Some questions regarding Phase One Sensor+ technology

[yashima]

Is there any gain in dynamic range or colour depth using Sensor+, or is it simply about sensitivity/high ISO performance?

If I'm about to compare IQ180 Sensor+(20mpx) to P25+ (22mpx), would the noise level at Sensor+ 800 ISO comparable to P25+ ISO 100 or 200? And apart from Dalsa vs Kodak colour, what other significant difference would I find?


I really enjoy using Sensor+ on my new IQ160 upgrade. Is there some technical reason that this can not be done with CMOS sensor, especially the recent IQ3-100? That would make it alot more attractive for me, and some others I would imagine. For many work I would not want to wrestle with 60MB/100MB files all the time.

[Paul2660]

In my usage, with IQ180, IQ280, and IQ160, and IQ260, sensor plus does give you more dynamic range in any ISO setting past 200, and in some cases, even better than 200.  Even in good light, with 400 and 800, you will start to see loss in color saturation and shadow range.  In low light the loss in both is extreme and the sensor plus files will be superior in all respects besides resolution. 

I can see really no reason for Phase to use Sensor plus with their CMOS tech, as the main reason for sensor plus is for lower noise @ higher iso albeit at 3/4 loss in overall resolution.  To cripple a 100MP back down to 25MP, when the images so far @ 400, 800, and 1600 are quite good, vastly superior to any CCD back I have used over the years, makes no sense.  At least to me.   You then open the entire discussion to lets compare the output from 100MP CMOS 25MP to a DSLR/Sony, Nikon etc. 

Sensor plus was always a loosing proportion for my work, and I really never used it, as the resulting loss in overall resolution was just too much, where as if it could have been done by only a 1/2 loss, it might have been a great solution. 

Paul C

[ErikKaffehr]

Hi,

Yes, Sensor+ would help with DR. What it does is essentially to increase the number of photons captured per pixel, while keeping readout noise constant. With CCDs readout noise is relatively high while modern CMOS can have pretty low readout noise. In theory, Sensor+ would give the same DR at ISO 1600 as the te full resolution sensor at 400 ISO. DR essentially translates into clean shadow detail.

With CMOS there is less need of tricks like Sensor+ as readout noise is low to begin with.

CMOS has a trick in it's sleave, sometimes called the Aptina patent. It effectively reduces the well capacity of the sensor at a set ISO, thus increasing the voltage fed to the ADC.

Bill Claff has a very nice figure demonstrating this: http://www.photonstophotos.net/Charts/PDR.htm#Nikon%20D810,Pentax%20645Z,Sony%20ILCE-7RII

Of the three sensors the Sony A7rII uses the Aptina trick. As you see, DR has a small jump at 640 ISO, this is the effect of the Aptina trick. At 640 ISO the sensor is only using a sixth of it's full well capacity, so cutting that Full Well Capacity in half (or so) actually helps DR.

The Nikon D810 doesn't use the Aptina trick so it lags behind the A7rII at high ISO (at least according to Bill's data). A problem with Bill's data is that he uses nominal ISO, but vendors take a lot of liberties with nominal ISO, which the norms may allow.

If we see a next generation of CMOS sensor's in MFD it may be probable that those sensor's would use that Aptina trick. With some probability Sony either licenses the Aptina patent from ON Semi or has a cross license arrangement. Licensing costs and issues may prevent Sony from using that trick on MFD.

As a side note, I would ask myself if we are going to see a rapid development of MFD sensors now that almost all vendors are using Sony CMOS? Will MFD follow with sensor updates for each iteration of Sony technology or will MFD staying at a longer development cycle?

Best regards
Erik

In my usage, with IQ180, IQ280, and IQ160, and IQ260, sensor plus does give you more dynamic range in any ISO setting past 200, and in some cases, even better than 200.  Even in good light, with 400 and 800, you will start to see loss in color saturation and shadow range.  In low light the loss in both is extreme and the sensor plus files will be superior in all respects besides resolution. 

I can see really no reason for Phase to use Sensor plus with their CMOS tech, as the main reason for sensor plus is for lower noise @ higher iso albeit at 3/4 loss in overall resolution.  To cripple a 100MP back down to 25MP, when the images so far @ 400, 800, and 1600 are quite good, vastly superior to any CCD back I have used over the years, makes no sense.  At least to me.   You then open the entire discussion to lets compare the output from 100MP CMOS 25MP to a DSLR/Sony, Nikon etc. 

Sensor plus was always a loosing proportion for my work, and I really never used it, as the resulting loss in overall resolution was just too much, where as if it could have been done by only a 1/2 loss, it might have been a great solution. 

Paul C

[Graham Welland]

I'm sure that there's no reason at all why CMOS sensors couldn't be pixel binned in the same manner as Phase One does with their Sensor+ approach. At the end of the day it's all about picking which pixels to bin together - I don't know if Phase One have a patent or not for their binning pattern on Bayer or not. However, given the number of pixels we have to play with these days I expect this to be the optional norm in the future actually. I'd take a 50/100/200 mp sensor and pixel bin down to superior DR at 12/25/50mp resolution.

[Christoph B.]

IIRC Dalsa has the patent on hardware based pixel binning and I doubt they'll let someone else have it.
But let's be realistic - there isn't really much use for it with CMOS due to the high ISO performance.

It's like adding the option for a dark frame subtraction - nice idea and someone might find it useful, but I guess it wouldn't be feasible to implement such a feature because the people who would pay a premium for it are scarce.

[Doug Peterson]

IIRC Dalsa has the patent on hardware based pixel binning and I doubt they'll let someone else have it.
But let's be realistic - there isn't really much use for it with CMOS due to the high ISO performance.

It's like adding the option for a dark frame subtraction - nice idea and someone might find it useful, but I guess it wouldn't be feasible to implement such a feature because the people who would pay a premium for it are scarce.

The Patent was filed and is owned by Phase One.

[Theodoros]

IMO, it is rather surprising that no binning method is applied from any of the makers on the Sony 100mp sensor... Other than the extensive LL ability (which admittedly isn't as attractive as with CCD sensors since LL performance is already much better - but still it would be there if ever needed), the ability to only use 100mp when it is really needed would be very welcomed indeed...

Further more, if one could use a 25mp pixel file but at "true colour" (by having a mixed result out of all 4 pixels binned, i.e. a process of the GRGB information), that would be even more tempting... Is there a technical reason why no binning is applied to the Sony Cmos 100mp sensor?

[yashima]

In my usage, with IQ180, IQ280, and IQ160, and IQ260, sensor plus does give you more dynamic range in any ISO setting past 200, and in some cases, even better than 200.  Even in good light, with 400 and 800, you will start to see loss in color saturation and shadow range.  In low light the loss in both is extreme and the sensor plus files will be superior in all respects besides resolution. 

I can see really no reason for Phase to use Sensor plus with their CMOS tech, as the main reason for sensor plus is for lower noise @ higher iso albeit at 3/4 loss in overall resolution.  To cripple a 100MP back down to 25MP, when the images so far @ 400, 800, and 1600 are quite good, vastly superior to any CCD back I have used over the years, makes no sense.  At least to me.   You then open the entire discussion to lets compare the output from 100MP CMOS 25MP to a DSLR/Sony, Nikon etc. 

Sensor plus was always a loosing proportion for my work, and I really never used it, as the resulting loss in overall resolution was just too much, where as if it could have been done by only a 1/2 loss, it might have been a great solution. 

Paul C

Hi Paul,

Do you mean that Sensor+ files at ISO 800 can have even better DR than full resolution files at ISO 200 (noise should be the same, that's what Phase One says anyway)? If that's true then thats quite a big plus.

Regarding CMOS sensor, of course its high ISO performance is great, however there are still degradation from ISO3200 in comparison to ISO800, and similarly ISO800 to ISO200. So it still make sense to use Sensor+ to shoot at high ISO and get file quality of low ISO files at the same time. Especially in circumstances where we dont need huge files.

In any cases, its one of Phase One unique features, if they had collaborated with Sony to make exclusive Phase One - Sony sensors that had Sensor+, that would had given them an USP and differentiate from Hasselblad/Pentax products. But I guess the collaboration process for making sensor can be too complicated.