K3ii vs OMD-EM1 vs Foveon


Link Posted 10/11/2015 - 14:55
The "improvement" in the K3ii I find most interesting is the so-called "pixel shift". The reviewers I've read don't seem to think so, but I wonder if they've read all that much about Foveon. I've been reading Foveon articles for a long while, the idea of an alternative to Bayer, the Foveon three-level layers that manage to bring out more and more life-like color has been fascinating.

I woke up when I heard that Pentax was doing something like that with their K3ii, but was it as good as the Foveon original? The typical review comment was that you can get some really good photos but only stills and you must have a tripod on a clear day that isn't windy and nothing is moving -- and wait a long time. Sigma doesn't tell us that about their Foveon cameras. In fact they have a range-finder-like edition. Does Sigma expect users of their DX1 & DX2 to carry a tripod around with them when all they say is that you can carry the DX1 or the DX2 in your pocket and get some excellent shots with them. There is no turning the "pixel shift" (I forget what they call it) off for Foveon.

And then there is the Olympus OMD-EM1 which includes this capability, but in more than 4 layers. How well does the EM1 do? Is it better or worse than the K3ii, or than the Sigma Foveon? I don't know -- can't find a discussion. There should be a concussion somewhere about how well the K3ii with its "pixel shift" stacks up against the Sigma Foveon, or the EM1. If someone knows of such a discussion I would be grateful for the reference.



Link Posted 10/11/2015 - 15:25
The Sigma Foveon cameras have RGB photosites for each pixel (on the 4.6mp and 15mp cameras anyway) so they only take a single image, hence you don't 'need' a tripod to use them.

Both the Pentax 'Pixel Shift' and Olympus 'High Res/Sensor Shift' are different. They only output the interpolated image into a Jpeg file.

The Pentax uses the SR to take 4 separate images and merge them together to sample the Red, Green and Blue pixel information and merge them together to create a 24mp image with pixel level detail and RGB values.

The Olympus uses the Image Stabilisation to take 8 photos, and merge those images together to create a ~40mp image from the 16mp image sensor.

The Sigma cameras (SD9-SD1M and DP1-DP1 Merrill) cameras have a layered sensor that has a Red, Green and Blue sensor for each individual pixel, so taking 3 colour readings for each pixel, and no Bayer Interpolation. You only get a single image, in both the 3XF RAW file format and Jpeg if you so wish.

The newer Sigma cameras, with the Quattro sensor are slightly different.

I have had both the Sigma DP1X with the 4.6mp Foveon Sensor, the Sigma DP3M with the 15mp Foveon Sensor and a K-3II with the Pixel Shift.

I think for pixel level sharpness and RGB colour information, the Sigma Foveon is still the best approach, as you only take a single image so no issues with multiple exposures. I have yet to find a situation to use the K-3II Pixel Shift, other than Product Photography, even then you only get it on Jpeg output, rather than a single RAW image.
D-FA28-105 | FA43 | K55 | DA100 Macro | DA*300
Pentax K1 | Pentax K-01 | Pentax MZ-S | Pentax MZ-5n
Flickr Photostream
Last Edited by hoopsontoast on 10/11/2015 - 15:27



Link Posted 10/11/2015 - 16:13
Thank you, that was very helpful. If the K3ii or EM5 produced anything as good as the Sigma Foveon I might have be tempted, but I take most of my photos hiking and have no room for a tripod -- nor patience since I hike with two dogs. So if I were to get anything along this line it would have to be the Faveon DP1 or DP2. The DP1 at $799 from Amazon is almost tempting enough.



Link Posted 10/11/2015 - 19:09
hum there is no comparison is there because correct me if im wrong but isant one of them mirrorless
and if so they are totally differant


Link Posted 10/11/2015 - 19:22
You've chosen three very different beasts there, so no real comparison would be possible IMHO. Likewise, the idea of the Pixel Shift Resolution is totally different in concept to a Foveon sensor.

Fuji and Sigma have both used different types of sensor to the standard Bayer pattern design, but to be honest the Bayer images look better to me, despite the theory suggesting the others could have advantages.
Best regards, John


Link Posted 10/11/2015 - 20:00
Hmmm. Comparison. Hmmm. Not sure what you call it. All three had a system that purported to produce better color than the simple Bayer system. Did they manage? Did Olympus and Pentax achieve something by their multiple Bayer layers that might render the Foveon complexity and weirdness unnecessary?

With just the brief reviews of the Olympus & Pentax systems that I found, they seemed to be achieving what Foveon had achieved, but I learned that wasn't the case. As Hoopsontoast and some others explained, Olympus & Pentax require multiple Bayer shots with the end result being a JPEG. Sigma's Foveon takes a single shot, hence not absolutely requiring a tripod or a subject that remains still. The end result, however, is a sort of Raw file that only a cranky Sigma software system can manipulate. To get it into Photoshop (if you have Photoshop and if you don't, don't bother) requires that it be converted to a Tiff file first.

So the answers I got tell me that if I want Foveon quality I'm not going to get it from either Pentax or Olympus, but if I do choose a Sigma camera, even if only the Rangefinder-like DP1 (that doesn't require additional Sigma lenses) I'd better begin with a lot of patience and not complain when nothing functions smoothly.



Link Posted 10/11/2015 - 21:12
I tried to comment on this last night on my phone, the page kept throwing an error.

I suspect the pixel shift will be used as much as the simulated AA filter option. In my opinion they are nice to haves, but I doubt they'll get much use by most users.
They seem ideal for macro/static product work, and for selected landscape work where there is essentially zero movement between shots.

stu62 wrote:
hum there is no comparison is there because correct me if im wrong but isant one of them mirrorless
and if so they are totally differant

Not at all really. The mirror relates to how you see the image prior to taking the shot, and for some focussing system types.
When the shot is taken, they're essentially identical, with a mechanical shutter.

Olympus opted to make a higher res photo, but essentially it's just a larger version of the same smaller photo. There is not much difference in image quality at all, however you do get some inherent noise reduction going on when you scale it back down to a smaller size, so there are real benefits in that.

Pentax said the same in one of their interviews, in that yes they could make the image larger in resolution, but it ultimately looked exactly the same and they couldn't see a use for it. Other than printing really large, or heavily cropping, resolution doesn't add much. If you're going to view on a 1920x1080 monitor, well that's roughly 2MP in size, so lets keep it in perspective.

So Pentax went the improved colour route and this does look pretty good. You're limited by the static subject matter requirement, but it looks good when suitable. Maybe there are some examples somewhere to compare these with the Sigma Foveon approach?

Olympus don't seem to be in that game at all, instead focussing on a larger image. Only Pentax are offering something new to compare with the Foveon.


Link Posted 10/11/2015 - 22:20
mr fleur
can you manualy focus a mirrorless camera without switching it on

i think the awnser to that is no
can you manualy focus a mirror camera in without switching it on yes
ie they are nothing alike
Last Edited by stu62 on 10/11/2015 - 22:20


Link Posted 10/11/2015 - 22:29
stu62 wrote:
mr fleur
can you manualy focus a mirrorless camera without switching it on

My point is that has nothing to do with the pixel shift technology in discussion in this thread.

Regardless, when you want to take the photo, you will need to have the camera switched on in either platform.


Link Posted 10/11/2015 - 22:41
Fair enough Mr Fleur.
My Guides to the Pentax Digital Camera Flash Lighting System : Download here from the PentaxForums Homepage Article .... link
Pentax K7 with BG-4 Grip / Samyang 14mm f2.8 ED AS IF UMC / DA18-55mm f3.5-5.6 AL WR / SMC A28mm f2.8 / D FA 28-105mm / SMC F35-70 f3.5-4.5 / SMC A50mm f1.7 / Tamron AF70-300mm f4-5.6 Di LD macro / SMC M75-150mm f4.0 / Tamron Adaptall (CT-135) 135mm f2.8 / Asahi Takumar-A 2X tele-converter / Pentax AF-540FGZ (I & II) Flashes / Cactus RF60/X Flashes & V6/V6II Transceiver


Link Posted 10/11/2015 - 23:54
My feeling is that this is a post for another forum (although the mods may have their own opinion). IMHO if you really want a 3 layer sensor, the Foveon is the only solution, anything else is 'second best' at best. As a film shooter I have three (sometimes four) colour layers to play with.

If you're already a Pentax shooter the question is irrelevant, the difference will be minuscule. If you're undecided about Pentax/OMD/Foveon try them in a shop and pick the one you prefer.

PPG Flickr


Link Posted 11/11/2015 - 08:57
It's been well known since film days that true colours are very DULL and they had to be made more vibrant to be accepted by the public. Pentax colours are generally acceptable straight out of the camera.

You can get Test Cards and software that will correct both RAW and JPG colours to a known standard. e.g. QPCARD


Half Man... Half Pentax ... Half Cucumber

Pentax K-1 + K-5 and some other stuff

Last Edited by Algernon on 11/11/2015 - 09:00


Link Posted 12/11/2015 - 13:25
Read this thread with interest... I've always been impressed by Sigma's persistence in trying to improve on the familiar Bayer arrangement. It's persistence that hasn't yet paid off, in terms of delivering a camera that can be used as flexibly as any Pentax or almost any other camera. Its technical achievement is real -- the quality is there, if the circumstances are right -- but its usability is restricted. The following review is knowledgeable and balanced (it also has some complimentary words about the Ricoh GR in passing!)
Regards, Christopher



Link Posted 12/11/2015 - 13:54
Not sure if my thread on a replacement for the Bayer filter is relevant here but here's the link again just in case .


Link Posted 12/11/2015 - 16:28
Gwyn wrote:
Not sure if my thread on a replacement for the Bayer filter is relevant here but here's the link again just in case .

Very interesting (I'd missed the original thread, Gwyn). As might be expected, the paper itself carries some caveats that are not prominent in the journalist's summary; but even so, the possibilities are intriguing. I see the research was sponsored by NASA -- always good to hear of a potential new entrant into the world of photography
Regards, Christopher

Add a Comment
You must be registered or logged-in to comment.