid to provide support; whether they sit at a desk with a Dealer logo above it or a Manufacturer logo above it is fairly irrelevant. This is especially true when you're looking to provide support on a variety of inter-related products which are not manufactured by the same company. A dealer can help you with your H4X, your P1 back, your profoto lights, your pocketwizard sync, your eizo monitor, and your capture one software - each manufacturer typically only supports their own part of the chain.

Also I'd like to think Value Added Dealers increase net sales, and greater volume means less R+D cost has to be born by each unit sold. They do this by providing expertise, the ability to demo the unit, the inventory of accessories (especially 3rd party items like tech cameras), the support they provide if/when there are issues (even if the issue isn't with the gear), and by providing rentals/replacement/loaner etc. There are countries where a particular brand is very dominant, largely because the dealer in that country is especially good/liked, even if that brand is not as popular in the broader world market.

[hjulenissen]

It's not the same with motion. Two reasons:

• Motion blur and camera movements reduces resolution. The lower the frame rate the worse it gets. 60p would be just about okay.

Excessive motion blur and camera movement would not be good arguments for increasing sensor resolution, though.

• Interlacing reduces resolution.

If we are talking about a hypothetical, top-notch video camera, why would we even think about doing interlacing?

• Chroma subsampling reduces resolution.

Yes, but for for natural images it tends to be acceptable (it is also used for JPEG still-image compression at moderate and low bitrates and kind-of in the Bayer CFA)

• Compression destroys resolution. Contrast that to fine art printing, where the goal is to preserve and even 'upgrade' resolution as much as possible. Video is the other way around.

Agreed.

• Poor display technology and viewing conditions reduces perceived (and real) resolution.

But why then should the recording be improved even further?

-h[/list]

[BJL]

... it shouldn't be difficult to offer a 30x45mm version of it…. Let's not forget that Leica's current sensor is oriented from a currently available MF sensor of other dimensions.

Maybe. In the past, it was reportedly far easier to take a basic CCD photosite design and use it in a variety of different sensor shapes, perhaps because a CCD has a very simple repetitive structure and read-out method. Active pixel CMOS sensors instead seem to need more custom designing for each different size and shape. One possible reason is the fact that active pixel CMOS sensors do read out by direct transfer from each photo-site to the edge of the sensor, and so increasing the sensor size changes the signal path lengths and timing.  So Leica's custom CCD size might be hard to replace by an active pixel CMOS sensor in a large custom size.

I do know that Dalsa have made custom sizes much larger than 56x56mm, but I don't know what the cost overhead is, it may be massive. Possibly the size could worsen noise characteristics too? Or otherwise break suitability as a photographic sensor.

It used to be that all the huge sensors for astronomy and such either (1) had huge pixel sizes, allowing fabrication on "large format" steppers not suitable for the pixel size needed in MF, or (2) were arrays of several smaller sensors butted together, leaving visible join lines that are tolerable in X-rays, but not for a MF camera. I am not sure what ht state of the art is, but:

1) Teledyne-Dalsa will make custom CCDs and CMOS sensors at up to "wafer scale" (up to 98x49mm, so only one sensor per 150mm diameter wafer), but pixel spacing is a huge 96 microns. Interestingly, this design was derived from an earlier design of exactly half the linear dimensions and half the pixel pitch, so the same pixel count. This suggests that pixel pitch is is limited by the angular resolution of the optical system used in fabrication, and limits sensors to the same maximum pixel count.
See http://www.teledynedalsa.com/public/corp/pdfs/papers/Very_Large_Area_CMOS_Active-Pixel_Sensors.pdf

2) Teledyne-Dalsa also makes the 94 x 78mm, 250MP CCD sensor for the Z/I DMC IIe 250 aerial mapping camera with a more MF-friendly 5.6 micron pixel pitch. However, I do not see a price quoted anywhere.
ADDED LATER: this camera weighs 66Kg and needs a 350W power supply! http://www.aerial-survey-base.com/wp-content/uploads/2012/05/ZI_DMC250_DS_en3.pdf

[ErikKaffehr]

Hi,

Think IMAX, think 4K and think 8K, perhaps a reasonable. It seems that a lot of films are made on IMAX. Sony is a dominant player in the cinema business.

Best regards
Erik

yeah! …a cinema camera that would cost 200k, one would have to throw all his lenses away, have only a batch of maybe 10 dedicated lenses, develop "new" style of direction for (almost) NO DOF when one "moves" inside the frame, eliminate camera operators to maybe 4 people worldwide and use 50mp to take …8 (!) eight out of it, to use in 4k raw video… S35 (and APS-C for starters) is more than enough for video… sensor size increase is the last thing that video will ever need.

[EricWHiss]

If it it's quite easy to make a basic digital back thanks to devkits, which it might be, I find it sort of surprising that 1) DHW does not make a 56x56 Hy6 back, 2) Alpa/Arca/Cambo/Linhof does not make a tech cam digital back to go with their own cameras.

I think you'll see huge mega corporations create a bevy of product lines to fill a vertical, but small companies typically stick to their area of expertise and try and do it as well as they can.

That said, I would love to see them or anyone build a 56x56mm back - but as pointed out this can only happen if there are sensors made for it.   There are some big square sensors but I think mostly even bigger approx.  8cm on a side and for astronomy and space telescopes.     

[Sareesh Sudhakaran]

Hi,

I know little about motion. I have seen some IMAX much more than cinema, actually.

It's nice to hear about 8K, but right now I see 39 MP from my P45+ on a 2MP projector. Si I am waiting for 4K projectors to be affordable. Well I ca buy 3 of them for an IQ 250, everything is relative.

I hope 4K goes down in price, I am sure it does.

Best regards
Erik

Ps. Thanks for the, I have not read it yet, but I will! Your writing is always interesting!

Pps. Case of beer posted ;-)

Thank you humbly for the beer and the kind words!

[Sareesh Sudhakaran]

Excessive motion blur and camera movement would not be good arguments for increasing sensor resolution, though.

You used the word "excessive", I didn't ;-). It could well be that a 1080p image at 60p has an equivalent perceptible resolution to 4K at 24p. The strange quirks of video.

If we are talking about a hypothetical, top-notch video camera, why would we even think about doing interlacing?

Thankfully we don't.

Yes, but for for natural images it tends to be acceptable

Unfortunately it is always combined with compression, sampling and a color gamut transformation - which is why the effects are not visible. If you take a TIFF image and compress to 4:4:4 vs 4:2:2, all other things being equal, the difference is visually noticeable. Therefore, it is a non-negligible factor.

But why then should the recording be improved even further?

For its own sake?

[hjulenissen]

Unfortunately it is always combined with compression, sampling and a color gamut transformation - which is why the effects are not visible. If you take a TIFF image and compress to 4:4:4 vs 4:2:2, all other things being equal, the difference is visually noticeable. Therefore, it is a non-negligible factor.

I believe that for natural images at sensible resolution/distance and contrast, sRGB directly _tends_ to be visually indistinguishable to sRGB that has been converted to YCbCr 4:2:x and back.

For non-natural images (and some corner-case natural images) the difference may be clearly visible.

For its own sake?

"Quality improvements" that does not lead to improvements for the end-viewer is not a worthwhile goal in my opinion.

-h