CMOS Look, CCD Look in MF Sensors

[fotografz]

Welcome to the forum Lobalobo! You sure picked a hotly debated subject for your first post It isn't a settled discussion by any measure, and there are those who are firmly in one camp or another.

Some swear that they can not see a difference … and if they can see it initially, usually feel that the differences can be mitigated through digital post processing. This may well be true, and is a valid point for those who wish to achieve a certain look through post work.

Personally, I'm in the camp that desires an out-of-camera response that is initially closer to my own image criteria … and have even chosen between different CMOS cameras to achieve that, let alone CMOS verses CCD MFD cameras.

My current MFD system is CCD, and was selected for many of the attributes that you mentioned, and others here have commented on.

However, attributes like expanded DR and higher ISO performance are of less importance to me than they are to others. I use MFD in fat light, either natural or by means of lighting. In fact probably 75% of my MFD work involves strobe work where ISO and DR are less of an issue because I control both through placement, quantity and quality of the light.

We have to remember that the beginnings of digital backs were aimed at studio shooters as the digital age crashed down on them and all media abruptly went digital. My first digital back provided any ISO as long as it was 100:ROTFL: … and required being tethered to a computer or a giant battery. Hardly "field friendly".

I can fully understand why a natural landscape photographer using a field camera would hunger for all the attributes that a CMOS back would provide … and then expertly work on hard earned key images in post no matter how much time it may take.

- Marc

 

[voidshatter]

 

[ErikKaffehr]

Hi,

I wouldn't say CMOS is simpler to make, but it is the dominating process for photographics sensors now, so most effort is going there.

CMOS has a few key advantages above CCD

• More electronics can be integrated on the chip
• Each pixel can be read directly, with CCD the pixels are popped from pixel to pixel several thousands times before readout
• CMOS allows for columnwise analogue digital conversion, this is what gives Sony sensors their DR advantage.

CMOS is probably more expensive to produce but readout can be made on the sensor itself, so there is no need for an analogue readout board. Also, DSLR sensors is just a very tiny bit of the CMOS sensor market (a few % by volume), so much of the development is driven by small devices.

Basically, the CCD is a very simple device, but a lot of design effort seems to be needed to make it work really well.

Weather CMOS or CCD, the devices do absorbs photons (quanta of light) and convert them to electron charges stored in capacitors. The major difference between CCD and CMOS is the ways the stored electrons are counted.

CMOS can read pixel values in a nondestructive way, so they implement something called correlated double sampling (CDS). Essentially reading the sensor pixels pre and post exposure. CDS and the lower readout noise on modern CMOS sensors gives images that have far less noise than CCD.

Canon sensors to CDS but have analogue readout, that is the reason that they don't achieve high DR at base ISO.

The present generation of CMOS used in MFDs has on sensor analogue digital conversion, with sensor coming from Sony or using the CMOSIS design on the Leica S (Type 007).

CFA designs and microlenses differ between different makers of sensors.

Best regards
Erik

As Leica has shown us with the S 007? even though the only real change was a move to a mature sensor design they've been selling for years now? or the price differential between the Phase, Hassy and Pentax variants on the same sensor?

Let's not be fooled into thinking that just because CMOS technology is somehow easier to manufacture, companies won't take advantage of the fact to improve profit margins. A typical pro DSLR body with sensor and all costs around $3500, or up to $6000 for those with a vertical grip, so Pentax isn't being unreasonable with their offer, a sensor that's 1.8x as large needs a like increase in camera size, and the price scales to reflect that. (Although the 645Z body could in theory support up to a full-size sensor meaning that it's bigger still than it needs to be.

 

[carstenw]

For cameras that cost more than sports cars cheapest available inputs is not a strategy likely to attract customers, particularly demanding pros (I'd think, though I'm not a pro). If CMOS wins in the long run, it will, I'd imagine (and hope) be on terms that provide options for the look the consumers demand, including the look of current CCD sensors.

I don't think the consumers are demanding the CCD look in CMOS sensor, which is likely why we don't have it already. Consumers are demanding high FPS, high ISO, high MP, Live View, and all sorts of other things, but very, very, very few people are really interested in subtle colour differences, sadly.

Going back to your original question, I think all the information is there in the new CMOS sensors, but you have to be a lot better at post-processing to pull it out, whereas with CCDs with older-style "fat" CFAs, the look is already pleasing. The M9 vs. M240 is another example of this difference. The M240 files have more meat on them, but they come out flatter looking and need some tweaking.

If you want great images out of the camera, then get a CFV-16 II or something like that. If you want the most information possible in the file, then get one of the newest backs with CMOS sensors, but make sure you brush up on your post-processing skills first.

 

[Lobalobo]

You sound like a CCD salesman and you say you currently shoot 4x5 film; are you familiar with digital post processing?

EDIT: I edited this post because I had a suspicion the OP may be trolling the forum up-selling CCD over CMOS, and my apologies if I am wrong, but the OP's timing could not have happened any quicker than after the CFV-50c hit the market as a win-win for photographers and for Hasselblad.

Wow. That's quite a response. I am an amateur who shoots 4 x 5 film and scans on an old Epson 1680, then has the product printed at a commercial shop. I am an academic, a law professor, but it's good to know that I could have a future as a CCD salesman, though I'm not sure for whom I'd work, as it seems quite clear that CMOS is the future. My interest was purely academic (an occupational hazard) as it seemed that technology might have been moving backwards in one respect to move forward in another, a surprising move, but am glad to hear that this is not the case (glad for the day I might splurge for a divorce-risking purchase of a digital back).

What most amazes me about the venom in your post is that I've repeatedly in this thread that I know little and am seeking information from those who know more. Not much of a sale pitch.

 

[Lobalobo]

I don't think the consumers are demanding the CCD look in CMOS sensor, which is likely why we don't have it already. Consumers are demanding high FPS, high ISO, high MP, Live View, and all sorts of other things, but very, very, very few people are really interested in subtle colour differences, sadly.

Going back to your original question, I think all the information is there in the new CMOS sensors, but you have to be a lot better at post-processing to pull it out, whereas with CCDs with older-style "fat" CFAs, the look is already pleasing. The M9 vs. M240 is another example of this difference. The M240 files have more meat on them, but they come out flatter looking and need some tweaking.

If you want great images out of the camera, then get a CFV-16 II or something like that. If you want the most information possible in the file, then get one of the newest backs with CMOS sensors, but make sure you brush up on your post-processing skills first.

Interesting, and if you don't mind offering a bit more education to the ignorant, I am curious about the metaphor "fat" versus "thin" CFAs. Does "fat" refer to higher contrast, or is it something more? Thanks.

 

[SHV]

"If you want great images out of the camera, then get a CFV-16 II or something like that."
*****
I recently started using my CFV-16 again after it had spent about 2 1/2 years in storage. There is just "something" about the "fat" pixel files that is special but difficult to describe. The biggest downside of the CFV-16 for general use, IMO, isn't resolution but low ISO. Since I hadn't used it for a significant period of time, I was considering selling it. I changed my mind about that.

Steve

 

[bensonga]

Interesting, and if you don't mind offering a bit more education to the ignorant, I am curious about the metaphor "fat" versus "thin" CFAs. Does "fat" refer to higher contrast, or is it something more? Thanks.

The reference to "fat" pixels is to sensors/backs with photo sites in the 9 micron range, such as the Hasselblad CFV-16, Phase One P20/25, Leaf Aptus II 5 and several others from that generation of medium format digital backs. Most have resolutions in the 16-22 megapixel range.

There are examples here:
http://www.getdpi.com/forum/medium-format-systems-digital-backs/44225-fat-pixel-digital-backs.html

Gary

 

[Lobalobo]

The reference to "fat" pixels is to sensors/backs with photo sites in the 9 micron range, such as the Hasselblad CFV-16, Phase One P20/25, Leaf Aptus II 5 and several others from that generation of medium format digital backs. Most have resolutions in the 16-22 megapixel range.

Gary

Actually, the quote I asked about was about fat "CFA"s, not fat pixels. Another matter, I think.

As for fat pixels, that's another hot-button topic, right? DPReview used to include pixel density as a specification for cameras on the theory that too many pixels on a sensor degrade image quality because small pixels clip highlights too readily and because of noise. This created a fire storm of protest by technically-minded photographers who said that clipping can be avoided either through technique (with blocked shadows rescued in software) or pixel binning and that while noise was a theoretical issue, density in practice is nowhere near high enough to be a problem. The consensus was, I thought, that while sensor size matters (more total info for the image) pixel density does not. This was for smaller sensors, of course, but the principles should be the same, no (cross-talk issues with movements aside)?

I have to say, my uninformed prejudices were to support both theories: that CCD looked different from (and to my eye better than) CMOS and that large pixels looked different (richer, more saturated) than small pixel-images; that is, my subjective belief that there was a price to pay both for high ISO performance and extra resolution and that if shooting at base ISO where huge prints were not going to be made, it was hard to beat a CCD sensor with large pixels. But the people who know the technical truths say otherwise, mostly. Perhaps, though, others disagree.

 

[Lobalobo]

Although I had read it before, I realize that I neglected something in Doug Peterson's excellent LuLa article "CMOS Fully Realized". Doug said this:

As one example [of a tradeoff], the selection of a CFA, the color pattern put in front of the sensor, is a choice between quality of color, and ISO performance. If the CFA allows each pixel to see a broader spectrum of color (e.g. for the green pixels to see a bit further into yellow) a camera’s ISO range can be modestly increased. The resulting loss in color quality is subtle – subtle variations in color are missed and a handful of specific colors become difficult to photograph.

Perhaps the answer is just simply that the Sony CMOS backs are catering to those who care about higher ISO capabilities and so the CFA has been designed to be just slightly less favorable to color subtleties (such as appear in the Morgan Library images). This would be sensible as those who care more about color subtleties than ISO can now choose a CCD back. If that's right, when the CCD backs disappear, there may be different flavors of CMOS backs, some of which emphasize higher ISO, some of which favor color subtleties. That would be a good result and would be particularly great if CMOS brought the cost down, for pros and for hobbyists (like me) who might spend $10,000 on a digital back but could not spend five times that.

All this said, there is something else to note about Doug's LuLa article: it contains samples with spectacularly beautiful but natural looking colors, at least on a computer screen (a big "at least" I know):

The Phase One IQ250 CMOS Fully Realized

This does indeed support Doug's conclusion in the article that at least the Phase One version of the Sony CMOS colors are awfully good even from the perspective of those who care enough about small detail and subtlety to spend a small fortune on a camera back.

All in all, seems easy to see why there is a debate here, over what is ultimately a subjective judgment. Glad those who know are willing to share their views.

 

[darr]

Wow. That's quite a response. I am an amateur who shoots 4 x 5 film and scans on an old Epson 1680, then has the product printed at a commercial shop. I am an academic, a law professor, but it's good to know that I could have a future as a CCD salesman, though I'm not sure for whom I'd work, as it seems quite clear that CMOS is the future. My interest was purely academic (an occupational hazard) as it seemed that technology might have been moving backwards in one respect to move forward in another, a surprising move, but am glad to hear that this is not the case (glad for the day I might splurge for a divorce-risking purchase of a digital back).

What most amazes me about the venom in your post is that I've repeatedly in this thread that I know little and am seeking information from those who know more. Not much of a sale pitch.

No venom intended. I am just a little tired of sales pitches to the uninformed, but you cleared that up, and again, my apologies. :angel:

What is your current post processing workflow once you scan your 4x5" film?

 

[Lobalobo]

No venom intended. I am just a little tired of sales pitches to the uninformed, but you cleared that up, and again, my apologies. :angel:

What is your current post processing workflow once you scan your 4x5" film?

No offense taken. (And by the way, your Hasselblad 50c images are beautiful, the lighthouse at sunset in particular.)

To answer your question, my workflow has changed. I used to scan (typically Provia 100) on the Epson 1680 with Silverfast to JPEG, working hard to fine-tune the image with that software then making minor adjustments in some rudimentary processor such as PaintShop Pro. I've shifted though away from fine-tuning on the scanner, using VueScan at basic settings, again creating a JPEG (or occasionally a TIFF), then ultimately adjusting in Paintshop, frequently with an assist from Topaz.

Not high tech at all in either case, and I hope to increase my Post Processing skills along with my other skills. Of course, Lightroom or Photoshop or both would be a good start, but shooting film, my approach has been to try and get it right in the camera then replicate the film with a scan.

 

[richardman]

The sensor manufacturers will move to CMOS and in fact have mostly done so already. So even if there is a "look," it's like bemoaning the passing of Ektachrome or Kodachrome.

"Oh well."

 

[richardman]

...I've shifted though away from fine-tuning on the scanner, using VueScan at basic settings, again creating a JPEG (or occasionally a TIFF), then ultimately adjusting in Paintshop, frequently with an assist from Topaz.
..

Save it as TIFF. JPG loses some quality. With 4x5 Provia scans, unless you have to fight to lift shadow details, just a few tweaks will get you incredible images.

Oh, get the V700/750 or the new 800/850, or if you are on a budget, the old 4990 in used market. I have made 24x32" prints from 4x5 with a V700 that looks quite good.

 

[Lobalobo]

Save it as TIFF. JPG loses some quality. With 4x5 Provia scans, unless you have to fight to lift shadow details, just a few tweaks will get you incredible images.

Oh, get the V700/750 or the new 800/850, or if you are on a budget, the old 4990 in used market. I have made 24x32" prints from 4x5 with a V700 that looks quite good.

You are right, of course, regarding JPEG versus TIFF, but for 11 x 14 prints (as large as I generally print) I have not noticed the difference so I sometimes cut that corner. (If I were printing larger, I'd likely pay for a drum scan anyway.) Also, I agree that few tweaks are needed, which has been my experience and why I've stopped fiddling so much with Silverfast. What will make a difference, I believe, is the Epson 800 scanner, which is on my buy list (once I come to grips with the fact that I'm not likely to buy a MF system, at least not until my kids are out of college). The improvement I expect from the new scanner is not so much from the new light source but from the fact that it is designed to be used with a provided 4 x 5 film holder that flattens the film between glass, which is advertised to work without causing Newton rings. We'll see.

 

[Lobalobo]

The sensor manufacturers will move to CMOS and in fact have mostly done so already. So even if there is a "look," it's like bemoaning the passing of Ektachrome or Kodachrome.

"Oh well."

Yes, but the Kodachrome look, e.g., is still available in digital (with a CCD sensor anyway ). The hope is that options are not eliminated, just shifted to newer technologies.

 

[bensonga]

Actually, the quote I asked about was about fat "CFA"s, not fat pixels. Another matter, I think.

As for fat pixels, that's another hot-button topic, right?

Wow, for a newby, you sure know a lot more than I do. Admitedly, I'm just an amateur with too much gear.

I had never heard the term "fat CFAs". How are backs with fat color filter arrays different from backs with "fat pixels"?

Gary

 

[Lobalobo]

Wow, for a newby, you sure know a lot more than I do. Admitedly, I'm just an amateur with too much gear.

I had never heard the term "fat CFAs". How are backs with fat color filter arrays different from backs with "fat pixels"?

Gary

Just to be clear, I don't know what "fat CFAs" means, either, which is why I asked. An earlier post used that term. So maybe we'll both find out. I think, having done some further research, that the poster probably meant by "fat" purer colors (fat within the color spectrum rather than at the thin ends of that color) but I could have that entirely wrong.

 

[richardman]

4x5 film are thick and quite flat even in the standard Epson holder. I have tested it against the Better Scanning glass "holder" and the difference is minor. Silverfast is great if you are scanning color negs. For color pos and B&W, Vuescan is great.

I used a fat pixel P25+ for a few months, and it's definitely quite nice. Not where near 4x5 nice, but better than M9 nice.

If anyone wants to throw away their ANCIENT CCD back, let me know and I will properly dispose it for you XD.

 

[richardman]

Ben, you know where to send your obsolete Pentax 645D and other useless gears