I have an old 900mm lens from the 70’s. There is often not enough light to achieve decent sharpness when you pixel peep, even under ideal conditions. It’s always a trade off between iso noise, low aperture blurriness, and shake at slow shutter speeds. At f5.6 the edges are blurry even the center too. f11 is much sharper and no chromatic aberration. Since there is no image stabilization 1/1000 is ideal to maximize sharpness. On a sunny day iso 3200 adds noticeable noise. None of this is ideal. It is less sharp than cropping a modern 250mm. It is a lot of work to get a moving subject in focus closer than 50ft. It’s hard to hold still enough to reduce blur, without a tripod.
I think I’ve found a solution! Can you think of a better idea? Am I wrong that it APPEARS sharper?
Alright maybe not sharper, but it appears AT LEAST AS SHARP. Depending on your skill level. With much less effort. Much skill is required for an old lens and a dynamic subject. It even works with a tripod at near dusk.
Change the resolution setting to M or S not L. This is my “pixel binning”. I’m happy to hear why it’s not really. This allows me to choose High NR at high iso, without reducing the sharpness of M resolution. Of course it effects L res. Multi NR works even better with a static subject with or without a tripod. Now I can use iso 6400 and it looks the same sharpness as 1600. F22 is my choice now focusing is easy even at dusk. Iso 12800 has no noise because of the NR. It’s as sharp as f5.6 was before. Camera shake at 1/400 is not so critical as before. How is this possible? Simply by changing pixel resolution?
I’ve created a point and shoot camera. With a very long lens. It is so much more fun to shoot when birds are far away. When they are very close no need to focus anymore. I enjoy this more than a modern 250mm with IS. The results seem comparable in sharpness. You can add a 2x extender which is otherwise impossible to get decent results.
Why are so few people doing this? There are so many disadvantages to overcome compared to a modern long lens with AF and IS. Lower your expectations. Lower your resolution.
Here’s my famous quote: The best way to avoid disappointment, is to lower your pixel resolution. It still looks sharp on a laptop screen!
Am I missing something? What are the disadvantages to this method? It appears as sharp as L resolution using an old lens. At a much lower cost $20 on eBay. You can’t print billboards.
There is a lot to unpack here. Some of it is a bit obscured by terminology. Since I don’t know what camera you have, I have no idea what S,M,and L are when it comes to resolution. It could go different ways. Small pixels versus Large pixels? And how big?
The unknown question is what amount of processing is occurring in camera, and exactly what parameters are being used. If your camera has a Raw mode to save pictures in, it would be useful, as you could get the image before any processing had occurred, and learn a great deal more about what is going on by reproducing processing steps yourself on a computer.
Cameras have all manner of oddities going on. You start with the Bayer pattern of coloured pixels, and the processing needed to infer pixel colours. Dropping to a lower resolution alone can skirt past some of the artefacts there.
Left to itself, your camera probably applies some image sharpening to the result. To understand how this affects the result, again, try working with a raw image.
Averaging to improve pixel depth in the face of noise is a very real thing, and is why pixel binning has value. In general it can’t improve spatial resolution by itself, except by improving contrast lost to noise. It can’t, by itself, improve losses due to movement or limitations of the lens’s modulation transfer function (aka how good the lens is.)
I’ll bet you are seeing a mixture of the above, and that after the fact image sharpening is yielding the illusion of a sharper result. Which if it looks good, arguably, is good.
The best way to improve a blurry image by downsizing is to do it yourself in post and use the “nearest-neighbour” option instead of bicubic/bilinear etc. this works by undersampling, which would ordinarily result in aliasing but with a blurry image you’re effectively discarding parts of the point spread, reducing the effective blur diameter.
I’m taking this to the next level by doing this in camera in advance. This makes the post process trivial as there is nothing more to do!
It just sounds like a crappy lens, if you can get the same result with less fuss with 250mm and cropping - because that’s essentially what you are doing. Instead of cropping a sharp 250mm picture from a modern lens (with image stabilization), you are taking a picture with half or less resolution full frame (we assume) from a 1000mm lens that is not as sharp. Cropping to select 1 pixel on the sensor vs. binning multiple pixels from a less sharp image… (At least with a 250mm you can pick the area to crop - so exact framing of the image is less of an issue.)
The key is what the binning gives you - So if your sensor is say, 250mm at 24Mp and you crop to 6Mp, or you take full frame 1000mm but combine 4 pixels-to-1 to get a 6Mp image - do you get better light sensitivity at least? Or do you have to reduce the ISO to get less noise anyway? Can you use a shorter exposure than the 250mm?
But then, with no Image Stabilization function, you have to be sure your heavier configuration is more stable.
Not sure what a 900mm looks like, but one with f/5.6 makes me think it’s a monstrosity. On safari I took a Tamron 18-400 (with a Canon M3, so not full 35mm sized frame) with a max at full zoom of f/6.3 and that was as big as I wanted to travel with. 900/5.6 gives me 6.4 inches diameter for the front glass alone if I’m remembering this right. Not something fun to travel with. Maybe for local scenery shots… Or if you budget is very heavily constrained.
(But about 15 years ago I was in NYC passing by where they were doing a film shoot, and some of the paparazzi across the street had those white Canon lenses bigger than that.)
Like everything in photography, it’s a trade-off. Change one thing, compensate elsewhere.
Anyway, I think you figured it out – you would probably get better results reducing the resolution in post processing manually, but this is a quick workaround that avoids that kind of post processing.
You still might be better off using a sharper, newer lens and cropping, depending on how much you have to crop and what S, M, and L correspond to.
There’s something about “circle of confusion”, and I think if you end up with effective pixels bigger than that, you’ll get rid of motion blur, but I’m at the limit of my knowledge there.
I have a few old lenses and I’m shocked at how much sharper newer lenses are. My old lenses were consumer grade (as are my new lenses) – maybe pro lenses of old were sharper.
Modern camera resolution is just a lot higher than most films, and the lenses have had to adapt, I guess.
Once I took some snapshots of humans on 35mm T-MAX 3200 and the film grain made it look awesome.
That may have been a particular aesthetic choice, but I am suggesting that “pixel binning” or “noise reduction” or heavy AI photoshopping does not always make a photo look better. However, I agree with @Francis_Vaughan that you are better off disabling all that on the camera and doing all the processing on the computer; it is much more powerful and then you can really experiment with getting the right effect.
Because it’s fun. A friend of mine had an old 9mm lens that he wanted to use for some video thing he was doing on his modern mirrorless. The number of adapters he had to set up was hilarious!
As a person whose photographic “skill” consists of choosing “Auto”, aiming carefully, and pressing the button, I have to say I’m in awe of the photographic knowledge in this thread.
For truly soft images, I use Topaz Sharpen AI. It can sometimes perform miracles. Their Gigapixel AI for upsizing works similarly great. There are limitations and it sometimes makes makes mistakes, so best to do it as a layer in Photoshop and tweak, but it’s my solution for where various standard sharpening techniques are insufficient.
Exactly! the point with this would be to be able to use something that should be obsolete and superseded, just for fun. Like making something with hand tools instead of power tools.
My father-in-law told me he had a friend who, growing up in the late 50’s, absolutely loved the look of the 57 Chevy and really wished he could own one. Then around 2000 the guy finally got a chance to drive one. His take-away from the experience? “What a piece of junk!!” The good old days were more old than anything.
What I suspect you’re doing is stopping your lens down to the point where the aberrations are minimal, and then relying on the post-processing in the camera to compensate for any issues that might cause. They do this in astronomy/astrophotography all the time to reduce mirror and lens aberrations.
You’re probably best off like others have said, doing that post processing in a computer app, and not relying on the built-in camera functions. They’re almost certainly optimized for speed to some degree, and are fairly blunt instruments (L, M, H), while the computer apps have a lot more fine tuning and aren’t as concerned with having it applied and saved fast enough so that the photographer can take another shot right away.
My earliest raw files still read fine (late 2005). I have a D1 (1999) as well, but that camera didn’t even have an option for raw files, just JPG or TIF. I’d be surprised if Lightroom/Photoshop/Adobe Camera Raw didn’t continue to support those legacy formats. I think they just keep adding more and more support and don’t deprecate the old ones. It would suck if they did, as many photographers (like me) keep an archive of all their photos that span back to the beginning of digital cameras (and before, in my case, in binders full of negs.)
Why must everything be “vinyl record”-ed? Not all old stuff is good.
I have this Nikon f4 500 mm lens. I can’t imagine what a 70’s 900 mm looks like. “Ginormous” comes to mind. Unless it is a mirror lens?
While it has nice optics, with no aberration, it weighs a ton, and is only manual focus. And combine that with the fact modern DSLRs don’t have prism focusing screens makes it difficult to get a proper focus, especially in anything other than bright sun.
“Lower your expectations, lower your resolution”? No thanks! I like high quality pictures. Though I do hear 480 line resolution CRTs are making a comeback.
I was just wondering… I have an Adobe cc:mail archive from my first job out of college (late 1990s), and try as I might, there seems to be NO way to actually convert that thing into a PST or anything else that’s of this century. Nor are there any copies of cc:mail floating around to install, so I can export it.
