[FozzFoster]

I feel intimidated when I hear about 50 and 60 MP files (100MP! Yikes!).. post-processing and storage of those files is definitely a different beast then processing my 20 MP files while can already take awhile to load off my 5600 RPM HDD..

[biz-engineer]

I feel intimidated when I hear about 50 and 60 MP files (100MP! Yikes!).. post-processing and storage of those files is definitely a different beast then processing my 20 MP files while can already take awhile to load off my 5600 RPM HDD..

Unfortunately, the number of mega pixels is only part of the story for photo printing. The pixels will only ever reflect what the lens is capable of producing optically. With more mega pixels on the sensor, you just get more resolution of image details , but you also get more resolution of those red/blue fringes in high contrast area of the image and more resolution of those smeared corners of that wide angle lens. Personally, I don't believe in the 61Mp on FF hype of Sony, it's just that the camera user is almost never going to get true 61Mp out of the camera, due to lens limitations, diffraction, camera motion, subject motion, higher ISO setting etc. Getting to larger prints require increase of both mega pixels and sensor size, becoming awfully expensive on digital.

[MJKoski]

Non-shaken, @ISO100, pixel-shift allows 1.5m wide print which can be inspected at 50cm distance without issues if lens is not a total potato. Done such print now 3 times. My usual end product is 1.25m wide.

[biz-engineer]

without issues if lens is not a total potato

sometimes the lens is a lemon....

[normhead]

So true. An article a while back was talking about how 60mp files weren't big enough for really, really big prints. 100mp or more was needed. As I remember the details, the author printed huge prints (like 4x6 feet) of 100mp and down sampled 60mp, and 36mp files. Had people view them at average distance and vote for the best version. He claimed that even at distances where the eye shouldn't be able to tell them apart, the 100mp pic garnered the most votes. Fascinating.

Unfortunately I had the article link saved on a tablet that recently died.
Thanks,
barondla

And that is relevant to people who print 72 inches by 48 inches. If you're one of those people, listen to what he says, if not ignore it. a 16 Mp camera is 5000 x 3400 approx. A 36 Mp camera is apron 7000 x 5000. You double the MP count, but you don't double the pixels dimension.In fact instead to 2x they are only .4x

A 100 MP file is 11,000 by 9000 approx. 3 times the MP but not even double the dimensional pixel cost. So yes, the amount of MP needed for large files increases exponentially, but, it's impossible to say what that translates to.

But a few quick calclations. 72 inches at 100 MP (11,600 pixels wide) gives you a pixel density of 161 per inch.
7000 (K-1) divided by 161 equals 43.4 inches.

SO to match the pixel density recommended by this guy for maximum display you could print with a K-1 to 43.5 inches.
My own research suggests that for my own preferred medium, printing on canvas, I could go considerably larger, although I do try and print at least 150 pixels per inch with 200 being a preferred minimum and 300 or 360 being ideal.

But, my point is, if you have an image you really like becuse of its composition, colour values etc., you may be quite pleased with the image printed to as low as 72 pixels per inch which would be 100 inches on a K-1 image. It might not look as good as the same image printed from a 100 MP camera, but that fact will possibly be irrelevant to you as you sit there on a comfy sofa drinking a glass of wine admiring your image.

My 42 inch image hanging on my wall is 142 pixels per inch, less than my desired standard, yet the image looks great, (and for those of us who remember, digital prints started taking over from analogue when printers started to reach 144 DPI.) Everyone who comes in my place goes over and looks at it, some have offered to buy it, but not at the price I'm willing to sell it for. And it's not canvas so the usual argument about canvas being more forgiving doesn't apply.

This is not about what looks the absolute best, this is about what might make you happy. In my experience 110 pixels per inch can be sellable. It can make people happy enough to pay good money fo it, and we never sold those images for under $300 a print.

[biz-engineer]

And that is relevant to people who print 72 inches by 48 inches. If you're one of those people, listen to what he says, if not ignore it. a 16 Mp camera is 5000 x 3400 approx. A 36 Mp camera is apron 7000 x 5000. You double the MP count, but you don't double the pixels dimension.In fact instead to 2x they are only .4x

That's why I have a K1 and plan to have it for very long.
Given that to double the enlargement, I would need 4x the number of Mpixels, means 4 x 36 = 144 Mp, which is even more Mpixels than the $10 000 Fuji GFX100 can deliver. If ever I need to print a few meters long on a wall, I'd just stitch a panorama from K1 files and it's cost nothing but a little bit of processing time on computer.

If I'd look at a 5DSr (50Mp), a D850 (45Mp), a Z7 (45Mp) or A7rIII (42Mp) or even A7rIV (61Mp), or S1R (47Mp), the improvement vs Pentax K1 is so negligeable that the few additional Mp of those camera models aren't worth at all the money they cost.

[normhead]

That's why I have a K1 and plan to have it for very long.
Given that to double the enlargement, I would need 4x the number of Mpixels, means 4 x 36 = 144 Mp, which is even more Mpixels than the $10 000 Fuji GFX100 can deliver. If ever I need to print a few meters long on a wall, I'd just stitch a panorama from K1 files and it's cost nothing but a little bit of processing time on computer.

If I'd look at a 5DSr (50Mp), a D850 (45Mp), a Z7 (45Mp) or A7rIII (42Mp) or even A7rIV (61Mp), or S1R (47Mp), the improvement vs Pentax K1 is so negligeable that the few additional Mp of those camera models aren't worth at all the money they cost.

Exactly my evaluation as well.

I wouldn't go over a K-1 unless I had access to a market that would pay me $3000-$5000 print (and I do know of a few guys in that market.). The craft show market tends to be one tenth of that, at it's very best and lots of folks are looking at 8x10s or 5x7s in the $25-$75 range.

---

Last edited by normhead; 11-25-2019 at 01:14 PM.

[biz-engineer]

Exactly my evaluation as well.

That also led me to think that I might be better off using an apsc camera for macros (more magnification) and wildlife (benefit = faster frame rate, better AF coverage), and stitch a bunch of vertical shots when I want a high resolution landscape. The downside with the K1 is that you get a lot less frame rate traded for a little more resolution. That's why I'm likely to get the new Pentax apsc model and use my DFA28-105 and DFA150450 with it.

[RobA_Oz]

My largest prints to date have been 600mm by 400mm, on a variety of media, using both the K-1 (with FA31) and K-3 (with DA18-135). If I had more wall space, I’d happily go larger with either, although the K-1/31 would require less work to satisfy me.

Viewing distance is as important as anything, obviously, but when people like an artwork, they almost always lean forward to look more closely, so dpi matters to some extent. Unless you’re producing something forensic, though, the original image resolution isn’t that important: you can always upsample using several editing tools to interpolate.

[barondla]

All interesting points so far. Wouldn't a well done Pixel Shift image allow further enlargement than a standard shot? If I'm not mistaken, the image height and width stays the same, but the file is cleaner and can be up sampled more?
Thanks,
barondla

[normhead]

All interesting points so far. Wouldn't a well done Pixel Shift image allow further enlargement than a standard shot? If I'm not mistaken, the image height and width stays the same, but the file is cleaner and can be up sampled more?
Thanks,
barondla

I would go with a pixel shift image, just on the chance it might be better. But i haven't worked out the parameters. Sometimes a pixel shift image is noticeably better, sometimes they aren't much different. There might be a bit more resolution but the biggest advantage is colour gradation and clarity. It's a better picture, but probably not a bigger picture. But someone who has done the testing should answer.

[RobA_Oz]

All interesting points so far. Wouldn't a well done Pixel Shift image allow further enlargement than a standard shot? If I'm not mistaken, the image height and width stays the same, but the file is cleaner and can be up sampled more?
Thanks,
barondla

I would go with a pixel shift image, just on the chance it might be better. But i haven't worked out the parameters. Sometimes a pixel shift image is noticeably better, sometimes they aren't much different. There might be a bit more resolution but the biggest advantage is colour gradation and clarity. It's a better picture, but probably not a bigger picture. But someone who has done the testing should answer.

Not having shot much with pixel shift, I couldn’t be certain how much extra resolution is achieved, either. Ricoh’s Pixel Shift web page isn’t helpful in clarifying that, unfortunately.

[Take-5-JB]

CAVEAT: I am not an expert, a physicist, or optical engineer. I am a retired industrial chemist.

I have a small photography business, I don't make a lot of sales, I am a juried exhibiting artist in photography and exhibit in a gallery in Charleston, SC. I started photography in 1977 with 35mm film. All that considered, see caveat.

For a commission in 2018 for a landscape of a lake with mountains and clouds during an afternoon storm, I judiciously used denoising, sharpening, and my favorite interpolating enlargement algorithm and made a 300 dpi 72x96 inch image from the 16x24 inch 300 dpi image that is generated from 35mm full frame digital. I that took down to 48x72 inch at 450 dpi. I sent this to my canvas printing partner and they produced the 48x72 inch canvas stretched in gallery mount. My client hung it above the couch in her den that had a 20 foot ceiling. She loved it and had a party shortly after where it was a hit.

Could you inspect it with a loupe? Not to make experts happy, but it had the artistic value she desired with the event she requested. She gladly paid for the work.

The image was made with my K-1 and A35-105mm with a circular polarizer, a tripod, and hand held shutter release. Again, see caveat.

JB

[StiffLegged]

Could you inspect it with a loupe? Not to make experts happy, but it had the artistic value she desired with the event she requested. She gladly paid for the work.

Exactly. Keyboard warriors aren't going to pay for that, but good clients do!

[RobA_Oz]

CAVEAT: I am not an expert, a physicist, or optical engineer. I am a retired industrial chemist.

I have a small photography business, I don't make a lot of sales, I am a juried exhibiting artist in photography and exhibit in a gallery in Charleston, SC. I started photography in 1977 with 35mm film. All that considered, see caveat.

For a commission in 2018 for a landscape of a lake with mountains and clouds during an afternoon storm, I judiciously used denoising, sharpening, and my favorite interpolating enlargement algorithm and made a 300 dpi 72x96 inch image from the 16x24 inch 300 dpi image that is generated from 35mm full frame digital. I that took down to 48x72 inch at 450 dpi. I sent this to my canvas printing partner and they produced the 48x72 inch canvas stretched in gallery mount. My client hung it above the couch in her den that had a 20 foot ceiling. She loved it and had a party shortly after where it was a hit.

Could you inspect it with a loupe? Not to make experts happy, but it had the artistic value she desired with the event she requested. She gladly paid for the work.

The image was made with my K-1 and A35-105mm with a circular polarizer, a tripod, and hand held shutter release. Again, see caveat.

JB

As much as I like printing on canvas, it does tend to mask a lot of very fine detail, so on that score alone some “experts” wouldn’t be happy. By the way, do you find you have to work with contrast and micro-contrast, to get the best out of the A35-105? I like my copy, but that’s the one aspect that I sometimes have an issue with. I’m told it’s because of the lack of MC on the back of the rear element.

Also, which is your favourite interpolating enlargement algorithm?