A load of old pixel shift. Why I just don't care for high-res modes(dpreview.com)
dpreview.com
A load of old pixel shift. Why I just don't care for high-res modes
https://www.dpreview.com/opinion/6915548723/a-load-of-old-pixel-shift-why-i-just-don-t-care-for-high-res-modes
9 comments
Professional product photographer here, primarily jewelry. Can confirm, pixel shift is a thing of pure beauty in this application, at least on the Hasselblad medium format system our studio used for years. Was a pain to wait for 16 captures to fire off, we certainly didn’t use it every day, but for high end work it gave beautiful tones and pixel resolution. I’ve never tried such a thing on landscapes, I can see the reasons to avoid it in that case for sure!
I just bought an OM-D EM-5 MKII used from a local shop - the first digital camera of my adult life (I've been shooting film as a hobby for years). It has this feature. I used it a couple times and abandoned it, for the same reasons the author lays out.
The folks running the shop are obviously huge camera geeks, and they were really psyched about that feature. To their credit, it's damn impressive that a camera that small, and a sensor that small, can produce a true 40MP image where every pixel actually captures a geometrically unique location on the focal plane.
As a key selling point over an E-M10 though? Nah. It's a marvel of engineering, but extremely limited in its practical value - a hallmark of "features for nerds." Sadly, in today's consumer landscape you kinda have to satisfy the spec sheet gawkers if you're going to sell a product. That's another reason features like this exist.
16MP on a sensor of that quality can still be blown up to enormous sizes without it looking crunchy. The real reason to spring for an E-M5 is the weather sealing and bigger viewfinder.
The folks running the shop are obviously huge camera geeks, and they were really psyched about that feature. To their credit, it's damn impressive that a camera that small, and a sensor that small, can produce a true 40MP image where every pixel actually captures a geometrically unique location on the focal plane.
As a key selling point over an E-M10 though? Nah. It's a marvel of engineering, but extremely limited in its practical value - a hallmark of "features for nerds." Sadly, in today's consumer landscape you kinda have to satisfy the spec sheet gawkers if you're going to sell a product. That's another reason features like this exist.
16MP on a sensor of that quality can still be blown up to enormous sizes without it looking crunchy. The real reason to spring for an E-M5 is the weather sealing and bigger viewfinder.
I think these 4/3rds camera were dismissed by armchair photographers in the early years because of the sensor size, and I think these pixel shift type features were added so that they could add a higher resolution mode to the headline features of the camera, to compete with higher resolution sensors.
However, as you have probably found, 16mp is actually a very useful size - a good compromise between image size and resolution. I actually shoot a 16mp camera (Fuji X1) most of the time, even though I have much higher resolution cameras to hand.
However, as you have probably found, 16mp is actually a very useful size - a good compromise between image size and resolution. I actually shoot a 16mp camera (Fuji X1) most of the time, even though I have much higher resolution cameras to hand.
It was never about resolution. 4/3s was dismissed because it’s very hard to get nice depth / bokeh, poor ISO, and great lenses were rare. It works great for niches like wildlife, news, street photography, not so for others.
I owned an E-M10 for a while and ended up selling because it had nothing much to distinguish itself from my iPhone, and with all the computing power the phone photos were sharper and with better dynamic range. Portrait mode also evolved to do a pretty decent job for family pictures.
APS-C is at the sweet spot of enabling very small bodies and lenses, but still getting FF-like images with a fast lens and decent low-light performance. My current Fuji kit is actually lighter than the old MFT one and about the same size, while taking far better pictures.
I owned an E-M10 for a while and ended up selling because it had nothing much to distinguish itself from my iPhone, and with all the computing power the phone photos were sharper and with better dynamic range. Portrait mode also evolved to do a pretty decent job for family pictures.
APS-C is at the sweet spot of enabling very small bodies and lenses, but still getting FF-like images with a fast lens and decent low-light performance. My current Fuji kit is actually lighter than the old MFT one and about the same size, while taking far better pictures.
Bokeh and depth are entirely a product of the properties of the lens. Among the great things about m4/3 is the tiny sensor, and tiny distance between the lens and the sensor, which together facilitate a low barrier to entry for high-quality fast glass.
Sure, take APS-C and get rid of the mirror and you get the same effect, but you still have physically larger lenses, significantly higher costs, and a whole lot of pixels that are doing literally nothing unless you're regularly making large prints.
Your whole comment reads like "what are ya, POOR?"
Sure, take APS-C and get rid of the mirror and you get the same effect, but you still have physically larger lenses, significantly higher costs, and a whole lot of pixels that are doing literally nothing unless you're regularly making large prints.
Your whole comment reads like "what are ya, POOR?"
There are few MFT lenses providing great bokeh, they cost thousands of euros.
I don’t understand how you could take my comment that way. MFT gear is not cheap at all. APS-C these days can be physically similar in size, and cheaper while having better image quality. Hence why I think there is little reason to choose four thirds.
I don’t understand how you could take my comment that way. MFT gear is not cheap at all. APS-C these days can be physically similar in size, and cheaper while having better image quality. Hence why I think there is little reason to choose four thirds.
>very hard to get nice depth / bokeh
F1.2 lenses are quite widely available for micro 4/3 systems, which is equivalent to f2.4 on full frame in terms of depth of field. Not enough to get you absolutely razor thin depth of field, but perfectly sufficient to blur the background.
F1.2 lenses are quite widely available for micro 4/3 systems, which is equivalent to f2.4 on full frame in terms of depth of field. Not enough to get you absolutely razor thin depth of field, but perfectly sufficient to blur the background.
That wasn't the case for a long time, and good lenses are extremely expensive, in the 1000+ euro range. Laowa / 7Artisans etc were not a thing until a few years ago. Plus bokeh quality is subjective and not just about blurring the background.
Not saying you can't enjoy an MFT system. My 20mm/1.7 Lumix took great pictures. Truth is the original premise has not held. The best MFT setups are just as heavy as APS-C while having slightly worse specs due to the smaller sensor, and at the same time a higher price tag.
These are very real limitations and much more than "dismissed by armchair photographers in the early years".
Not saying you can't enjoy an MFT system. My 20mm/1.7 Lumix took great pictures. Truth is the original premise has not held. The best MFT setups are just as heavy as APS-C while having slightly worse specs due to the smaller sensor, and at the same time a higher price tag.
These are very real limitations and much more than "dismissed by armchair photographers in the early years".
I have an old manual Nikon 50mm F1.2 with adapter on my E-M5. If you shoot portraits, you run into the "your eyes are in focus but not the tip of your nose" problem. Surely you don't need any shallower DoF than that.
That's 100mm/f2.4 equivalent, of course you get razor thin DoF - even on a 100mm/f4 that's a handful cm at portrait distances.
Maybe I shouldn't have said 'hard to get', cause you certainly can get decent DoF like that, but it's a real physical limitation. You won't get it that easily with classic 28mm/35mm/50mm focal lengths. And there will never be a practical f0.6 lens.
Maybe I shouldn't have said 'hard to get', cause you certainly can get decent DoF like that, but it's a real physical limitation. You won't get it that easily with classic 28mm/35mm/50mm focal lengths. And there will never be a practical f0.6 lens.
F-mount lenses with a speed booster on a micro four thirds body is a seriously underrated setup.
Yeah. Armchair photographers fighting over cameras by looking at the specs and comparing photographs under a microscope. There’s a tradeoff here between making a small, portable, inexpensive camera that you want to carry with you, and making a bigger, clunky, expensive camera that takes high-quality pictures. Armchair photographers looking at the specs will always pick the bigger, more expensive equipment.
I know some people that bought into the 4/3 system, and I personally bought into larger systems. At the time, maybe that made sense for me. But if the camera is smaller, you actually bring the damn camera with you and get a shot. The 4/3 is a beautiful compromise where the shots are probably going to be noticeably better than your phone, and you’re going to bring the camera a lot more often than the APS-C or full-frame cameras.
The resolution differences are on a square-root scale when you count megapixels. 24 MP has 22% better resolution than 16 MP — √(24/16). There’s a whole complicated story here involving aperture size, optical resolution, sensor efficiency and all sorts of other factors, which means that in the end, you don’t sacrifice as much image quality as you expect when switching to a smaller camera.
I know some people that bought into the 4/3 system, and I personally bought into larger systems. At the time, maybe that made sense for me. But if the camera is smaller, you actually bring the damn camera with you and get a shot. The 4/3 is a beautiful compromise where the shots are probably going to be noticeably better than your phone, and you’re going to bring the camera a lot more often than the APS-C or full-frame cameras.
The resolution differences are on a square-root scale when you count megapixels. 24 MP has 22% better resolution than 16 MP — √(24/16). There’s a whole complicated story here involving aperture size, optical resolution, sensor efficiency and all sorts of other factors, which means that in the end, you don’t sacrifice as much image quality as you expect when switching to a smaller camera.
It's a funny thing about the small bodies. My Olympus is too small to handle. I use both the optional grips to make it a decent size.
I made a heftier grip for my E-M5 just using black Sugru. Textured it a bit so it looks damn near stock, it's held up well for six years now.
My other regularly-used cameras are an SRT-102 and an RB67. The smallness is refreshing.
I always imagined that after they added the voice coils for body stabilization, the high resolution shooting modes could be added at no cost, so they just added them.
I started shooting on a 12mp sensor after spending most of my time on 24mp (Sony A7 II to a Nikon D700) and I was surprised at how much detail was in the 12mp images. Unless you're cropping really heavily or printing really large, you really don't need the resolution most of the time.
Canceling out the Bayer pattern to get full color is clever but I’m more excited for the release of Sigma’s full frame Foveon camera with their X3 sensor that doesn’t use the Bayer method. https://en.m.wikipedia.org/wiki/Foveon_X3_sensor
I borrowed a Sigma DP1 which has this X3 sensor a long time ago. The colors are unbelievably good even today when viewed on a 4K OLED TV. However it was not without its problems, the photo quality degraded very quickly once you increased the sensitivity above the base ISO, and the autofocus was quite slow. Still, it was worth it for the amazing photos it produced.
is there some tangible news that it's coming?
“In [a Feb 2024] interview with PetaPixel, Sigma’s CEO Kazuto Yamaki explains that it hasn’t moved past stage two of three in the sensor’s development — a stage it has remained in since 2022.”
https://petapixel.com/2024/02/22/sigmas-full-frame-foveon-ca...
https://petapixel.com/2024/02/22/sigmas-full-frame-foveon-ca...
I would have expected pixel shift to be more useful in studio settings, e.g. product photos, macro photos, etc. This article doesn't seem to mention those.
In general images are often limited by depth of field, camera movement, subject movement or lens resolving power. Some of these can be improved upon with better software. I have been working on software to do this myself as for fun. For the rest of this, though, there are other techniques that will work better than pixel shift.
48MP sensors will be the spot on APS-C and larger especially. High resolution when you need it, drop to 12. otherwise
These are not exactly the type of shots that would benefit from increased resolution.
I don’t know what happened to the Sony images but they look soft even for a standard shot, don’t think they are representative of what the technology can do either. The Panasonic one has a lot more detail.
I don’t know what happened to the Sony images but they look soft even for a standard shot, don’t think they are representative of what the technology can do either. The Panasonic one has a lot more detail.
Back when I was deciding on my first DSLR, I deliberately chose the Nikon D40, at 6 megapixels, over the new D40x, which had 10 megapixels. I did this because I wanted the larger pixels, and thus better low light performance, of the former.
I wore the shutter out on that camera, it was awesome.
I wore the shutter out on that camera, it was awesome.
Yeah I don't know what they are going on about. They sort of dismiss the Olympus implementation at the beginning, by mentioning it's useful and one of the best. The only universal truth in this article is that the author doesn't want to shoot static scenes.
Having tried this feature on my own Olympus camera, I'm not enamored with it either. It's only really good for a pretty narrow selection of "static" scenes.
Light breeze comes through as you take a landscape shot? Welp, I guess that tree branch is going to look like a glitch in the matrix.
It's at its best indoors on a tripod. Architecture and still lifes are where the Olympus implementation really shines.
Light breeze comes through as you take a landscape shot? Welp, I guess that tree branch is going to look like a glitch in the matrix.
It's at its best indoors on a tripod. Architecture and still lifes are where the Olympus implementation really shines.
I think their point is quite clear: very few things are actually static enough for the feature to work properly.
If, however, I was a photographer shooting still life in a studio, for example, product photography for a high end jewellery company, then this is exactly the sort of technology that will work, and will give a resolution boost that may be useful. With a tethered camera, and possibly spending hours getting the light and reflections just right, the extra time to get the shot is of no consequence. Older digital photographers in that market will have used medium format scanning backs, so it's pretty clear that they have patience and that ultimate image quality may trump convenience.
The comments about beyer sensors is interesting, and may be something that the casual photographer doesn't realise. Camera sensors have a patter of red/green/blue filters over each site on the sensor, so if the resolution of your camera is 16mp, you'll have 16 million sensor sites, but each will only give you the level of 1 channel (red, green or blue) and the camera will interpolate the colour resolution across the image. By shifting the sensor across exactly the width and height of 1 pixel site, the camera can capture the correct colour information at the same resolution as the output image, and hence give better colour rendition.
As someone who normally photographs people, and often converts to B&W, this lack of colour resolution isn't important to me, but there are other people where this might be a significant concern.