MF CMOS -- now it gets interesting

[Vladimirovich]

You always have interesting insights into Pentax Radu.  Can you explain #5 in your post please?  No grip to protect?

may be he was talking about mount (camera/back mount) ?

[BJL]

"Never" is not a word I don't use often but I think the chances that Pentax will make a digital back ever are close to 0%.

I agree, and propose one more argument for your list: with Pentax's 645 system, the cost of a 645D is dominated by the cost of what would go into the digital back, as opposed to the part that would go into separate "front" (body), so the cost savings in being able to upgrade just a Pentax digital back rather than buying a complete new Pentax 645D camera would be modest, and would for most customers be outweighed by the advantages and the integrated design.

Note that despite endless forum pots explaining what a modular system with interchangeable sensors in 35mm format would be a good idea, no SLR maker has done that, and in fact SLR modularity has decreased over the years, such as with the loss of options like interchangeable viewfinders and focusing screens. And even with MF film, Pentax based its lower cost, lower size market position around an integrated design.

[Theodoros]

Now we know… this is a "basic" MF product, to attract DSLR users to MF…

http://www.youtube.com/watch?v=IvEEB7GYbY0

Obviously, Hasselblad will price it lower than an H5D-40, (I suspect that Leaf will discontinue the 5ii  because of it) and we'll all live happily ever after…  So much for the "Cmos revolution" …now we're talking!

[LKaven]

Now we know… this is a "basic" MF product, to attract DSLR users to MF…

http://www.youtube.com/watch?v=IvEEB7GYbY0

The video mentions "14 stops of dynamic range at all ISO settings."

I noticed this is something that Arri also says about the Alexa.  You always get 14 stops.

That's a weird way of speaking.

I understand that the sensor takes in 14 stops (or so) regardless of what the gain setting is.  So the question is what would they do with that dynamic range during the gain and A-D steps?

[ErikKaffehr]

Hi,

It is a marketing truth, know in common speak as BS. 

Best regards
Erik

The video mentions "14 stops of dynamic range at all ISO settings."

I noticed this is something that Arri also says about the Alexa.  You always get 14 stops.

That's a weird way of speaking.

I understand that the sensor takes in 14 stops (or so) regardless of what the gain setting is.  So the question is what would they do with that dynamic range during the gain and A-D steps?

[Radu Arama]

Hi Tom and thank you!

What I meant is that there is no separate (battery) grip as an accessory so Pentax can modify any dimension of the future 645D2 without worrying about backwards compatibility. Anyway mid February the CP+ show begins and we will find out more about the system as a whole (my best hope is for new lenses to materialize and a new lens roadmap to foresee what lies ahead).

Best regards,
Radu  

You always have interesting insights into Pentax Radu.  Can you explain #5 in your post please?  No grip to protect?

Tom

[LKaven]

It is a marketing truth, know in common speak as BS. 

After years of touting their 12 bit sensors as "16 bit" we finally get an almost-16 bit sensor.  For a moment, I thought "facts" were converging on reality.  We need new lies?

[Theodoros]

Hi,

It is a marketing truth, know in common speak as BS.  

Best regards
Erik

I've yet to see in pictures the 1.5 stops DR difference in favour of Nikon than Canon that DXO finds… and that's although I'm using Nikons for FF….

[Theodoros]

After years of touting their 12 bit sensors as "16 bit" we finally get an almost-16 bit sensor.  For a moment, I thought "facts" were converging on reality.  We need new lies?

…for sure! …remember "De Lorean"? …it became "back to the future".

[LKaven]

I've yet to see in pictures the 1.5 stops DR difference in favour of Nikon than Canon that DXO finds… and that's although I'm using Nikons for FF….

Lift the shadows at base ISO and you'll see it.  You can lift the shadows on the D800 6 stops without pattern noise or read noise becoming intrusive.  You cannot do this on the 6D/5DII/5DIII.

[ErikKaffehr]

Hi,

That may depend on Canon's having enough DR. Another factor is lens flare, it is normally about 0.5-1% on modern lenses and it reduces image contrast.

Upgrading my Sony Alpha 900 to the Alpha 99 would give me 1.5 stop of advantage but it took me 2 months to find it. It took shooting a repro  of a Velvia slide with extreme contrast in "totally dark" room, I needed a flashlight to find the camera.

As a side note, the Canon sensors are very good, it is the readout noise that kills DR, adding noise in the deep shadows. Increasing ISO reduces readout noise but also reduces exposure.

This pair of images shows shadow noise on P45+ (left) and Sony Alpha 99 (right) both exposed almost identical and pushed 4EV.


These are procentually similar crops scaled down to same size, so the size advantage of the P45+ is included in the small samples shown. Here the Sony Alpha image is a bit brighter although it has a bit more headroom in the highlights (around 1/3 stop less exposure based on raw histogram data), that difference may come from lens flare. Full info is here: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/80-my-mfd-journey-summing-up?start=1

Raw images are here: http://echophoto.dnsalias.net/ekr/index.php/photoarticles/80-my-mfd-journey-summing-up?start=2



Best regards
Erik

I've yet to see in pictures the 1.5 stops DR difference in favour of Nikon than Canon that DXO finds… and that's although I'm using Nikons for FF….

[Theodoros]

Lift the shadows at base ISO and you'll see it.  You can lift the shadows on the D800 6 stops without pattern noise or read noise becoming intrusive.  You cannot do this on the 6D/5DII/5DIII.

I have a D800 (it's an "E" actually… and I didn't say "I can't find any difference… I said I've yet to see the 1.5 stops of difference...

Another thing though… D800's extension of DR is a joke, it's what remains after processing than what is recorded on sensor that matters...

[Bart_van_der_Wolf]

The video mentions "14 stops of dynamic range at all ISO settings."

Yes, I've also noticed that, and it will be debunked before long. The 14-stops are likely (although TBD) for native sensitivity, which may be ISO 100, but it will drop as ISO gain is boosted or the image is underexposed. Just plain physics, currently there is no way around it.

Cheers,
Bart

[Vladimirovich]

What I meant is that there is no separate (battery) grip as an accessory s

but never say never, may be Pentax will make a battery grip for 645DmkII ?

[ErikKaffehr]

Hi,

I think it can work if you don't manipulate the raw file. Shooting 6400 ISO corresponds to -6 EV exposure. It will give very noisy shadows but 6 stops of extra protection in the highlights.

Best regards
Erik

Yes, I've also noticed that, and it will be debunked before long. The 14-stops are likely (although TBD) for native sensitivity, which may be ISO 100, but it will drop as ISO gain is boosted or the image is underexposed. Just plain physics, currently there is no way around it.

Cheers,
Bart

[jerome_m]

I think it can work if you don't manipulate the raw file. Shooting 6400 ISO corresponds to -6 EV exposure. It will give very noisy shadows but 6 stops of extra protection in the highlights.

The H3D2-50 works like that. It does not have an amplifier it its CCD, so it actually takes all pictures at ISO50 and manipulates the data numerically. I can insure you that ISO 400 is not what you want to use on that camera.

Since you like dxomark, you can check what they measured about that particular camera. It is in their database.

[Theodoros]

The question still remains though… "why would one spend that much money on an MF system that its only advantage will be higher Iso performance, when he can use a DSLR using a fast lens at a fraction of the cost, gain one or two stops of Iso performance (due to the faster lens) and "equalise" DOF  (again because of the faster lens?". Obviously, when buying an MF system, one should concentrate on near base Iso performance… everybody that buys MF has an additional DSLR… no? I am quite happy using Cmos for 35mm FF and CCD for MF… If I was to choose again today, I would do the same… I find no point in this release, unless it can compete with CCD at near base Iso…

[ErikKaffehr]

Hi,

Many cameras scale the RAW data with ISO, in which case they shift data to the right and thus throwing away data. You need a raw image to check. DxO mark doesn't publish raw data.

The image will be garbage of course, but the dynamic range will be there.

Best regards
Erik

The H3D2-50 works like that. It does not have an amplifier it its CCD, so it actually takes all pictures at ISO50 and manipulates the data numerically. I can insure you that ISO 400 is not what you want to use on that camera.

Since you like dxomark, you can check what they measured about that particular camera. It is in their database.

[BJL]

The question still remains though… "why would one spend that much money on an MF system that its only advantage will be higher Iso performance... Obviously, when buying an MF system, one should concentrate on near base Iso performance ...

1) The advantage will almost certainly not be only in low light/high ISO performance; the Sony Exmor CMOS sensor should also have an advantage at base ISO speed through lower dark noise levels and thus somewhat greater dynamic range.  Phase One reports a 14 stops DR for the IQ250, vs 13 for the IQ260 and IQ280, and this 14 stops fits with third party measurements of other Sony Exmor sensors of similar pixel size. (This is an inherent and well-evidenced advantage of CMOS sensors, and in particular Sony's Exmor sensors, over CCD.)  By the way, I recommend ignoring the 16-bit spec.: 16-bit conversion of a 13 stop signal simply adds "three bits of marketing noise", as has been explained many times.

2) Some people will appreciate convenient 24fps Live View for careful manual focusing.

3) Some people like the ability to use medium format lenses rather than the ones available for 35mm format.

4) Compared to options like Sony's 36MP 36x24mm Exmor sensors: even in situations where 36MP is enough, 51MP downsized to match the resolution and detail given by a 36MP sensor gains in "per pixel" SNR and DR, and thus in fineness of tonal gradations, shadow noise, etc.  For image shapes like 4:3 and 5:4, the image comes from almost twice the area on a 44x33mm sensor compared to using a 36x24mm sensor (90% more to be pedantic), and with the same core sensor technology, this is likely to deliver a visible difference in some situations.

[siebel]

I for one, am an exception to the generalisation that everyone has a DSLR system as well. I do NOT own a DSLR, nor do I intend to get one, most certainly not to shoot work for my clients. My needs, and those of my clients, are  currently fully met by my MF system, based around an IQ180 and a P45+ back on DF, Fuji GX680II and ALPA STC systems. Yes, I do own a 4/3 system as well, but it's purely my happy-snap toy. For this reason, I do follow very closely any new developments in MF-land. For the moment, the IQ250 has no interest for me, primarily because I will not sacrifice the angle of view I currently have with my favourite combo of the 23HR on the IQ180. I do, however, find the opportunities offered by having better LV and high ISO very exciting. When I can have it in a full-size sensor that works with a tech cam, then that will be my Holy Grail. The nature of my work and my clients makes me very much an outlier, or as a friend of mine describes me, an edge walker, so I don't think my criteria reflect the larger part of the market. I think this back will suit a great many shooters and will do very well. It also remains to be seen what the prematurely announced and one-day-to-be-delivered (Sorry Hasselblad, your over-promise and under deliver history, as manifested by your 60MP back launch, among others, remains a red flag) is going to deliver. I'm sure it's going to be a capable device also. Given that P1 still have to show us their new body (PLEASE!), and Blad are openly saying they have more to come soon, it's shaping up to be a big year. Reports of the demise of MF seem to be greatly exaggerated.