Mirrorless design

[Offler]

The only real reason to use a large sensor camera is if you wish to print at larger sizes for framing. If your only going to display images digitally, there really isn't much need anymore for anything more than a camera phone. Unless it's to future proof or some kind of specialized imaging like sports, wildlife or astro

While if you prefer to take photos using standard control with manual focus, exposition and aperture control, get a small camera with replaceable lenses.

There is a lot of myths surrounding APS-C and M43 systems regarding "less light" where people who dont know how optics work claim that aperture has to be adjusted by crop factor (not it does not) or that smaller sensor size means "less light", while light intensity is exactly the same.

The only difference is that 25mm M43 vs 50mm FF will have different perspective and bokeh at F2,8 for 25mm M43 will be about same as F5,6 at 50mm FF.

 

[Destination: Moon]

While if you prefer to take photos using standard control with manual focus, exposition and aperture control, get a small camera with replaceable lenses.

There is a lot of myths surrounding APS-C and M43 systems regarding "less light" where people who dont know how optics work claim that aperture has to be adjusted by crop factor (not it does not) or that smaller sensor size means "less light", while light intensity is exactly the same.

The only difference is that 25mm M43 vs 50mm FF will have different perspective and bokeh at F2,8 for 25mm M43 will be about same as F5,6 at 50mm FF.

Not sure what your trying to say. Bokeh is a hazy concept that describes the way a lens renders out of focus elements. It varies from lens to lens of the same focal length but of differing design. And of course it's aperture dependent. Do you mean Depth of Field?

As for light gathering. It is dependent on a bunch of things but it's mention regarding sensor size relates to pixel size. A sensor with larger pixels can accumulate more photons in the same time span. So if 2 sensors have the same pixel dimensions then the larger one can gather more light if fully illuminated. Which will reduce pixel noise

 

[rdenney]

Correct.

I use my phone for documentary photography. But it lacks editing range and control.

I like the idea of rangefinders and have tried a number over the years. They are all frustrating in practice, though, and I find myself relaxing back into ground-glass viewing.

I guess I’ve never been particularly sensitive to bulk and weight. Compared to my Sinar, my Pentax 645z—which most of you would think a beast—is remarkably portable and convenient. And unlike software-driven cameras, actually has physical controls and buttons I can use by tactile feel while my eye is focused on that ground glass.

Nobody who has used digital cameras from iPhones to medium format doesn’t understand the advantages of larger sensors, by the way. My iPhone is pretty good at masking noise, but the Pentax doesn’t create it in the first place, in the same conditions. The bigger advantage, however, is dynamic range.

Rick “a ground-glass guy, with or without a pentaprism” Denney

 

[Offler]

Not sure what your trying to say. Bokeh is a hazy concept that describes the way a lens renders out of focus elements. It varies from lens to lens of the same focal length but of differing design. And of course it's aperture dependent. Do you mean Depth of Field?

As for light gathering. It is dependent on a bunch of things but it's mention regarding sensor size relates to pixel size. A sensor with larger pixels can accumulate more photons in the same time span. So if 2 sensors have the same pixel dimensions then the larger one can gather more light if fully illuminated. Which will reduce pixel noise

Basically what you said. Most photographers I know tend to forget about pixel surface and the go by "sensor generation". Two sensors with different sizes but same pixel size (and construction) will have exactly same ISO range but grain/noise is little bit more complicated.

Higher pixel count reduces noise because there is more pixels on the image, however it increases it at the same time because the pixel surface is smaller. Large sensor can have small pixel surface, and in general capture less light per pixel, regardless the whole sensor has larger surface compared to M43.

If the sensor is capable to capture images in 8, 10, 12 or 14bit depths also affect noise - more bits = higher precision.

Thats why "smaller sensor = less light" cannot be taken seriously.

 

[rdenney]

Basically what you said. Most photographers I know tend to forget about pixel surface and the go by "sensor generation". Two sensors with different sizes but same pixel size (and construction) will have exactly same ISO range but grain/noise is little bit more complicated.

Higher pixel count reduces noise because there is more pixels on the image, however it increases it at the same time because the pixel surface is smaller. Large sensor can have small pixel surface, and in general capture less light per pixel, regardless the whole sensor has larger surface compared to M43.

If the sensor is capable to capture images in 8, 10, 12 or 14bit depths also affect noise - more bits = higher precision.

Thats why "smaller sensor = less light" cannot be taken seriously.

Don’t confuse precision with accuracy.

And sensor generation doesn’t take the same big steps it used to. And even that is more about software correction that sensor capture, particularly with iPhones.

My wife’s D500 sensor is newer tech than the Sony sensor in my Pentax 645z, and has 20-odd megapixels. But the older 645z absolutely blows it away in low-light capability. That’s clearly large sensor size at work.

Rick “who works in raw” Denney

 

[Destination: Moon]

Don’t confuse precision with accuracy.

And sensor generation doesn’t take the same big steps it used to. And even that is more about software correction that sensor capture, particularly with iPhones.

My wife’s D500 sensor is newer tech than the Sony sensor in my Pentax 645z, and has 20-odd megapixels. But the older 645z absolutely blows it away in low-light capability. That’s clearly large sensor size at work.

Rick “who works in raw” Denney

Larger pixels is the primary factor. The signal needs less amplification - so there is less heat and noise generated - much like audio in some regards I'm guessing. More pixel density gains resolving power but at the expense of noise and lower sensitivity. In daylight these things don't collide for images that are relatively small files. It's astounding how small the sensor and lens systems are in camera phones - and the superb images they can generate with the power of advanced processing.

If you print a file to a 20x30 inch print, and view it from 6 feet or less the defects and nuances become clearly visable. That's when the bigger sensors really show their advantages in image quality. Viewed digitally, a well exposed image file from a good phone or a FF sensor is usually barely discernable unless your pixel peeping or the original image is severely cropped

 

[rdenney]

Larger pixels is the primary factor. The signal needs less amplification - so there is less heat and noise generated - much like audio in some regards I'm guessing. More pixel density gains resolving power but at the expense of noise and lower sensitivity. In daylight these things don't collide for images that are relatively small files. It's astounding how small the sensor and lens systems are in camera phones - and the superb images they can generate with the power of advanced processing.

If you print a file to a 20x30 inch print, and view it from 6 feet or less the defects and nuances become clearly visable. That's when the bigger sensors really show their advantages in image quality. Viewed digitally, a well exposed image file from a good phone or a FF sensor is usually barely discernable unless your pixel peeping or the original image is severely cropped

Mostly, yes, but it all depends on use case.

Here’s what I want: a print that even when approached and viewed closely does not undermine the illusion of unlimited detail—that the only reason more detail can’t be seen is because the viewer can’t get closer.

Any camera/lens/capture system can create that impression up to a certain print size. I can make snapshot-sized prints from an iPhone, or prints three feet by four feet from my 645z while maintaining that illusion similarly.

That’s my objective for resolution and detail, and I’ve been chasing that objective for nearly 50 years, using cameras up to 4x5”.

My objective for tonality is to avoid any impression of tones being discontinuous. That is actually more difficult than the resolution and detail. Software generally wants to smooth over noise and increase the appearance of edges. This creates the impression of sharpness with smaller enlargement ratios. The more software processing (in camera or in post) that is required, the smaller the resulting print can be and sustain the illusion of continuous tone when viewed closely. That’s no small accomplishment.

My iPhone photos look great on iPhones, and only okay on computer monitors—most of those are documentary and it’s good enough. A 16x20” print? No way. I don’t go larger than 12x18” prints with my Canon 5DII, because colors wash out and the limits of detail show themselves. My big Pentax will much more than double that, but the Canon is certainly not recent tech.

Bigger sensors support both of my objectives in part for the reasons you mention, but only in part. Newer small sensors attempt to do so mostly by use of greater software power, and to some extent by improvements in the sensels themselves. But the worry about aliasing, even with very high sensel density, and use a filter to cut off detail near the Nyquist limit, just as we do in audio.

But the 33x44mm sensor in my 645z, which came out in 2014 but used 2013 sensor tech, does it better than my wife’s Nikon D500 DX-format camera, which came out three years later. That’s true even in low light and high ISO. The sensel density is high but the cells are larger and either avoid the aliasing problem or leave it to post-processing, which means no AA filter.

Lenses are part of the problem—a big part. I can attain my objectives with most lenses Pentax made for medium format, even film-era lenses now ridiculously cheap in the secondary market, when used carefully. The lenses needed to get the most out my wife’s D500? Bring your checkbook.

So, technological advancements do not replace the importance of size, at least not a generation or three, because sensor tech is only part of the problem in the real world.

(This isn’t an issue related to mirrorless or not, of course.)

Rick “who actually makes bigger prints, and isn’t measuring outcomes based on pixel-peeping” Denney

 

[Destination: Moon]

These are a couple reasons I went with Fuji after years with Canon film and then digital. I always found huge sample variations in Canon L lenses. Which made buying them a big pia.....

After years of sorting canon lenses, I bought 2 identical Fuji zooms and they were both excellent and virtually identical image quality. I'd never seen anything remotely like that with Canon. Fuji quality has suffered a bit now that they moved production to the Philippines but still pretty good.
For the money, I don't see anyone else making glass as good as Fuji in apsc format. I like smaller lighter gear as I'm mostly shooting nature while hiking or skiing

The other thing Fuji has is the lack of aa filters in a small very compact size across the line. I was hooked. I'm still not totally sold on xtrans. It can show artifacts - I doubt you'd be happy with it given the sizes and demands your placing on the images. But it's more than up to the task for my needs.

I'm not a pro, just a serious amateur that still can't believe the prints I can make at home on my old Epson 1280..... After sitting unused for months, it'll perform without any problems when asked, without a bunch of cleaning. Coming from my old BW darkroom days it's a dream come true

 

[mrbungle]

These are a couple reasons I went with Fuji after years with Canon film and then digital. I always found huge sample variations in Canon L lenses. Which made buying them a big pia.....

After years of sorting canon lenses, I bought 2 identical Fuji zooms and they were both excellent and virtually identical. They've suffered a bit now that they moved production to the Philippines but still pretty good.
For the money, I don't see anyone else making glass as good as Fuji in apsc format

The other thing Fuji has is the lack of aa filters in a small very compact size across the line. I was hooked. I'm still not totally sold on xtrans. It can show artifacts - I doubt you'd be happy with it given the sizes and demands your placing on the images. I'm not a pro, just a serious amateur that still can't believe the prints I can make at home on my old Epson 1280..... After sitting unused for months, and then perform without any problems when asked. Coming from my old BW darkroom days it's a dream come true

Same here, amateur super happy with the Fuji S10. I recently splurged on the 18mm/f1.4. It’s just the perfect grab and go camera now. For me at least.

 

[Offler]

Larger pixels is the primary factor. The signal needs less amplification - so there is less heat and noise generated - much like audio in some regards I'm guessing.

That sums it up nicely.

Mostly, it gets overrated, but only up to full frame. When you introduce medium formats to such person, he could not justify his previous claim, and it tells that the dispute is not about technology, but about favorite format.

This is why I check "native" and "extended" ISO ranges. The first one is done solely by the sensor, the latter requires processing and is more likely to introduce noise. And having 100-6400 native range is enough for me.

 

[PierreV]

This is why I check "native" and "extended" ISO ranges. The first one is done solely by the sensor, the latter requires processing and is more likely to introduce noise. And having 100-6400 native range is enough for me.

Everything is done on board in the sensor. ISO is a bit meaningless - it is a convenience for the photographer.
Dynamic Range per sensel depends only only well capacity and read noise (for short exposures) and shot noise if you do too short low light exposures.
There might be a couple of "native" ISOs to choose from.
Here's an example for the Sony IMX455

 

[michael-b]

While if you prefer to take photos using standard control with manual focus, exposition and aperture control, get a small camera with replaceable lenses.

There is a lot of myths surrounding APS-C and M43 systems regarding "less light" where people who dont know how optics work claim that aperture has to be adjusted by crop factor (not it does not) or that smaller sensor size means "less light", while light intensity is exactly the same.

The only difference is that 25mm M43 vs 50mm FF will have different perspective and bokeh at F2,8 for 25mm M43 will be about same as F5,6 at 50mm FF.

Not really. I started with Olympus a few years ago, then got a Canon R6 full frame. The Canon handles low light much better. You're right about the bokeh difference, but the M43 sensors -- all of which are dated -- are not up to the performance of today's full frame sensors. Could they be theoretically? I guess. Are they? No. Will they catch up? Economics are against Panasonic and OM System having the resources to do so.

 

[Destination: Moon]

Same here, amateur super happy with the Fuji S10. I recently plunged on the 18mm/f1.4. It’s just the perfect grab and go camera now. For me at least.

If you can find a Japan copy of the 18-55 kit lens, jump on it. It's the best travel lens I've seen. It lacks some zoom range, and speed compared to Fuji's bigger Med zooms but it's 2.8f at 18mm and sharp out to the edges across most of the zoom range

Everything is done on board in the sensor. ISO is a bit meaningless - it is a convenience for the photographer.
Dynamic Range per sensel depends only only well capacity and read noise (for short exposures) and shot noise if you do too short low light exposures.
There might be a couple of "native" ISOs to choose from.
Here's an example for the Sony IMX455
View attachment 174683

There a word or a term to describe what your saying. Inversion or something like that? I read the concept over at luminous landscape. It's an old piece by Michael Reichman. It basically made the point that driving sensors at very high iso to capture in low light wasn't the best practice as better results were possible using post. It showed how there was good - usable detail in very underexposed images shot in near dark at low iso. It was an interesting article. I'll try to find it again later

 

[JeffS7444]

Not really. I started with Olympus a few years ago, then got a Canon R6 full frame. The Canon handles low light much better. You're right about the bokeh difference, but the M43 sensors -- all of which are dated -- are not up to the performance of today's full frame sensors. Could they be theoretically? I guess. Are they? No. Will they catch up? Economics are against Panasonic and OM System having the resources to do so.

But Sony does, and this is the new (as of July 2021) ultra-fast M43 sensor, specifically intended for consumer still cameras:
https://www.sony-semicon.co.jp/products/common/pdf/IMX472-AAJK_Flyer.pdf
How fast? 120 frames per second at 12 bits/pixel, full resolution.

 

[michael-b]

But Sony does, and this is the new (as of July 2021) ultra-fast M43 sensor, specifically intended for consumer still cameras:
https://www.sony-semicon.co.jp/products/common/pdf/IMX472-AAJK_Flyer.pdf
How fast? 120 frames per second at 12 bits/pixel, full resolution.

Would be great if it came to be, but my understanding is that OM System is moving toward video, commercial applications. The industry is training people to embrace full frame, for better or worse, and it's hard to imagine how the shift away from M43 would be reversed.

 

[PierreV]

There a word or a term to describe what your saying. Inversion or something like that? I read the concept over at luminous landscape. It's an old piece by Michael Reichman. It basically made the point that driving sensors at very high iso to capture in low light wasn't the best practice as better results were possible using post. It showed how there was good - usable detail in very underexposed images shot in near dark at low iso. It was an interesting article. I'll try to find it again later

Yeah, I remember. Most of it would be outdated today though given the evolution of sensors. ETTR (expose to the right) is still valid, of course: fill the wells as much as you can and you maximize DR. The rest was too sensor dependent to stay relevant over the years.

Basically, what they were after, indirectly, on Luminous Landscape, was the minimum read noise of the sensor. You wanted the best signal/read noise ratio. So, on a camera that had a good ratio at 100 ISO but a bad one at 1600 ISO, an underexposed image was indeed much better than a "properly exposed at first sight" one. It made sense then, given that sensors data-sheets and implementation details weren't necessarily known.

The best site for technical information about sensor implementations in cameras is currently this one

Photons to Photos

www.photonstophotos.net

The photographic terminology actually makes things much more complicated than they actually are. In the widely used IMX455 sensor I mentioned above, if you have a bright subject you would maximize DR at 0 gain because your well capacity would be higher but if you don't have enough light, using a gain of 100 would make more sense because what you lose in well capacity (or more correctly how your well filling level is measured) you gain back in read noise.

In practice, how the camera firmware handles those things under the hood is what matters to the customer though, and it can be a bit random.

 

[JeffS7444]

Would be great if it came to be, but my understanding is that OM System is moving toward video, commercial applications. The industry is training people to embrace full frame, for better or worse, and it's hard to imagine how the shift away from M43 would be reversed.

New enthusiast-grade camera has already been teased as of late October 2021, starting at about 8:20 of this video:
https://om-digitalsolutions.com/octannouncement/en/
It's expected to be an EM-1-series camera, and I speculate that they'll attempt to release it in time for Japan's Golden Week holidays which begin 29 April, 2022.

 

[JJB70]

I think use case is critical to just about any technical evaluation. The example I like to use is cars, I used to love winding car enthusiasts up who's get into heated debates about whether the latest Porsche 911, Ferrari, McLaren or whatever was the best car in the world by saying my Toyota MPV was better than any of them. They'd invariably take the bait and erupt yet my opinion was absolutely correct for my own use case - what good is a two seat supercar if you have two kids and sometimes need to take grandparents or friends along too, and need a big boot? A Ferrari, however good it is on a track, would be useless to me (and no faster on public roads anyway). Cameras are the same, as a discretionary purchase for many people it is easy to overlook the importance of what feels right to the user, tactile feel, control layout, industrial design and yes, brand image, make a difference. Beyonf that, do you need interchangeable lenses? Do you need something lightweight, weather sealed? Are you a studio professional? What will images be used for, do you really need a 60MP FF sensor? Are the necessary lenses and other accessories you need available?
Digital cameras are a mature technology and sensors reached a point whereby almost any of them will meet the needs of most people (including most professionals). If people doubt this, ask yourself how many of your smartphone shots would be rejected for technical reasons? If anything smartphone cameras are ahead of dedicated cameras in some areas simply because that's where most R&D spend goes and they can do amazing things with software systems. Ultimately I suspect in most cases camera choices are made based on personal preference, and there's nothing wrong with that.

 

[JJB70]

Something which has become very obvious very quickly is the value of an exposure compensation dial or assigning that function to a dial or control. One of the biggest advantages of mirrorless is being able to use exposure compensation in real time so to speak, it is quite transformative and has joined shutter speed and aperture as the third big control function for me.

 

[keith_h]

I'm firmly in the traditional SLR bodied camp for the long term. The current crop of mirrorless bodies offer amazing results but the cost for me to migrate is prohibitive.

I currently have four bodies and a suite of lenses, plus an RX100II which means pretty much every situation is covered. Up until 2013, I was shooting events with a pair of Sony A700 which remain in use even now. Great cameras capable of superb images then and now. Newer cams have improved focus and sensors have come a long way, but these cameras handle well, I can use them entirely by feel, I can get exactly the result I need from them. Why would you change?

But size I hear you say. Once you start getting some serious glass on the front of an A7 for example, the size argument goes out the window. I have an A200 body which makes a good travel kit in combination with the appropriate lenses. Compact, light, convenient. Not as full featured as an A700 for sure, but quite capable of great images. I took this camera to Italy for a month and didn't regret the decision at all.

The RX100 exists for the same reason everyone else has one, superb images without drama in a small rugged form factor. Saved the day on an overseas trip when my main lens bit the dust. There are things that you can do with a DSLR you cannot with the RX100, but in many situations its going to be good enough.

The icing on the cake for me is now that Sony have canned the A mount, lenses are very affordable. And while I appreciate the E mount optics are cutting edge, I like the image quality of the legacy glass as well as the way the old lenses work. It's all good for me on that front.

And the camera in a mobile phone while great for snapshots just isn't going to get the job done in every situation. The number of shots I've missed with an iPhone is what led to acquiring the RX100. Problem solved.

Gear