I read a study of data collected watching patrons in museums view paintings, and it was noted that a person tends to view any image from a distance equal to 2x the diagonal of that painting (or photo).
Thank you Keith for this excellent video and so generously sharing your extensive knowledge on printing and photography. After lots of experience printing in the darkroom (years ago) I am relatively new to digital printing, and very much appreciate your discussions on all various aspects. Cheers! GB
I stumbled on your channel thanks to Google's algorithms. I am barely an amateur photographer (I have had to learn how to take photos of my art by necessity and they are not close to what a photographer would do). I also own an Epson SC-P900. I wanted to thank you for providing such amazingly good and to the point information. Every word that comes out of your mouth has meaning, and at multiple levels. Listening to you reminds me of reading Robert Henri's "The Art Spirit." Your knowledge was undisputed but in this video I realized that I am not just listening to a true expert, but that you are also an amazing artist. Thank you
I also find media and intent surely come into play here. e.g. a canvas print of gouache styled art doesn't need to be super sharp and it still looks good.
I find the science of viewing distances oddly satisfying! I remember looking up the near point (the closest distance your eyes can resolve details) and learned it was ~8 in on average. When I grabbed a ruler and tested, I found it to be roughly true. So for most people's eyes, that chart doesn't need to go lower than 8 inches or 858 ppi.
Do you print often at 600 PPI and editing the images with ACR? I have been using 400 PPI and will experiment with 600 PPI on my EPSON P 900. Great Video awesome work!
It depends on the source image and the size - I print at whatever convenient resolution is there - so no specific number [ACR is part of my Photoshop edit workflow, but nothing to do with sizing) For a lot more info on the P700/900, read my article about this very topic www.northlight-images.co.uk/driver-settings-and-print-detail/
Of my couple of thousand negatives, maybe a couple of hundred are worthwhile. But most of them have sentimental value, and are important to me. Scanning medium format negatives at 600 ppi was a big blunder. Printing at 300 dpi would span 4 x 4 prints. I now am gradually scanning again, at 4800. The TIF file size is typically 500 MB. So I can print a 13 x 19 nicely at 500 dpi. The prints look nice (to me) from eight feet away. Looking at them from one foot away seems silly. Overkill? Sure. But I'm happy.
Most scanners only have one scanning resolution, it's maximum. When you choose a lower res it simply downsizes the file before outputting it. You therefore always scan at the highest or native resolution of any scanner, it's not something you need to calculate or even think about. You then resize the file if you like, or not.
@@johnsmith1474 Depends on the job at hand. I tend to differentiate between bulk 'cataloguing' type scanning and scanning individual images at optimal - storage and time constraints vary. There is no better way of effectively culling/curating slides/negs than having to scan them. I've just got an Epson bulk print scanner to test. The speed goes down as resolution increases. I'll do a quick into for it here...
Group f/64 would beg to differ, but even Ansel backed off from the premise that all must be in sharp focus. People who may actually buy prints don't view them through a loupe--they either like the image or they don't. It just needs to "work" (and part of that is the viewer not noticing technical details, but rather just enjoying the image)...
Some time ago I did some research on this topic. I was wanting to see what resolution I should use for large panoramas of 2m or longer. Normal visual acuity is 1 minute of arc. That's a 60th of one degree. This means at 1m we can expect to define pixels at about 90ppi or less. The chart you showed recommending 191ppi at 36" makes sense. Double the resolution to make sure you can't see pixels in an image expected to be viewed from 1m. But here's a thought. What about viewing of a screen? On my 27" 4k computer screens, viewed from a typical distance of 600mm when at my desk, I'm looking at 164ppi and no hint of pixels. Likewise when sitting just 1500mm from my 65" 4K TV, as I am tight now, I'm looking at just 68ppi and again no hint of pixels. I do start to get a hint of rastered text closer than 1.5m though. I think prints from A3 up only need 200ppi at most. I've just received my ET8550 and am just learning to drive it. I have a couple panoramas' I will try printing A4 sections of at as low as 100ppi to see for myself what I can get away with as I would like to print them at least 2m wide, Yes I will have to do multiple prints on this machine. It'll be a fun learning curve I think. I can't thank you enough for the amount of info I have gained from your numerous ET8550 vids as well as others like this one. You're a treasure!
Thanks - always glad to contribute. Not sure about screens, but I view my 4k 32" screen from 70-80cm - but that depends on which pair of glasses I'm wearing...
Good stuff. I recently discovered that combining various levels of sharpness creates depth/three dimensionality... even for portraits. It's hit and miss... more art than science... but well worth experimenting with.
re: push button sharpening in LR: I agree that it's important to sharpen artistically on purpose, but there is a whole medium/size/resolution part to print sharpening that was integrated off some old plugins. It's a part of image making that most creators will not know the particulars of and the technical knowledge is getting harder and harder to find online. When I'm asked, I usually give people a crash course on manually sharpening the original image to completion, then using the baked in presets at print. 😅of course the real answer is more complicated, but everyone has a different appetite for technical complexity
Yes - The subject can soon get quite complex. In many respects I only expect people to take away a few aspects of videos like this. I've been asked to write a book on the subject of printing, but the tilt/shift one was hard enough!
@@KeithCooper I was recommended Image Sharpening by Fraser/Shewe in a printing class and finally picked it up. I def think a full book like your tilt shift one would be a real challenge to write 😅 there's so much about the craft that benefits from hands on real life teaching
@@JohnPurcell Actually I'll admit that there are two specific things putting me off right from the start - one that I'd reluctantly have to include Lightroom, and the second that I've not used a Windows PC this century ;-) Both would require considerable effort/time for me to get my understanding up to a level where I could include them in any more general book. If I stuck with Photoshop, Macs and a few other bits of software, it's a possible, but the whole field is changing far faster than the world of tilt/shift lenses. The complexities around when to sharpen, when to resize and how this relates to original files are a chapter alone...
To my recollection, 300 DPI comes from the Apple LaserWriter. I think it's worth noting the print density of your specific printer, multiples of 90 might work better than multiples of 100. It's also possible that my speculation is correct, but the effect not noticeable because the liquid inks tend to blend a little. LrC can enhance resolution, I pressed the button once to see what happened.
Yes '300' goes back a long way... So far, that almost any time a potential client uses the number I flag it as a potential marker for someone who does not understand what 'resolution' means ;-) 'Best' settings are very much dependent on the actual printer and driver - hence my specific testing of this for many printers.
"Why would I sharpen clouds in the sky?" That's what I said to my first client :) But it needed showing him oversharpened clouds (and remembering him that some of audition could be younger and could see it much faster) to show the effect.
I met this wildlife photographer who does big outdoor exhibitions worldwide and he was very "megapixels don't matter", he used to blow things up really huge, was about 2006 I met him.
@@adamarmfield1069 I have some seriously good lenses, but sharpness, chromatic and other aberrations, vignetting were never a consideration in my choice. I chose zooms, whenever possible, with a fixed maximum aperture because I don't want my exposure to change when I zoom. I used Canon's fantastic plastic F1.8, that's fast enough and the lens is cheap enough you could carry a spare or two. I have TS-E lenses because i want the shift
Informative information Keith - What are your thoughts though on actual final output PPI for printing? I let the PPI go where it wants to go, I don't resample at all up or down. I have made prints at 190 ppi which are excellent in quality. I think a lot comes down to lens quality too, and like you say intended viewing distance. What do you do with small prints, as you get some massive PPI if you don't resample at all - is this using more ink in not interpolating?.....would like to know your thoughts on final output if you let the PPI go where it wants to go or change it. I use my own printer and leave it go where it wants' to, and my professional printer does this as well.
Yes, I've made good prints at low ppi in the past, but with modern software, I'd likely consider resampling up to 450 or more. However, that depends on the source image and print size, and as you say, lens characteristics are part of the mix. My large 47 foot long panoramic was printed at 108ppi [due to a file/driver size limit] but resampled in the driver to 300 [Canon iPF8300] - People walked right up to that one ad no-one noticed any lack of detail and many commented how much detail there was... There are no correct answers in this sort of stuff - something I know makes some uncomfortable ;-) As to ppi and ink, it makes no difference - the driver resamples anyway, as part of the dithering and conversion from RGB pixels to ink dots
@@KeithCooperthanks Keith - At the end of the day then if I let the PPI goes where it wants to go then it’s the same as setting to an arbitrary PPI? I haven’t don’t elaborate tests on this. But have been happy with the results. I go into photoshop and turn off resample and set the print dimension size which gives me the PPI. I then make a new canvas with the PPI I got WITHOUT resampling. I will have a look at your website posts too, I believe you said you’d done some tests, would also be interested in knowing your upsizing software you recommend too.
Yes - the key is always to test and decide if it's worth the effort - Most of the time with my newer [more MP] images I do just let the PPI be whatever it is. It's older, lower MP images where I'll take a look at scaling, but it so much depends on the source image [the original raw file] I tend to use gigapixel AI these days, albeit not the current version, for upscaling - see the various links in the notes to this video
Hi Keith. Why not just print at highest possible ppi despite size of image, as we are not ultimately sure what distance some of the images may be positioned at at home/show, etc? at highest ppi, even the larger image is then viewable at "any" distance..
Yes, in general, that's why I suggest printing at the full available real resolution above 300. The tests for different printers show that this introduces some visible benefits up to say 6-700. [there are differences - hence my set of articles looking at this for different printers] With modern resizing software I'll consider 'creating' more detail for this, but it does very much depend on the original image, size and the printer. However, I'm not assuming any particular software, which people might or might not have available for their use, so as to make this a bit more generalised for people
@@KeithCooper I shoot at 50MB (Sony A1) all the way down to 12MB (DJI Drone Mini 3 Pro) and aim to get max resolution in all originals, but am only now thinking about a great printer for all this. You have been inspirational and very informative , so huge appreciation for taking the time to educate. At the end of watching many videos and also your impressions, Im now looking at a Canon Pro300 for A3+ and down, and would send any larger print formats to a professional print house. Im based in Cape Town, South Africa. Thanks Keith!
Thanks, Keith. Interesting and helpful, as usual. I have a couple of questions. Perhaps not strictly on topic.- 1) I'm told that printing from an Adobe rgb file is better than printing from sRGB. Because both contain 256 shades of each color, and the Adobe RGB color space is bigger the differences between colors is greater. Are we more likely to see banding in the transitions on a large print when we make a print from Adobe RGB? 2) Are our cameras capable of recording 100% of the Adobe RGB color space when we select to shoot in Adobe RGB as jpg? Same question when we opt for RAW or sRGB. Thanks.
Thanks Not topics for concise or simple answers ;-) 1 No - banding is quite a complex issue, not primarily related to this. Better? - depends on the gamut and content of the image. 2 The camera setting only affects the jpegs - no effect on raw files other than flagging the info, and [maybe - never tested it] the embedded jpeg thumbnail Given the camera can capture a very wide range of colours, it depends on how the camera maker decides to handle/process raw sensor data, as well as how raw conversion is carried out. Note that in both cases I might also query the actual questions ;-) So, for 1 who says it, what evidence do they put forward and under what print conditions? ;-) For 2 - what camera, what sensor technology.
@@KeithCooper Thanks, Keith, for your very speedy reply. Question #1 was prompted by my curiosity. Adobe RGB is a larger color space than sRGB, yet each is described by the same number of shades. It seems reasonable to me that colors in sRGB are more tightly defined and smooth transitions would be easier. Also, it would seem to me that any deficiencies in the reproduction of gradients might be more obvious in a large print, especially when the viewing distance is intentionally close. Typically, I print no larger than 13X19 and only from sRGB files. So I do not have any first hand experience with this concern. Question #2: I have a couple of old cameras. Sony NEX 6 and Sony SLT A55. I searched the manufacturer's and other sites, but couldn't find information about how these camera convert color, or how RAW processors convert color out of the camera. Thanks again for your attention and considered insights.
1 dithering and the print driver... lots of factors, it's not a problem but when it is, it's likely not for this reason. 2 - you won't it's proprietary info ;-) For both [and lots more ;-) ] get a copy of 'real world color management' by Fraser et al. Also, this has a lot of links: en.wikipedia.org/wiki/Raw_image_format BUT, be prepared to delve into a lot of stuff...
Hope you don’t mind me jumping in… 1) Adobe RGB vs. sRGB for printing: Note that the 256 levels per channel is assuming 8 bits per channel. If captured and edited at a higher bit depth such as 16 bits per channel, the number of steps per channel is in the tens of thousands, so if you still get banding at that setting it’s not because of the color gamut but because of something else, such as an image with poor original image quality being stretched too far, or other incorrect editing technique. If done properly, and depending on the specific paper/ink combination, Adobe RGB can be better. The reason is that today, many pro photo printers can print colors outside of sRGB. To fully use the ink color gamut of such a printer, you’ll want to print an image in a gamut larger than sRGB. However, many commercial printing services and lower-end printers are still standardized to receive image data in sRGB to keep things simple, so for them, sRGB is better. (“It depends”) 2) A camera sensor has no inherent color space, but good sensors today can capture some colors well outside of Adobe RGB. That is why some photographers convert raw to the even bigger ProPhoto RGB gamut. Keep in mind that the 3D shape of a color gamut is not uniform, and not close to a sphere or cube…the shape is highly irregular, like a rock. So in reality, although the total volume of Adobe RGB is larger than sRGB or CMYK, the shape of sRGB means there are actually little color bits of sRGB and CMYK that extend beyond some parts of Adobe RGB. When you shoot in raw, the camera keeps the file raw, but the embedded preview typically uses the color gamut set in camera. You complete the conversion of the raw data to RGB in a raw processor. When you shoot in JPEG,the camera starts from its raw file and completes the conversion to JPEG RGB in camera using the color gamut and adjustment settings in the camera. Those are both true for practically every popular camera made going back to the earliest, simplest ones. All must record the raw sensor data first. The main difference is that early digital cameras offered fewer options. They often converted straight from raw to JPEG and only to one color gamut like sRGB. The difference now is those are switchable so you can suppress the in-camera raw conversion, get the actual raw file out of the camera, and on your computer, convert that to any form of RGB, CMYK, or other color gamut you want.
Thanks for another video Keith, I was wondering when yo're using gigapixel AI do you enlarge straight to the desired print size you want or use a step up method of doubling the size each time until you reach your final print size. I had seen a few people saying that it gives better results that way especially if you are making a significant enlargement.
Ah, it does depend on the source image. It's something I might experiment with, however raw processing and pre-enlargement sharpening is critical in this - see the linked article which explores this in some detail. Even choice of raw converter can make quite a difference.
@@KeithCooper Thanks Keith, It does sound like something worth experimenting with. I cant say I've seen many direct comparisons between the two methods beyond some people just preferring one over the other.
My suspicion is that any differences will be very image/source dependent - that makes any more formal comparison very difficult to do in any meaningful or generalisable way.
Hi, Keith! I love your channel. I admire the breadth of knowledge that you have on this subject. If you would allow me to pick your brain a bit, I'd like your opinion and help in choosing a TS lens. I have a GFX100s and an X-T5. Both are outfitted with Fringer Pro EF adapters. Back when I was much younger, I mostly used a TechniKardan 23S. But that was in the pre digital days. I am thinking about getting the Canon 24mm TS-E lens, which could be used with either camera. I am also thinking about waiting for the GF30mm TS to go on sale and springing for that. I am not a professional photographer. I'm retired now. I do travel a lot, and enjoy photography. What do you think? Just ordered your book from Amazon, btw.
Thanks! I suspect you will wait for quite some time for that 30mm TS sale ;-) The TS-E24 is 'equivalent' to 17mm on the GFX and 36mm on the crop - so an interesting option. I just got a TS-E50 for my GFX, works very well and the ~40mm equiv is very nice on the GFX.
The GF30mm ƒ5.6 TS is nowhere to be found. I found a used GF110 ƒ5.6 TS at B&H for $2,764, and decided to take a chance on it. I should have it in a few days. I am still thinking of the wide end of the spectrum. Perhaps the Canon TS-E 24mm? I'm just concerned that it might be too wide on the GFX100S. I could always use it on the X-T5. Any advice? I can get one for about $1,100.
@@michaelmaklakov2111 The TS-E24 works well, but is obviously quite wide. Needs stopping down [10-11] and considerable care in focusing when shifted much. Benefits from sharpening off axis when shifted with something like Sharpen AI - then again, so does the Fuji 30 [but less] Yes, the 30mm is very difficult to find - saw a used one in the UK the other day, but near enough full price
'Recommend' doing some very heavy lifting there ;-) That [as in ppi not dpi] does relate to the pitch of some of their print heads [some are now 300]. Actual testing shows that it's nothing special these days - much of the attachment to such numbers [from both Epson and Canon] is no more than a holdover from years gone by in different industries
@@KeithCooper yes, that was the recommendation on the Epson Print Academy (many years ago on Epson 3800). Still have that printer. Actually multiples of 360 or even 180 dpi when sizing images. I guess as much down to the print driver as the pitch of the print heads.
@@ianbeacham7704 Yes, it had some validity in the past, but it's often taught just because it's an easy answer, and satisfies some parts of the audience's desire for precise numbers ;-) Even in the days of the 3800, I'd have said it was outdated advice - 25 years ago, more reasonable
This is why beginning with the end in mind cannot be understated.
Yes - very true
I read a study of data collected watching patrons in museums view paintings, and it was noted that a person tends to view any image from a distance equal to 2x the diagonal of that painting (or photo).
Thanks - do you have any references for that?
Thank you Keith for this excellent video and so generously sharing your extensive knowledge on printing and photography. After lots of experience printing in the darkroom (years ago) I am relatively new to digital printing, and very much appreciate your discussions on all various aspects. Cheers! GB
Glad it was helpful!
Thanks, Keith! Yet another instructive and encouraging video.
Thanks
I stumbled on your channel thanks to Google's algorithms. I am barely an amateur photographer (I have had to learn how to take photos of my art by necessity and they are not close to what a photographer would do). I also own an Epson SC-P900. I wanted to thank you for providing such amazingly good and to the point information. Every word that comes out of your mouth has meaning, and at multiple levels. Listening to you reminds me of reading Robert Henri's "The Art Spirit." Your knowledge was undisputed but in this video I realized that I am not just listening to a true expert, but that you are also an amazing artist. Thank you
Thanks - that's much appreciated
I also find media and intent surely come into play here. e.g. a canvas print of gouache styled art doesn't need to be super sharp and it still looks good.
Of course, but I'm principally discussing this as a photographer
I find the science of viewing distances oddly satisfying! I remember looking up the near point (the closest distance your eyes can resolve details) and learned it was ~8 in on average.
When I grabbed a ruler and tested, I found it to be roughly true. So for most people's eyes, that chart doesn't need to go lower than 8 inches or 858 ppi.
I'll let the people in 2005 who produced it know ;-)
Of course I've also seen people view 6x4s with a magnifying glass...
A baby can focus on the tip of it's own nose.
You reminded why the yellow panel on the front of a train is 1 square metre. It corresponds to 1 (human retina pixel) at 1 km
Thanks - not sure how you make use of this? ;-)
Thank you🙏🏽
Thanks
Thank you for another practical/helpful printing video.
Glad it was helpful!
Thank you Keith
Thanks!
Do you print often at 600 PPI and editing the images with ACR? I have been using 400 PPI and will experiment with 600 PPI on my EPSON P 900. Great Video awesome work!
It depends on the source image and the size - I print at whatever convenient resolution is there - so no specific number [ACR is part of my Photoshop edit workflow, but nothing to do with sizing)
For a lot more info on the P700/900, read my article about this very topic
www.northlight-images.co.uk/driver-settings-and-print-detail/
Of my couple of thousand negatives, maybe a couple of hundred are worthwhile. But most of them have sentimental value, and are important to me.
Scanning medium format negatives at 600 ppi was a big blunder. Printing at 300 dpi would span 4 x 4 prints.
I now am gradually scanning again, at 4800. The TIF file size is typically 500 MB. So I can print a 13 x 19 nicely at 500 dpi.
The prints look nice (to me) from eight feet away. Looking at them from one foot away seems silly.
Overkill? Sure. But I'm happy.
Yes - I'd scan at such resolution too
Most scanners only have one scanning resolution, it's maximum. When you choose a lower res it simply downsizes the file before outputting it. You therefore always scan at the highest or native resolution of any scanner, it's not something you need to calculate or even think about. You then resize the file if you like, or not.
@@johnsmith1474 Depends on the job at hand. I tend to differentiate between bulk 'cataloguing' type scanning and scanning individual images at optimal - storage and time constraints vary. There is no better way of effectively culling/curating slides/negs than having to scan them.
I've just got an Epson bulk print scanner to test. The speed goes down as resolution increases.
I'll do a quick into for it here...
Group f/64 would beg to differ, but even Ansel backed off from the premise that all must be in sharp focus. People who may actually buy prints don't view them through a loupe--they either like the image or they don't. It just needs to "work" (and part of that is the viewer not noticing technical details, but rather just enjoying the image)...
Yes, a variety of views on 'total sharpness' ;-)
Thankyou very much for this one.
Glad it was of interest
Some time ago I did some research on this topic. I was wanting to see what resolution I should use for large panoramas of 2m or longer. Normal visual acuity is 1 minute of arc. That's a 60th of one degree. This means at 1m we can expect to define pixels at about 90ppi or less. The chart you showed recommending 191ppi at 36" makes sense. Double the resolution to make sure you can't see pixels in an image expected to be viewed from 1m. But here's a thought. What about viewing of a screen? On my 27" 4k computer screens, viewed from a typical distance of 600mm when at my desk, I'm looking at 164ppi and no hint of pixels. Likewise when sitting just 1500mm from my 65" 4K TV, as I am tight now, I'm looking at just 68ppi and again no hint of pixels. I do start to get a hint of rastered text closer than 1.5m though. I think prints from A3 up only need 200ppi at most. I've just received my ET8550 and am just learning to drive it. I have a couple panoramas' I will try printing A4 sections of at as low as 100ppi to see for myself what I can get away with as I would like to print them at least 2m wide, Yes I will have to do multiple prints on this machine. It'll be a fun learning curve I think. I can't thank you enough for the amount of info I have gained from your numerous ET8550 vids as well as others like this one. You're a treasure!
Thanks - always glad to contribute.
Not sure about screens, but I view my 4k 32" screen from 70-80cm - but that depends on which pair of glasses I'm wearing...
Good stuff. I recently discovered that combining various levels of sharpness creates depth/three dimensionality... even for portraits. It's hit and miss... more art than science... but well worth experimenting with.
Thanks for noting that!
re: push button sharpening in LR: I agree that it's important to sharpen artistically on purpose, but there is a whole medium/size/resolution part to print sharpening that was integrated off some old plugins. It's a part of image making that most creators will not know the particulars of and the technical knowledge is getting harder and harder to find online.
When I'm asked, I usually give people a crash course on manually sharpening the original image to completion, then using the baked in presets at print. 😅of course the real answer is more complicated, but everyone has a different appetite for technical complexity
Yes - The subject can soon get quite complex.
In many respects I only expect people to take away a few aspects of videos like this.
I've been asked to write a book on the subject of printing, but the tilt/shift one was hard enough!
@@KeithCooper I was recommended Image Sharpening by Fraser/Shewe in a printing class and finally picked it up.
I def think a full book like your tilt shift one would be a real challenge to write 😅 there's so much about the craft that benefits from hands on real life teaching
@@JohnPurcell Actually I'll admit that there are two specific things putting me off right from the start - one that I'd reluctantly have to include Lightroom, and the second that I've not used a Windows PC this century ;-) Both would require considerable effort/time for me to get my understanding up to a level where I could include them in any more general book.
If I stuck with Photoshop, Macs and a few other bits of software, it's a possible, but the whole field is changing far faster than the world of tilt/shift lenses.
The complexities around when to sharpen, when to resize and how this relates to original files are a chapter alone...
To my recollection, 300 DPI comes from the Apple LaserWriter. I think it's worth noting the print density of your specific printer, multiples of 90 might work better than multiples of 100. It's also possible that my speculation is correct, but the effect not noticeable because the liquid inks tend to blend a little.
LrC can enhance resolution, I pressed the button once to see what happened.
Yes '300' goes back a long way... So far, that almost any time a potential client uses the number I flag it as a potential marker for someone who does not understand what 'resolution' means ;-)
'Best' settings are very much dependent on the actual printer and driver - hence my specific testing of this for many printers.
"Why would I sharpen clouds in the sky?" That's what I said to my first client :) But it needed showing him oversharpened clouds (and remembering him that some of audition could be younger and could see it much faster) to show the effect.
Yes - one of those things where if you can see it, it's too much ;-)
@@KeithCooper Like a makeup. If anybody see it, it's time to wash it off.
I met this wildlife photographer who does big outdoor exhibitions worldwide and he was very "megapixels don't matter", he used to blow things up really huge, was about 2006 I met him.
Yes - knowing just where your prints will be viewed can be a big help.
It's very easy to say "my camera isn't good enough" when it really is.
@@KeithCooper I suppose the lens and settings (f stop etc) become more important at larger sizes
To some extent, but composition and subject matter choices are the real drivers, once you've a basic technical competence.
@@adamarmfield1069 I have some seriously good lenses, but sharpness, chromatic and other aberrations, vignetting were never a consideration in my choice. I chose zooms, whenever possible, with a fixed maximum aperture because I don't want my exposure to change when I zoom. I used Canon's fantastic plastic F1.8, that's fast enough and the lens is cheap enough you could carry a spare or two. I have TS-E lenses because i want the shift
Informative information Keith - What are your thoughts though on actual final output PPI for printing? I let the PPI go where it wants to go, I don't resample at all up or down. I have made prints at 190 ppi which are excellent in quality. I think a lot comes down to lens quality too, and like you say intended viewing distance. What do you do with small prints, as you get some massive PPI if you don't resample at all - is this using more ink in not interpolating?.....would like to know your thoughts on final output if you let the PPI go where it wants to go or change it. I use my own printer and leave it go where it wants' to, and my professional printer does this as well.
Yes, I've made good prints at low ppi in the past, but with modern software, I'd likely consider resampling up to 450 or more. However, that depends on the source image and print size, and as you say, lens characteristics are part of the mix.
My large 47 foot long panoramic was printed at 108ppi [due to a file/driver size limit] but resampled in the driver to 300 [Canon iPF8300] - People walked right up to that one ad no-one noticed any lack of detail and many commented how much detail there was...
There are no correct answers in this sort of stuff - something I know makes some uncomfortable ;-)
As to ppi and ink, it makes no difference - the driver resamples anyway, as part of the dithering and conversion from RGB pixels to ink dots
@@KeithCooperthanks Keith - At the end of the day then if I let the PPI goes where it wants to go then it’s the same as setting to an arbitrary PPI? I haven’t don’t elaborate tests on this. But have been happy with the results. I go into photoshop and turn off resample and set the print dimension size which gives me the PPI. I then make a new canvas with the PPI I got WITHOUT resampling. I will have a look at your website posts too, I believe you said you’d done some tests, would also be interested in knowing your upsizing software you recommend too.
Yes - the key is always to test and decide if it's worth the effort - Most of the time with my newer [more MP] images I do just let the PPI be whatever it is. It's older, lower MP images where I'll take a look at scaling, but it so much depends on the source image [the original raw file]
I tend to use gigapixel AI these days, albeit not the current version, for upscaling - see the various links in the notes to this video
Hi Keith. Why not just print at highest possible ppi despite size of image, as we are not ultimately sure what distance some of the images may be positioned at at home/show, etc? at highest ppi, even the larger image is then viewable at "any" distance..
Yes, in general, that's why I suggest printing at the full available real resolution above 300.
The tests for different printers show that this introduces some visible benefits up to say 6-700. [there are differences - hence my set of articles looking at this for different printers]
With modern resizing software I'll consider 'creating' more detail for this, but it does very much depend on the original image, size and the printer.
However, I'm not assuming any particular software, which people might or might not have available for their use, so as to make this a bit more generalised for people
@@KeithCooper I shoot at 50MB (Sony A1) all the way down to 12MB (DJI Drone Mini 3 Pro) and aim to get max resolution in all originals, but am only now thinking about a great printer for all this. You have been inspirational and very informative , so huge appreciation for taking the time to educate. At the end of watching many videos and also your impressions, Im now looking at a Canon Pro300 for A3+ and down, and would send any larger print formats to a professional print house. Im based in Cape Town, South Africa. Thanks Keith!
Thanks, Keith.
Interesting and helpful, as usual.
I have a couple of questions. Perhaps not strictly on topic.-
1) I'm told that printing from an Adobe rgb file is better than printing from sRGB.
Because both contain 256 shades of each color, and the Adobe RGB color space is bigger the differences between colors is greater. Are we more likely to see banding in the transitions on a large print when we make a print from Adobe RGB?
2) Are our cameras capable of recording 100% of the Adobe RGB color space when we select to shoot in Adobe RGB as jpg? Same question when we opt for RAW or sRGB.
Thanks.
Thanks
Not topics for concise or simple answers ;-)
1 No - banding is quite a complex issue, not primarily related to this.
Better? - depends on the gamut and content of the image.
2 The camera setting only affects the jpegs - no effect on raw files other than flagging the info, and [maybe - never tested it] the embedded jpeg thumbnail
Given the camera can capture a very wide range of colours, it depends on how the camera maker decides to handle/process raw sensor data, as well as how raw conversion is carried out.
Note that in both cases I might also query the actual questions ;-)
So, for 1 who says it, what evidence do they put forward and under what print conditions? ;-)
For 2 - what camera, what sensor technology.
@@KeithCooper Thanks, Keith, for your very speedy reply.
Question #1 was prompted by my curiosity. Adobe RGB is a larger color space than sRGB, yet each is described by the same number of shades. It seems reasonable to me that colors in sRGB are more tightly defined and smooth transitions would be easier. Also, it would seem to me that any deficiencies in the reproduction of gradients might be more obvious in a large print, especially when the viewing distance is intentionally close.
Typically, I print no larger than 13X19 and only from sRGB files. So I do not have any first hand experience with this concern.
Question #2: I have a couple of old cameras.
Sony NEX 6 and Sony SLT A55.
I searched the manufacturer's and other sites, but couldn't find information about how these camera convert color, or how RAW processors convert color out of the camera.
Thanks again for your attention and considered insights.
1 dithering and the print driver... lots of factors, it's not a problem but when it is, it's likely not for this reason.
2 - you won't it's proprietary info ;-)
For both [and lots more ;-) ] get a copy of 'real world color management' by Fraser et al. Also, this has a lot of links:
en.wikipedia.org/wiki/Raw_image_format
BUT, be prepared to delve into a lot of stuff...
@@KeithCooper Thanks, Keith.
Hope you don’t mind me jumping in…
1) Adobe RGB vs. sRGB for printing: Note that the 256 levels per channel is assuming 8 bits per channel. If captured and edited at a higher bit depth such as 16 bits per channel, the number of steps per channel is in the tens of thousands, so if you still get banding at that setting it’s not because of the color gamut but because of something else, such as an image with poor original image quality being stretched too far, or other incorrect editing technique.
If done properly, and depending on the specific paper/ink combination, Adobe RGB can be better. The reason is that today, many pro photo printers can print colors outside of sRGB. To fully use the ink color gamut of such a printer, you’ll want to print an image in a gamut larger than sRGB. However, many commercial printing services and lower-end printers are still standardized to receive image data in sRGB to keep things simple, so for them, sRGB is better. (“It depends”)
2) A camera sensor has no inherent color space, but good sensors today can capture some colors well outside of Adobe RGB. That is why some photographers convert raw to the even bigger ProPhoto RGB gamut. Keep in mind that the 3D shape of a color gamut is not uniform, and not close to a sphere or cube…the shape is highly irregular, like a rock. So in reality, although the total volume of Adobe RGB is larger than sRGB or CMYK, the shape of sRGB means there are actually little color bits of sRGB and CMYK that extend beyond some parts of Adobe RGB.
When you shoot in raw, the camera keeps the file raw, but the embedded preview typically uses the color gamut set in camera. You complete the conversion of the raw data to RGB in a raw processor.
When you shoot in JPEG,the camera starts from its raw file and completes the conversion to JPEG RGB in camera using the color gamut and adjustment settings in the camera.
Those are both true for practically every popular camera made going back to the earliest, simplest ones. All must record the raw sensor data first. The main difference is that early digital cameras offered fewer options. They often converted straight from raw to JPEG and only to one color gamut like sRGB. The difference now is those are switchable so you can suppress the in-camera raw conversion, get the actual raw file out of the camera, and on your computer, convert that to any form of RGB, CMYK, or other color gamut you want.
Thanks for another video Keith, I was wondering when yo're using gigapixel AI do you enlarge straight to the desired print size you want or use a step up method of doubling the size each time until you reach your final print size. I had seen a few people saying that it gives better results that way especially if you are making a significant enlargement.
Ah, it does depend on the source image. It's something I might experiment with, however raw processing and pre-enlargement sharpening is critical in this - see the linked article which explores this in some detail. Even choice of raw converter can make quite a difference.
@@KeithCooper Thanks Keith, It does sound like something worth experimenting with. I cant say I've seen many direct comparisons between the two methods beyond some people just preferring one over the other.
My suspicion is that any differences will be very image/source dependent - that makes any more formal comparison very difficult to do in any meaningful or generalisable way.
Hi, Keith! I love your channel. I admire the breadth of knowledge that you have on this subject. If you would allow me to pick your brain a bit, I'd like your opinion and help in choosing a TS lens. I have a GFX100s and an X-T5. Both are outfitted with Fringer Pro EF adapters. Back when I was much younger, I mostly used a TechniKardan 23S. But that was in the pre digital days. I am thinking about getting the Canon 24mm TS-E lens, which could be used with either camera. I am also thinking about waiting for the GF30mm TS to go on sale and springing for that. I am not a professional photographer. I'm retired now. I do travel a lot, and enjoy photography. What do you think? Just ordered your book from Amazon, btw.
Thanks! I suspect you will wait for quite some time for that 30mm TS sale ;-)
The TS-E24 is 'equivalent' to 17mm on the GFX and 36mm on the crop - so an interesting option. I just got a TS-E50 for my GFX, works very well and the ~40mm equiv is very nice on the GFX.
@@KeithCooper Thanks! I wasn't thinking about that lens, but maybe I should take a look at one. I seems like it would be a better fit for me.
The GF30mm ƒ5.6 TS is nowhere to be found. I found a used GF110 ƒ5.6 TS at B&H for $2,764, and decided to take a chance on it. I should have it in a few days. I am still thinking of the wide end of the spectrum. Perhaps the Canon TS-E 24mm? I'm just concerned that it might be too wide on the GFX100S. I could always use it on the X-T5. Any advice? I can get one for about $1,100.
@@michaelmaklakov2111 The TS-E24 works well, but is obviously quite wide. Needs stopping down [10-11] and considerable care in focusing when shifted much. Benefits from sharpening off axis when shifted with something like Sharpen AI - then again, so does the Fuji 30 [but less]
Yes, the 30mm is very difficult to find - saw a used one in the UK the other day, but near enough full price
so its not the size it is, it's where you put it🙊
Indeed...
Will printing at a lower resolutions use less ink?
Not in any significant way - print in draft mode to use less ink and produce dismal looking pictures :-)
Epson recommend 360 dpi
'Recommend' doing some very heavy lifting there ;-)
That [as in ppi not dpi] does relate to the pitch of some of their print heads [some are now 300].
Actual testing shows that it's nothing special these days - much of the attachment to such numbers [from both Epson and Canon] is no more than a holdover from years gone by in different industries
@@KeithCooper yes, that was the recommendation on the Epson Print Academy (many years ago on Epson 3800). Still have that printer. Actually multiples of 360 or even 180 dpi when sizing images. I guess as much down to the print driver as the pitch of the print heads.
@@ianbeacham7704 Yes, it had some validity in the past, but it's often taught just because it's an easy answer, and satisfies some parts of the audience's desire for precise numbers ;-)
Even in the days of the 3800, I'd have said it was outdated advice - 25 years ago, more reasonable