The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

IQ180 vs Up-Res A7R2

fmueller

Active member
My comparison is between a Credo 60, tech cam, rodenstock 40 and 70 as compared to an A7r2, Loxia 50, Voigtlander 35 1.2 v2.

The Sony can be easily carried and used handheld and has significant low light capability and image quality is quite good.

When I have the time, inclination and opportunity to plant a tripod, the Credo 60 wins by a large margin but at significantly greater cost.

And with all that I recognize the two kits serve vastly different purposes to this amateur enthusiast.
 

Jack

Sr. Administrator
Staff member
Re color fidelity:

Not withstanding the approximate 0.25 stop more exposure to the IQ image, the lighter green leaves in the Sony image lean toward brown while the IQ light green remains in the greens; the pinks in the rose show far more range in the IQ image than they do in the Sony. Finally there is obvious superiority in the rendering of the "whites" of the small flowers, pot and background bricks and mortar. What I'm seeing isn't due to resolution, it's due to the sensor's ability to handle color and DR. Now which is more accurate color, tough to say without a color checker in the image, but I will bet the cost of an A7rII on the IQ180 :D
 

ShooterSteve

New member
I'd like to see the comparison with different lenses. There seems to be a huge difference in micro contrast but wonder if he had tried some other optics, like a Leica R or M, Zeiss etc. I have a set of Leica R's and am used to seeing wonderful textures that are not very evident in the first shot. Maybe it's the sensor, maybe it's the lenses...
 

rhern213

New member
I highly doubt it's the lenses.

On the A7R2 I used the 55 1.8, I don't think there's anything adapted from Leica or any other brand that will outresolve that lens to make any difference.

And for the IQ180 I used the 80 LS to match the same distance/perspective/compression as much as possible. So it's arguable that I could've gotten even better results from the IQ180 had I used the 150 2.8.


I just picked up a Mamiya 120mm Macro with manual aperture control. So I'm going to re-do the test with that same lens on both bodies to see if I get any different results.


I'd like to see the comparison with different lenses. There seems to be a huge difference in micro contrast but wonder if he had tried some other optics, like a Leica R or M, Zeiss etc. I have a set of Leica R's and am used to seeing wonderful textures that are not very evident in the first shot. Maybe it's the sensor, maybe it's the lenses...

I don't think there's much to compare here. The CMOS sensor in the A7R, D810, etc... had better dynamic range than the IQ180's CCD sensor. The IQ180 of course has cleaner images at base ISO, but the A7R2 should be at minimum maybe 1 stop or so better in DR.

I would really love to see DR comparison,
it will be great if someone can show side by side.

Thanks
 
Last edited:

torger

Active member
Re color fidelity:

Not withstanding the approximate 0.25 stop more exposure to the IQ image, the lighter green leaves in the Sony image lean toward brown while the IQ light green remains in the greens; the pinks in the rose show far more range in the IQ image than they do in the Sony. Finally there is obvious superiority in the rendering of the "whites" of the small flowers, pot and background bricks and mortar. What I'm seeing isn't due to resolution, it's due to the sensor's ability to handle color and DR. Now which is more accurate color, tough to say without a color checker in the image, but I will bet the cost of an A7rII on the IQ180 :D
We're probably not going to agree about the hardware vs profile issue, so I'll leave that.

I've seen over and over again that when images are scaled down, to remove the resolution advantage, and then blind tested things that were "easy" to see before suddenly becomes very hard. I've seen Sony sensor's go from plasticky to fantastic just by ending up in an IQ250. I've seen the P45+ go from fantastic to mediocre, just by showing how far behind it is in DR. There's a lot of bias in this world. I'm not immune myself either though, as I don't think there is a hardware difference I'm biased towards not seeing any of course. My standpoint is clear enough so I shall not elaborate further.

One thing could be stated for sure though -- Capture One's profiles are not about color accuracy at all. This is not a defect, it's just a design choice. They provide canned looks. And even if they would have chosen accuracy, it can only exist with a linear curve (as you use in reproduction). As soon as a curve is applied subjective design is necessary, as color appearance and contrast are tightly coupled. That cameras make so different looks is only partly due to sensor color filters and lenses, it's very much a profile design choice. Just see how different a Credo 50 and IQ250 looks, despite using the exact same sensor.
 
Last edited:

ErikKaffehr

Well-known member
Hi,

My take is really that once we upsize an image we are loosing image quality. More pixels can be a real advantage, once they are needed. So I don't think a 40 MP device will be able to replace an 80 MP, ever. That is if those pixels are needed. My experience is with DSLRs from 6-24 MP and 39 MP MFDB. What I can see is that something like 12 MP is good enough for A2-size prints. That is what I normally make. At that size (around 16 x 23"), I don't see any benefits from 39 MP MFDB, going larger I would say that the pixel resolution is pretty obvious.

Regarding colour I would say it is 90% colour profiles and 10% sensor technology. Initially I felt that Adobe/Sony colours were much more accurate (measured as DeltaE) than Capture One with P45+, but I have later found Capture One has a quite aggressive "film curve" as default. Using linear profile Capture One is quite accurate (in DeltaE terms).

Colour reproduction is much to taste, I would say that Adobe colours are a bit conservative while Capture One is more intensive. One thing I noticed is that C1 can push some very real colours (like blue) wide outside Adobe RGB. You of course don't see it on a screen as colour reproduction is clipped to either sRGB or Adobe RGB.

Best regards
Erik
 

Shashin

Well-known member
Uprezing does not change image quality. It simply smooths any pixelation that might be apparent in a large print at short viewing distances. Just as there is no detailed added by adding pixels, there is also none taken away.
 

torger

Active member
Are you saying Sony has better DR? Have you push/pull IQ file before? It will be nice to show samples.
That particular exercise have been done to death. Yes the Sony CMOS have had better DR than the Dalsa CCDs since the days of Nikon D800. I have an old example showing IQ180 vs D800, it's not my pictures though so I can't share, but it shouldn't be that hard to find surfing around a bit on this forum and/or LuLa.

One thing that makes it a bit difficult to compare is that C1 applies noise reduction even when noise reduction is disabled. A trained eye can detect noise reduction and their negative effects though (smearing, brownish pastel-looking colors). Users in general seems to be blind to that though which has caused many to believe that MF cameras has much better DR than they actually have. Hasselblad's Phocus does the same, my relatively noisy Kodak sensor looks really clean in Phocus, but I do see the adverse effects from noise reduction.

If you want to see "the truth" you can open up files in say RawTherapee which doesn't have noise reduction on unless you specify it.

That said the IQ180 has very good DR. My H4D-50 has a Kodak sensor which is at least a stop behind and I do well with that, but I do use grad filters from time to time, in a really complicated situation I may bracket and merge. It depends on your shooting style and subjects how dependent you are on the DR aspect. I happen to like soft light which does make it easier on the camera.

I certainly don't think the small DR difference between the IQ180 and a Sony A7RII would be a selling point in any case. If you're into long exposure though, then it's a different story.
 

torger

Active member
Sorry for the off-topicness coming in this message, it is about color fidelity though, and I thought I'd share some of my recent findings of the internals of C1 which some of you might be interested in; I've been working with my profiling project (which does both DNG and ICC profiles) and some of my users want to use it in C1 so I had to dig into it whether I like it or not ;)

I've been quite critical against C1 for using a simple RGB curve as their tone curve, and letting you change curves without changing profiles. As an S-curve RGB curve adds loads of saturation and shifts hues it's not exactly what you would think would be used in a high end raw converter. C1, like most established raw converters, was created in the 1990s though when computers were a lot less powerful than today so using simple algorithms like RGB curves was natural and this legacy is still in.

However, C1 designers have known what they're doing, the RGB curves found in the film curves are never a pronounced S, and will thus only cause small hue shifts. However such curves does not provide that much contrast either, in particular the shadow range becomes light and the overall look a bit dull. This has been solved by adding an additional fixed contrast curve in the ICC profile lookup table (LUT). By having it in the lookup table they can have used any type of curve algorithm processed offline, my guess (without analyzing in detail) is that they've used some sort of luminance curve to avoid hue shifts. This split approach to contrast makes the profile work better across several film curves.

However, it does of course also mean that if you apply "Linear Response" you still have the residual S-curve that the ICC applies, so it's not by any means an accurate colorimetric mode. The change of curve is still not totally immune to hue shifts either, and one can assume that the ICC profiles are optimized to look the best with the default curve.

I've mostly worked with P45+ as the test camera though, so I can't say if all C1 profiles work this way, but I'd guess so.
 

Jack

Sr. Administrator
Staff member
You do know that you can tweak C1's profiles in their advanced color editor and save them as camera-based pre-set ICC's?

I did this with my P65+ and IQ180 for studio flash to obtain near perfect color. It was not a feature of C1 in the days I owned the P45+, so I did not use it then.

Overall, I have found C1's base "flash" ICC profiles to be extremely accurate as re studio color, though admittedly not always perfect. But they also note that flash CT is not consistent enough and why they allow the tweaking in advanced CE.

You can call up any Phase ICC profile inside C1 -- in fact they should be auto loaded inside your system -- so you could delve into the LUT to see if the newer cam ICC's are Luminance-curved...
 

PSon

Active member
Uprezing does not change image quality. It simply smooths any pixelation that might be apparent in a large print at short viewing distances. Just as there is no detailed added by adding pixels, there is also none taken away.
Not taken away but perhaps diluted?
 

ErikKaffehr

Well-known member
Hi,

Think about one bottle of wine. You need two bottles of wine, so you either buy another bottle (stitch three images), or you add a bottle of water, mix, and pour up in two bottles...

When do dilute the wine it keeps it's ingredients, but it is not the same as another bottle.

Best regards
Erik

Perhaps you can explain...
 

Shashin

Well-known member
Hi,

Think about one bottle of wine. You need two bottles of wine, so you either buy another bottle (stitch three images), or you add a bottle of water, mix, and pour up in two bottles...

When do dilute the wine it keeps it's ingredients, but it is not the same as another bottle.

Best regards
Erik
The analogy has a flaw. The volume does NOT change. Think of a cubic meter. It has 2 million pink molecules and 1 million green molecules. However, you sampling device (eye/viewing distance) is too fine and so does not always sample 2:1 pink to green molecules (pixelation). You add more pink and green at a ratio of 2:1. The color of the gas does not change. The pressure, the file size, does, but the appearance remains the same and the sampling is smoothed so you collect the 2:1 ratio of gases (pixelations is eliminated).

Naturally, we do not need to make analogies. It is easy to describe what is happening with the image. This is why the wine dilution model does not work--you are not adding blank pixels (water), but pixels that are related to the wine. Averaging is not dilution.
 

narikin

New member
Like many here, I have both A7R2 and Phase 1Q 80mp system (and canon 5DSR, and, and...)

Even with the best lenses (Otus) on A7R, it is nowhere near the resolution or color of the MF camera. For sure, it can do things that my Alpa/Phase system cannot - hi ISO, great on chip AF, etc - and if that is what you need for the job, then it's perfect! But if you are shooting slowly in regular daylight, then no question, the MF system is much better.

I love the A7R2, and it's amazing technological developments, but can it replace my MF system? No, it cannot. If you want large super sharp images, it still has to be MF.
 

f8orbust

Active member
I can contribute one more comparison file ... tests & comments appreciated ... Christoph
Thanks for posting the RAWs - it just goes to confirm that 2015 is a great time to be a photographer.

As I see it, there are 3 winners: 1) The IQ180, on resolution; 2) It's pretty much a tie on color, but given that the A7RII costs 10% of an IQ180, I'll give it to the A7RII, and 3) The Canon TSE17 - what a nice lens.

Jim
 
Top