IQ180 vs Up-Res A7R2

[rhern213]

So the quick story behind this is that I got my A7R2 and now thinking that it's possible I could replace my IQ180 if there wasn't that much drop-off in the IQ of the final prints.

I'm planning on taking some real shots and doing real prints, but this was a quick monitor viewing up-res comparison to see what it spits out.

The exact same shot was taken with an IQ180 and 80mm LS lens, and an A7R2 with the 55mm 1.8. The IQ180 was at F11, and the A7R2 at F8. All of the necessary focus and stabilization methods were followed to ensure optimum sharpness.

Here's the original full comp:



Here's the A7R2 100% crop, RAW conversion in C1, no processing.



Here's the A7R2 100% crop, UP-Res'd in PS to match the IQ180 size, Un-Sharp Mask at 100%.



Here's the IQ180 100% crop, Raw conversion in C1, no processing.




As expected, the microcontrast from the IQ180 is clearly evident, the A7R2 just can't compete. Right away you can see so much texture in the rose that are simply non-existent in the A7R2. However, at more than 2x the sensor size, and 2x the MP, for the overall IQ of the subject and important details, I think it's very close. In the end my main concern is prints, also up-res'ing vs lower DPI is a whole other story, but I think the differences should be even less evident on print.

But I've already started thinking, well if I can up-res the A7R2, imagine how much I can up-res the IQ180 then.

So I'd like to hear your guys thoughts, it's not the best method of testing but it was the simplest to get a quick idea. Maybe you guys have comments on how I can make more accurate assessments with other quick steps in the process?

 

[Paratom]

I think the phase image shows so many more shades of pink. The Sony image looks artificial and dead compared to the IQ image.
I have not been 100% happy with color from my a7II as well, specially skin tones dont look 100% like I would like them to look.

 

[torger]

I don't see any more shades of pink in the IQ180 image. I made a quick color analysis of the images and the IQ180 contains 84724 unique shades and the A7RII contains 88844. Still I can agree that the IQ180 rose looks more refined in terms of highlight rolloff and to some extents in the midtones too.

I think this is however much due to that the subject is at the of resolving limit of the texture of the artificial leaves. With the IQ180 you see the texture, with the A7RII the texture is gone. The visibility of the texture makes the IQ180 look superior. I think if you had shot a smooth object where there would not so evident texture difference you wouldn't see much difference in tones. If the film curve is different the one with smoother rolloff can look more refined, but that is of course easily fixed.

When it comes to color I've worked the past six months quite extensively with my new camera profiling project DCamProf (DCamProf) and even more clearly now than before I'd say color sits 90% in the profile and only 10% in the hardware. Phase One has put more effort into the IQ180 profile for sure, and especially aim at satisfying the taste of MFD users so I would expect the color to be more refined. It's not impossible to make and tune your own profile though to satisfy your own taste, and it's not guaranteed that it's the same as the Phase One designers. The A7RII sensor is extremely capable in terms of color separation so if you just know what you want to achieve in terms of color you should be able to get there.

The big advantage of the IQ180 is the extra resolution, and if you like the result of the bundled color profiles you have it there ready to use.

The difference in resolution is exaggerated in this test when looking it at roughly 100ppi (screen) on a subject where we see texture in one image and not the other. If you print small enough to achieve 200ppi with the A7RII the resolution advantage will be less evident.

If I did this evaluation myself I would mostly consider my print sizes and what I would lose in nosing quality for those. I would not worry much about the color, but that is because I'm quite confident that I can get my color where I want it and I don't use bundled profiles at all as I prefer to be in control over color (one reason being that it becomes simpler to move between systems while keeping the same color).

Personally I couldn't afford an IQ180, I'm using a 50 megapixel H4D-50 mounted on a Linhof Techno with Schneider Digitar lenses. The A7RII exceeds my aging sensor in many aspects, but I wanted the "large format experience" with lens movements etc, without having to mess with film, and I've got exactly that which I'm pleased with. If I had shot with the MF-DSLR rather than a tech cam and only use it for the image quality I would have stepped back to 135 about now.
I'd probably choose the Canon 5Ds though, not as good as the A7RII in terms of dynamic range, but the ergonomics is better and I'm used to Canon. There are some unique aspects to the MF-DSLR shooting experience though, the larger viewfinder, the native 4:3 aspect ratio, leaf shutters so if those are important to you it should also be considered in the evaluation. To some those are "killer features", for me they aren't that important. Sure I prefer 4:3 over 3:2, but it's not really worth that many thousands of dollars. I've played a bit with my H4D-50 body and it's indeed well-designed and nice and all but as soon as I'm going hand-held to do something serious I use my Canon, and for tripod-mounted work I use my Linhof. Every user is different though.

 

[torger]

If you don't like skintones, an interesting experiment is to shoot some side by side shots compare A/B and then analyze why you like the skintones on one camera or the other.

Maybe it's the hue? Then you can color correct, RGB curves is a classic method that still works well. Maybe it's the hue separation, if you prefer a smoother look, then compress the gamut there. Maybe the skin looks more brilliant, then it could be the highlight rolloff, then adjust the contrast curve, or scale lightness in some range of the gamut. Does the shadow range look dull or grayish? Try adding some more saturation there. Spend some serious time on it. I'm quite confident that what you will end up with as the ultimate skintones will not be exactly as the bundled profile as tastes differ.

It doesn't help us to assume that color is an property mostly bound to the camera hardware. That is what manufacturers want us to believe as it's an excellent way of locking in customers. Leaf looks way different from Phase even when using the same sensors, and that is all about profile design. I have only had one Hasselblad so I don't know for sure, but I've heard Hassy has the approach to make their cameras look the same, regardless of sensor, which many pros appreciate as you can switch to a secondary camera in the middle of a shoot and still maintain the look, regardless if the sensor is a Sony, Kodak och Dalsa.

Hardware does have an effect though, color responses differ between sensors. What it means in practice is that while you can make a camera look virtually identical to another for a specific light (like studio flash), the looks will separate some when you shoot under a different light. The ability to separate color does differ a bit too, but all modern sensors do a good job for normal saturated colors, while there may be a significant dropoff in precision in ultra-saturated colors.

I find it highly unlikely though that the IQ180 Dalsa sensor would be superior than the A7RII concerning color in terms of hardware, but I find it very likely that more work has been put into the IQ180 profiles. And as we see in the first posted images the ability to resolve texture can make colors appear more refined.

I guess what I try to say is that if you find colors lacking there is a very good chance you can improve on that, but if you find important advantages bound to the resolving power then the A7RII won't be able to compete.

 

[f8orbust]

The printed image is a great leveler in many respects, so if I was doing such a comparison I would print the images at my usual size and then compare from a non-photographer viewing distance (i.e. not 1 inch). My guess is that most people, from a sensible viewing distance, wouldn't see much, if any, difference.

I used to say that if you print really big and/or make large crops and/or need more movements than a 35mm T/S lens will provide and/or love the workflow then MF still has a role. But since you can now get the last two by attaching a 35mm mirrorless camera to an Actus or Universalis, the list is reduced to: print really big or make large crops.

 

[jlm]

to f8's comment about tech cams:

what is the widest angle lens that will work with the Actus/A7R combo? i can get a 23mm using a MFDback, but in that case the sensor is very close to the standard. you have to figure in the minimum standard-to-standard thickness and the back standard to camera flange distance, and then consider the sensor is 14mm or so into the body of the A7

 

[torger]

As far as I understand this case is about the DF system, not a tech cam. But in any case, considering the wide angle options is indeed important as it's there you find the largest quality differences and also differences in flexibility (movements etc).

 

[rhern213]

My thoughts exactly.

From a standard 300 dpi print that's 30-40", even needing to up-res the A7R2 files maybe 150-200% as I have done above, I'm sure there's not going to be any real discernible difference.

My concern is going to be when I want to print upwards of 60", where a significant up-res and 200 dpi output or so is used, then the quality differences will be magnified. But then for that there's always the stitching option.

Shooting tech cams for wide angles is a different subject though.

The printed image is a great leveler in many respects, so if I was doing such a comparison I would print the images at my usual size and then compare from a non-photographer viewing distance (i.e. not 1 inch). My guess is that most people, from a sensible viewing distance, wouldn't see much, if any, difference.

I used to say that if you print really big and/or make large crops and/or need more movements than a 35mm T/S lens will provide and/or love the workflow then MF still has a role. But since you can now get the last two by attaching a 35mm mirrorless camera to an Actus or Universalis, the list is reduced to: print really big or make large crops.

 

[f8orbust]

what is the widest angle lens that will work with the Actus/A7R combo...

Once Cambo release their EF adapter (ACB-CA), my understanding is that you'll have access to all the high quality, large IC WA glass (see the threads by Stefan Steib), like the amazing 11-24, 17, 24 etc.

 

[Dan Santoso]

Can you post DR sample? Maybe push shadow and highlight recovery.

Thanks

 

[Paratom]

If you don't like skintones, an interesting experiment is to shoot some side by side shots compare A/B and then analyze why you like the skintones on one camera or the other.

Maybe it's the hue? Then you can color correct, RGB curves is a classic method that still works well. Maybe it's the hue separation, if you prefer a smoother look, then compress the gamut there. Maybe the skin looks more brilliant, then it could be the highlight rolloff, then adjust the contrast curve, or scale lightness in some range of the gamut. Does the shadow range look dull or grayish? Try adding some more saturation there. Spend some serious time on it. I'm quite confident that what you will end up with as the ultimate skintones will not be exactly as the bundled profile as tastes differ.

It doesn't help us to assume that color is an property mostly bound to the camera hardware. That is what manufacturers want us to believe as it's an excellent way of locking in customers. Leaf looks way different from Phase even when using the same sensors, and that is all about profile design. I have only had one Hasselblad so I don't know for sure, but I've heard Hassy has the approach to make their cameras look the same, regardless of sensor, which many pros appreciate as you can switch to a secondary camera in the middle of a shoot and still maintain the look, regardless if the sensor is a Sony, Kodak och Dalsa.

Hardware does have an effect though, color responses differ between sensors. What it means in practice is that while you can make a camera look virtually identical to another for a specific light (like studio flash), the looks will separate some when you shoot under a different light. The ability to separate color does differ a bit too, but all modern sensors do a good job for normal saturated colors, while there may be a significant dropoff in precision in ultra-saturated colors.

I find it highly unlikely though that the IQ180 Dalsa sensor would be superior than the A7RII concerning color in terms of hardware, but I find it very likely that more work has been put into the IQ180 profiles. And as we see in the first posted images the ability to resolve texture can make colors appear more refined.

I guess what I try to say is that if you find colors lacking there is a very good chance you can improve on that, but if you find important advantages bound to the resolving power then the A7RII won't be able to compete.

Thank you for the tips.
My only experience with custom tone curves was with a color checker passport and not leading to any really universal profile.
If there are cameras out there where I can get the color and look I like just with the standard profiles of c1 or lr and without having to do much post processing, than thats the way I prefer. Maybe I should spent more time at the computer but I prefer the time with the camera and therefore its an important factor for me.
Regarding the comparison image posted here...maybe there are not more shades of pink, but to me it looks like there are more, and it looks more alive. just my impression from looking at it.

 

[Jack]

If you look at color fidelity, there is simply no contest. Detail -- micro-contrast -- is clearly lacking in the uprez A7 file too, but it is subtle enough that I doubt it would bother most in a print. But the color tonality advantage of the IQ is very apparent and realtively huge...

 

[torger]

Color fidelity (and then I assume you don't mean colorimetric accuracy) and color tonality are qualities that are hard to pinpoint. The differences that may appear as tonality in the examples I think are related to the increased resolution.

Making two examples where both cameras are sampled down would make it easier to compare that particular aspect. And of course make the test blind, which you should do also when comparing prints. We are all biased, even if we try to be neutral knowing which is which will affect our judgment.

Small differences in exposure and contrast curve can have effects too. If we look for tonality a lower contrast curve with softer highlight and shadow rolloff will have an advantage, while a higher contrast curve can make an overall better impression giving more "punch", depending on subject.

I assume that the IQ180 have a number of more choices concerning both profiles and curves than the A7R2 in Capture One. If you don't want to mess around with trying to improve the defaults, which indeed is not an easy task, the looks you get there is what you get. But then keep in mind that you're mostly comparing profiles and curves rather than hardware. Actually, I believe that the Capture One software and the profile designer at Phase One is more important to the medium format color reproduction reputation than the hardware itself.

Personally I think it's possible to improve on the color reproduction Capture One offers, also for the IQ180. C1's color pipeline is not modern, it's filled with legacy from the 1990s most notable the RGB tone curve look with all its color shifts is built into the model. One can see how the C1 bundled profiles are partly compensated for that, but it isn't as bad as it could be as the curve is applied on the desaturated "raw data" before the ICC rather than after. Anyway, in order to apply different curves to the same profile (which really isn't a good approach as color appearance and contrast is tightly coupled) the profile color is tuned to not produce too bad colors with either linear response or the film curves, which of course compromises fidelity for both. That they do it like this is that it builds on film legacy, the idea that some color distortion bound to the curve is good as that was how film photographers were used to seeing the world.

I think it's better to design a profile with one fixed curve and only for that curve. Unfortunately profile design tools available to consumers today are mostly useless if you want to do something else than reproduction photography, which of course is great for the camera manufacturers. I have my own tool of course, but it's not really for everyone and it's still in development.

So in the end one will still have to treat the camera as the whole system with software, and the results you can get from that. If we don't have the power to change things it doesn't really matter if the "color fidelity" sits in the hardware or the software. As someone that has worked closely with the software for quite some time I believe that most sits in the software, and that there isn't really that much special about the MFD hardware. It's a big sensor with lots of pixels with great optics ahead of it which gives you lots of detail, which indeed can be a winning aspect.

 

[Paul Jameson]

IMO, the Sony looks bad in comparison. The two files are not even close and I would not be happy at this size, nor would my clients. I would imagine the IQ180 down sampled to the a7r would still look better too.

I've endlessly toyed with dumping my medium format and focusing the funds on other gear, usually when I listened to marketing hype of new cameras, or the idea of something more simple, or just new. But then I get a job that needs 3m prints and I forget about it. Even with high res 35mm bodies on offer, the only thing that gets closer is file size. I don't find the IQ very comparable all. In my quick testing of the a7r II, all things relative, i've found the output a bit disappointing. Flat and plasticky looking.

 

[GrahamWelland]

If you look at color fidelity, there is simply no contest. Detail -- micro-contrast -- is clearly lacking in the uprez A7 file too, but it is subtle enough that I doubt it would bother most in a print. But the color tonality advantage of the IQ is very apparent and realtively huge...

Like you Jack I see the differences immediately and I'm prepared to pay for them. However, they are pretty subtle and probably completely lost on the majority of owners of a print and for us the difference is 10x the cost. Is it worth the extra? Only the shooter can answer that. It's pretty scarily close though.

love my IQ150 and all the other sonys that I own, also my Nikon DF, but I still expect that I'll pick up an IQ180 or Credo at some point.

 

[Amin]

The difference in detail is not subtle at all!

While I've no doubt that the IQ180 also has advantages in color and tonality, I don't think those can be judged adequately from these samples. Ie, the Sony RAW could be processed to look much closer to what we're seeing in this processed IQ180 crop, and there is no such thing as "no processing" in C1. Sticking with the default processing just embraces the tone curve and color choices chosen as default by C1.

 

[PeterA]

Compelling subject matter is compelling....

 

[rhern213]

I admit I did just a basic Photoshop resize to the dimensions of the IQ180, and then unsharp mask to 100% and that was it. Actually I also bumped up the saturation on the A7R2 file to match the color intensity of the IQ180. The colors on the original raw output of the A7R2 were duller.

I'm going to try and do more fine detailing to see if I can get the A7R2 file a little closer.

Does anyone care to have the RAW file and give it a go themselves?

The difference in detail is not subtle at all!

While I've no doubt that the IQ180 also has advantages in color and tonality, I don't think those can be judged adequately from these samples. Ie, the Sony RAW could be processed to look much closer to what we're seeing in this processed IQ180 crop, and there is no such thing as "no processing" in C1. Sticking with the default processing just embraces the tone curve and color choices chosen as default by C1.

 

[vjbelle]

I have both an IQ180 and a Leaf Credo 50..... and also an 800e and had an A7r. I almost always print to 40 inches in one direction and many times to 48. A 36/42mp file simply will not compete with even a 50mp file..... especially a 50mp file from the Credo 50. I've also found that 50mp is the real sweet spot for file size vs print size when printing out to 40 inches. Even on screen I can't see much difference between 50mp and 80mp. Sure, there's a difference but its extremely difficult to see in the final print - if at all. Another issue is the quality of the lenses for each platform. There aren't many lenses - if any - that can compete with the Rody or Schneider Digitar's. The lens quality plus the quality/size of the sensor really widens the gap between the two platforms.

Victor

 

[Kolor-Pikker]

Do keep in mind that the A7 doesn't achieve it's small size without losses, and I'm not just talking about compressed Raw. If you use the A7 in certain shooting modes, the camera will automatically default to 12-bit color depth, losing 2 stops of dynamic range in the process. Specifically - silent shutter, long exposure/bulb and continuous shooting.

Compact mirrorless cameras are definitely the future, but for now I like to stick with the stuff that just works™, which means it can be as chunky and optical as it wants as if it means consistent performance in all circumstances.