That Medium Format 'look' - without medium format?

[LastLight]

Partly contradictory, I keep reading about how as cameras progress and pixel counts increase the difference between 'full frame' and medium format (digital) image quality is getting closer and closer. Pundits claiming that cameras like the Sony A7Rii and the newer Canon 5s are approaching the resolution (and image quality?) of cameras like the Pentax 645Z and Leica S.

On the other hand, we'll respected reviewers like Michael Reichmann and Kevin Raber are convinced that there is still a medium format 'look' as well as any print size advantage. And surely there must be an advantage in having larger pixels? After all, even 1" and APS-C sensors can now have massive pixel counts. Which got me thinking...

Would using something like a Sony A7S and stitching give the resolution of higher mega pixel cameras and, with less stressed sensors(?) something of that look they refer to? The lack of harsher transitions, artefacts and that 'waxy' look noted (by some) for digital cameras?

[Simpo Two]

How do you define 'the look'? There's more to it than just resolution I expect.

[JustinP1]

Yes, I agree.

What is the 'look'?

In my day job, the music world, it's commonplace for people to talk about a 'sound' without being able to explain it.

That doesn't mean it doesn't exist, however, often they are affected by the placebo effect of seeing the esoteric equipment used.

[fido]

You won't get the level of detail required for that medium format look with an A7S with 12MP.

But with an A7RII and a very good lens capable of resolving to that level of detail e.g. Leica APO or Zeiss Otus then you might get that look IMO.

[LastLight]

I'm not sure, I was just musing really and wonder what others felt. They often remark, sometimes in the same review - see the latest "Michael and Kevin's Toybox" video on Luminous Landscape - that the resolution and image quality of even smaller than full frame cameras is great and getting better (like their rave review of the Canon GX3), but then still talk about the 'look' and the benefits of a larger sensor.

I think they and aficionados of Leica S, Phase One, Leaf, Pentax etc. on various forums, talk about a different (and better) look from a bigger sensor and bigger pixel pitch; not so much resolution but in colour and tone and smoothness (without losing detail), the sort of thing oldies like me used to see when moving from 35mm film to 4"x 5" etc.

I'm not sure, but having only moved to digital a year ago (Nikon Coolpix, then Sony A6000, then A7ii and now 7Rii) find it difficult to explain, but think there is a harsher, harder look to the Coolpix and A6000 pictures whatever processing I do - more noticeable in colour than bw - and initially thought I'd never get on with digital as I didn't like the way edges/details etc. went when the resolution was exceeded/trying to print (too) big. It doesn't bother me so often now, but I don't know whether that is just familiarity or the higher resolution of the new camera? Instead of just a general blurring, smoothing and lack of detail/focus, more a jagged blocky look.

Ironically many prints even from film are now scanned and printed on ink jet printers so that must complicate things. I've a number of prints from film and digital sources/printers and I can usually tell which came form what. I was quite disappointed, for example, with some of the digital prints at the recent Masters of Vision show where grasses. twigs etc. looked 'bitty'. Yet the two prints that I got with the Paul Kenny book (and the printing of the book itself) done by Jack Lowe on a HP digital printer don't seem to suffer and look beautifully smooth, rich and still detailed.

[RobDickinson]

I hang out on LuLa forums a bit and have looked at a lot of MF images.

I really dont get it. Sure if you need the absolute best quality in a single frame then MF is the way to go. But the 'look' is down to the lens and processing IMO, not the sensor size.

In fact the current latest MF sensor is effectively 2 24mp a7s sensors bolted together. Comparisons to the old CCD and SLR stuff shows pretty much what I think. No 'special sauce'.

MF gets shallow DOF but doesnt have fast lenses, so the sensor is 1.7 times bigger but the lenses are f2.8 or slower.

[ExPat2B]

The medium format look is result of several things.

1. Physical : Longer lenses, and smaller apertures, and tilt shifting of view cameras.. This gives massive depth of field without diffraction, and wide angles without stretched corners.

2. Chemical. Highlight rolloff and its character is key to how we percive color and tone and film naturally has great highlight range. Digital just blows straight to white and looks very clinical. With skill you can make digital look good but it doesn't come easy.

3. Optical. Bigger piece of glass per sq mm of output medium means better colour, contrast and sharpness.

[LastLight]

In a way, that is it - the idea that MF is like 1.n or whatever times the sensor size of the smaller cameras, so for the same pixel count has larger pitch sensors (less strained?), so given the ability to stitch, 2, 3, 4 etc. frames form something like the A7 would match that resolution, detail and noise levels/colour/gradations/tones the 'experts' claim for the MF cameras.

I'm guessing there must be a reason why the A7s with it's 'big' pixels is so well liked, or why so many gurus claim this special quality of look for MF?

The main reason for the curiosity is that I've bought a Cambo Actus and mounts for the A7Rii. It can mount my Leica 100mm macro lens, and many others, and I have 2 Schneider lenses for it - a 47mm f5.6 Apo Digitar XL and a Schneider 120mm Apo Digitar Aspheric and want to use these with the Sony as a faux digital back. I was wondering if I added a second body whether one with fewer pixels crammed on the sensor, like the 7S, might be a good choice.

[RobDickinson]

No, basically.

The A7s is good at stupid high ISO, nothing more. There is no special measurable quality advantage in its larger pixels at anything below ISO 12800. And even then I think sony are pulling a fast one (NR on raws etc).

Smaller pixels (within reason) just gives you more detail (potentially).

If you took a 3 frame pano with the 24mp sony a72 or D610 etc and stitched (so long as the subject let you) it should look near identical to a 50mp MF shot if lenses are similar quality. I have seen this work out in reality.

Oh and I am currently stitching 50mp canon 5Ds files....

[DavidY]

LastLight said:

The main reason for the curiosity is that I've bought a Cambo Actus and mounts for the A7Rii. It can mount my Leica 100mm macro lens, and many others, and I have 2 Schneider lenses for it - a 47mm f5.6 Apo Digitar XL and a Schneider 120mm Apo Digitar Aspheric and want to use these with the Sony as a faux digital back. I was wondering if I added a second body whether one with fewer pixels crammed on the sensor, like the 7S, might be a good choice.

Unless you want to shoot in low light (handheld) and make use of the better ISO range of large pixels, then as Rob says no benefit.

You would gain more benefit for monochrome shooting using a dedicated mono sensor (such as the Leica Monochrom) but these are few and far between and cost ££££.

Lots of rumours flying around that the next Fuji XPro2 will be a MF rangefinder, would fit into Fuji's heritage but be a big step, as lots of new glass would be required, unless they made adapters from the off.

IMHO the biggest advantage of using MF is that the bigger the camera/lens, the slower you tend to operate so more care is put into the original photograph (but you can simulate this with slower sensors, just be more methodical!!)

[Andy M]

MF has a different depth of field to 135mm, often more pleasing to the eye even when using like-for-like focal lengths.

However, I reckon 99.9% of people wouldn't be able to tell the difference between MF and a shot taken with a A7r2/5Ds/810 + Zeiss Otus 85mm.

[RobDickinson]

Andy M said:

MF has a different depth of field to 135mm, often more pleasing to the eye even when using like-for-like focal lengths.

MF ys has shallower dof for the same focal length/aperture. But has few fast lenses compared to 35mm systems, and on the whole the digital MF sensors are not full sized so the advantage is somewhat less also.

Oh and its 135 format or 35mmm

[LastLight]

DavidY said:

LastLight said:

The main reason for the curiosity is that I've bought a Cambo Actus and mounts for the A7Rii. It can mount my Leica 100mm macro lens, and many others, and I have 2 Schneider lenses for it - a 47mm f5.6 Apo Digitar XL and a Schneider 120mm Apo Digitar Aspheric and want to use these with the Sony as a faux digital back. I was wondering if I added a second body whether one with fewer pixels crammed on the sensor, like the 7S, might be a good choice.

Unless you want to shoot in low light (handheld) and make use of the better ISO range of large pixels, then as Rob says no benefit.

You would gain more benefit for monochrome shooting using a dedicated mono sensor (such as the Leica Monochrom) but these are few and far between and cost ££££.

Lots of rumours flying around that the next Fuji XPro2 will be a MF rangefinder, would fit into Fuji's heritage but be a big step, as lots of new glass would be required, unless they made adapters from the off.

IMHO the biggest advantage of using MF is that the bigger the camera/lens, the slower you tend to operate so more care is put into the original photograph (but you can simulate this with slower sensors, just be more methodical!!)

Ironically I sold my M246 Monochrom last week, it was ace - every bit as sharp to my eye as the 7Rii to my eyes, but I couldn't justify the expense when software can get such good bw conversions.

The reason for the Actus is mainly to slow me down and get me to work more like I did with the Linhof as I've been struggling to avoid the tendency to rush and shoot too many pictures, just because I can!

So, you, Rob et al feel that it is resolution (inc. lenses) and pixel count rather than size of pixels, which is reassuring really, maybe the gurus do have some subjective bias and see what they want to be there?

[DavidY]

The colour filter on any colour DSLR reduces the spatial resolution of monochrome images to a quarter. So if you are using a camera for accurate measurement in monochrome, then to have the equivalent in a colour sensor after monochrome conversion you would need a sensor with 4 times the resolution!!! (I work in industrial machine vision - most of my work is in monochrome with industrial GigE cameras resolutions of varying resolution, though a surprising amount is still done at 640x480 especially if speed is of the essence, I did one system for a customer where we were processing the images and deriving a go/no go result 62 times a second!!!, but most of the stuff I do is at higher resolution but more complex inspections).

But, if you are not after ultra sharp images, then a colour sensor converted to monochrome will provide excellent results.

There is far more to photography than amazing technical equipment, as long as your equipment will meet your needs, then the rest is down to you. The art of photography is as still valid today as it has always been, and although machine gunning will fluke the odd good shot, proper technique and composition still counts.

[RobDickinson]

I may be a philistine though, I dont 'see' the Zeiss 3d look etc either. But hey whatdoIknow.

[DavidY]

RobDickinson said:

I may be a philistine though, I dont 'see' the Zeiss 3d look etc either. But hey whatdoIknow.

Its not a 3D look per se, but different coatings on lenses give different colour renditions, Zeiss are known to make certain colours 'pop' and some term this 3D.

It's not right and it's not wrong, its just different!!!

[LastLight]

RobDickinson said:

I may be a philistine though, I dont 'see' the Zeiss 3d look etc either. But hey whatdoIknow.

Despite being such a fan of Steve Huff?!

[LastLight]

DavidY said:

The colour filter on any colour DSLR reduces the spatial resolution of monochrome images to a quarter. So if you are using a camera for accurate measurement in monochrome, then to have the equivalent in a colour sensor after monochrome conversion you would need a sensor with 4 times the resolution!!! (I work in industrial machine vision - most of my work is in monochrome with industrial GigE cameras resolutions of varying resolution, though a surprising amount is still done at 640x480 especially if speed is of the essence, I did one system for a customer where we were processing the images and deriving a go/no go result 62 times a second!!!, but most of the stuff I do is at higher resolution but more complex inspections).

But, if you are not after ultra sharp images, then a colour sensor converted to monochrome will provide excellent results.

There is far more to photography than amazing technical equipment, as long as your equipment will meet your needs, then the rest is down to you. The art of photography is as still valid today as it has always been, and although machine gunning will fluke the odd good shot, proper technique and composition still counts.

Yes I know it isn't mainly about the gear, though the right equipment at the right time certainly helps, but I do like to see, or have made, big prints from time to time and some done on the Nikon and A6000 went too far and ended up a bit pixellated and jaggy on the edges branches etc. On film they'd have just got soft then blurry.

Looking back, of all the photos I've sold or won competitions with 90% were on 6 x 9cm or 6 x 12cm, or 4" x 5" slide film, despite taking way, way more 35mm shots. I think the time needed helps, excepting where sometimes time is of the essence of course and it's now or ever.

[RobDickinson]

LastLight said:

went too far and ended up a bit pixellated and jaggy on the edges branches etc. On film they'd have just got soft then blurry.

Then the file wasnt processed properly for the print size.

[LastLight]

RobDickinson said:

LastLight said:

went too far and ended up a bit pixellated and jaggy on the edges branches etc. On film they'd have just got soft then blurry.

Then the file wasnt processed properly for the print size.

I sent them as was - with minor tweaks in Lightroom (didn't have Photoshop then) - to Whitewall who did the prints and had noted that they would sort out appropriate processing for/advise on the size suited by the original. They weren't very high res (looking back, 2 were from an iPad) and included a couple of crops, but surely there is a limit however relevant the processing is?

I've now got an Epson 3800 to do my own and am (still!) part way through an article by Tim Parkin on processing and sharpening for different print sizes.

It's not that I always crave resolution or want big prints, just that this topic interested me and had me musing after reading what some say, like the aforementioned Huff when he first got all enthusiastic about the 7S.


BTW - are you using the 5Ds now? If so how do you find it?