How to print large files?

[Sweep]

Following on from my previous thread about Lightroom and viable alternatives for LF in B&W and colour, with the support of the forum members I managed to get a legitimate download copy of Photoshop CS2.
I can now do what I originally intended which is to correct the highlight and shadows etc. Pretty simple stuff I guess.
As I have been scanning at 4000dpi for maximum detail, which equates to 40,000 x 32,000 pixels on 10x8, I have been getting file sizes of around 1.2gb for B&W. I haven't actually scanned any colour yet but that would be limited to 2000dpi and 2gb by my scanner.
My new problem is that PS CS2, whilst loading fine, wont output to print an images with any 'length' above 30,000 pixels; the 'Print' button is just greyed out. I also downloaded Epson Print Layout (EPL) but that wont even load the file. EPL did manage to load a 5x4 Velvia scan at 4000dpi with a file size of 729mb.
Am I missing something or what? One of the reasons for me getting the 10x8 was to record huge amounts of detail, slow down my whole picture taking process, improve my technique, and have access to movements etc. Now it appears I can't print any of my pictures or, I presume, will have to reduce the resolution of my scans which defeats the object of doing 10x8. I may as well have bought a 5x4 :(
So, how do I set up to print one of these large files and have simple layout possibilities, such as image size on paper and margin size, and be able to load ICC profiles?

Thanks ...Sweep

[Greg]

Little bit of stepping back a bit here.... I figured out what the largest print size that I would be printing 99% of the time. Image files should be as large as to print at the chosen final print size at 300ppi. Large files as yours will contain a whole lot more information, but you won't see it in the final print. For the other 1% of the time when I will be making really large prints, I just rescan the film. Sure others will have different opinions....

[chassis]

I have printed (at a third party printer) 20" x 24" using the resolution and file size idea outlined by Greg, and am satisfied with the results.

[Pere Casals]

40,000 x 32,000 pixels on 10x8

You should send to printer an smaller file that just equals the number of pixels that the printer is to actually print. If you send more pixels than the printer is capable to put on the paper then the printer reduces the image size internally.

The image reduction algorithm the printer uses can modify the look... then it is better you reduce the file size in your computer to be sure about what you will obtain.

In photoshop yo go to main menu->Image->Size and a dialog appears. Then you select the new size, at the bottom of the dialog you have a combo box to select the resizing algorithm, don't use the default one, select "Bicubic, ideal for reductions", this will mantain a sharper look, (if you want sharpness).

After reduction to final size, go to Filter->Sharpening and see if some amount of digital sharpening still benefits your process or not.

There is some confusion between DPI and PPI, dots per inch vs pixels per inch. A printer may stamp several dots to build a pixel. So don't get wrong with that, know how many PPI is delivering you printer and make an image with same number of pixels that are to be stamped, one source pixel for each printed PPI !!!

Here you have a guide to do that: https://www.scantips.com/calc.html


ICC: There is the Soft Proofing technique... A printed image may look very different from what is seen in the monitor. First is that colors that a monitor and a printer may show are not the same, some printed colors can not be displayed by a monitor and the conunter, and then you have the calibration. Photoshop can be adjusted to more or less showing a simulation of the printed look, this is not difficult but it requires learning Soft proofing and ICC usage, and calibrating your monitor with a basic color calibration device.

One easy thing to start is to print a dozen of small same images in the same print, with different levels of saturation, contrast etc, you inspect it with a magnifier and decide. At the end, for critical jobs, it is something it is done before making a big print because soft proofing is not perfect.

If you don't use Photoshop, do it. Not necessary last version, CS2, 4 or 6 are very powerful. Learnig that is best investment you can do if you deal with digital images.

Regards.

[aclark]

Sweep, I am not sure why you are scanning your 10x8 negative at 4000ppi. A file this big will allow you to make a print about five feet wide with over 300ppi going to the printer. Do you really want to make prints this big? If not, then all you need do is scan at a smaller ppi number. If you scan at 1200ppi for eg, you will still have a file big enough to send 360ppi to the printer when doing a 30 inch wide print. If you think you won't ever be doing a print as big as 30" x 24" then scan at an even smaller ppi.

Alan

[paulbarden]

It would be very useful to know how large you want to print your images. Let’s start there?

[profvandegraf]

One thing to consider for final print resolution and size is the viewing distance. An 8x10 print at 300dpi will look great from a foot away, but a 30x40 print will be viewed from 6 feet or more. in my experience you will be fine with lower resolutions at larger print sizes. As always your mileage may vary.

Prof

[Pere Casals]

One thing to consider for final print resolution and size is the viewing distance. An 8x10 print at 300dpi will look great from a foot away, but a 30x40 print will be viewed from 6 feet or more. in my experience you will be fine with lower resolutions at larger print sizes. As always your mileage may vary.
Prof

It is not the same PPI than DPI, depending on system you may need 1200 DPI to obtain 300 PPI

https://www.andrewdaceyphotography.com/articles/dpi/

300 PPI is good, even 150 PPI may be OK, 300 DPI may be too low resolution.

[profvandegraf]

It is not the same PPI than DPI, depending on system you may need 1200 DPI to obtain 300 PPI

https://www.andrewdaceyphotography.com/articles/dpi/

300 PPI is good, even 150 PPI may be OK, 300 DPI may be too low resolution.

I was referring to the output resolution of the print. Of course different devices image differently.

Prof

[Pere Casals]

I was referring to the output resolution of the print. Of course different devices image differently.

Prof

Ok, but remember that if you speak about printers and you say dpi you are saying dots, that may be black or white, and may have no gray levels, if speaking on printers it is very different PPI than DPI. When speaking on scanners a PPI and a DPI is the same.

It is confusing because DPI term is normally understood different for scanners than for printers, so for printers one should make it clear, as printer manufacturers like to speak about DPI because the resulting number is more fancy.

Regards

[Sweep]

Hi Guys,
First of all, if it is not already apparent, I am a complete novice to all this digital printing stuff so all your help is very much appreciated.
To some it may seem insane to scan at such a high resolution (?) but I want to record, by the action of one single scan, the maximum amount of detail I can. This is no different to my film choice.
After that I can selectively choose any crop, or many, I want from the film and print these in the knowledge that I will have sufficient quality available. I don't really like the idea of scanning at ideal print resolution (?) of 300ppi as any crop would result in this figure dropping off. On top of that, mounting a piece of film to the drum takes just as much time for 300ppi as it does for 4000dpi; I just have longer to do other stuff when scanning at 4000 such as bake a cake, build a wall, read War and Peace etc. :)
Following Peter's advice I found the Image Size box in PS CS2 and had a play with it and, sure enough, by adjusting this, by trial and error, not maths, to 450ppi and making a copy it imports into Epson Print Layout and fills a A3 sheet with 10mm border top and bottom at 315ppi.
So, it looks like I am getting a little more of an understanding of things so I think my plan will be:
1. Scan at highest resolution available
2. Crop any part of the image that look like it might print nicely
3. Try figure out what ratio the crop size is to the paper size and adjust ppi accordingly
4. Save a copy of this in a new file
5. Leave original high resolution original scan in 'archive'

I can print up to 17" wide on the Epson SC-P800
I just had a look at one of the A3, 10mm top/bottom margin, prints I did by reloading into Epson Print layout and it shows at 2328ppi although what the printer actually does with this I don't know!
"Learning (Photoshop) is best investment you can do" PCasals. I have taken this on board.

Thank you all

[Mike Campbel]

Hi Sweep!
Nice and detail plans! I'll try something like you said. Could you give few examples of your results? I have Canon Pixma MG7550 and tried to print up to seventeen and have nice pictures.

[faberryman]

As I have been scanning at 4000dpi for maximum detail, which equates to 40,000 x 32,000 pixels on 10x8, I have been getting file sizes of around 1.2gb for B&W.
Which equates to a 9' by 11.5' print at 300. If you print smaller than that, you downsize and discard all the resolution you spent time scanning for.

[justiny]

Wow. Nice info. I print 20x60. What's the most effective ( sharpest) way to scan with an X1? I usually scan with max resolutions. Thank you in advance.

[Sweep]

Which equates to a 9' by 11.5' print at 300. If you print smaller than that, you downsize and throw away all the resolution you spent time scanning for.

Well, I can't argue with that but please read my reply at post 11.
What I don't understand, and I am a complete novice you understand, is why people would embrace LF photography and then not wish to record as much information as possible when scanning. I may as well just trade the whole lot in and get myself a cheap 35mm or low end digital. Life would be a lot easier, and definitely cheaper, and I'm sure the results would be exactly the same.

[Pere Casals]

Wow. Nice info. I print 20x60. What's the most effective ( sharpest) way to scan with an X1? I usually scan with max resolutions. Thank you in advance.

An X1 will deliver less than effective 2000 dpi for 4x5, as sensor has 8000 pixels and you have 4", so divide... your x1 image will have some 8000 pix for the 4" side of the 4x5" negative.

If (for example) you are to print 20cm at 300 ppi you have 7.874" , so multiply by 300 pix per inch, then you need an image with 2362 pixels for the side that has 20cm.

So make a file with 2362pix and print it 100% size at 300ppi (it can be higher dpi at the same time, more than one dot per pixel) and you'll have 20cm in that side.

Regards

[justiny]

I am sorry. It's 20 inches x 60 inches. so... 300 ppi times 20 inches. 6000 pixels and 18,000 pixels for 60 inches. Then x times y is equal to 108,000,000 pixels. It means 108mb with 300 ppi is optimal. Is that right? wait. no. how to convert or translate 108 mega pixels?
I got it. 12 megal pixels are equivalent to 36 mega bytes. so... 108 mega pixels means 324 mega bytes. Ok. So I need a 324 mb file to obtain the best result. Got it!! Thank you so much!!

[justiny]

Well, I can't argue with that but please read my reply at post 11.
What I don't understand, and I am a complete novice you understand, is why people would embrace LF photography and then not wish to record as much information as possible when scanning. I may as well just trade the whole lot in and get myself a cheap 35mm or low end digital. Life would be a lot easier, and definitely cheaper, and I'm sure the results would be exactly the same.

I agree with Sweep. I do the same. I can always downsize but not upsize without re-scanning. And I am not throwing away my time. I have a designated scanning station. I have other computers to do other things.

[EH21]

I don't think you throw away details by downsampling but the algorithm used may make differences in the results - eg. pronounce the grain, edges of certain frequency. The bicubic is one of the oldest routines, but there are newer methods such as S-spline and lanzos. You may also find when downsampling that doing it in steps can yield better results than doing it all in one go. Experiment and see what works for you.

[Pere Casals]

I am sorry. It's 20 inches x 60 inches. so... 300 ppi times 20 inches. 6000 pixels and 18,000 pixels for 60 inches. Then x times y is equal to 108,000,000 pixels. It means 108mb with 300 ppi is optimal. Is that right? wait. no. how to convert or translate 108 mega pixels?
I got it. 12 megal pixels are equivalent to 36 mega bytes. so... 108 mega pixels means 324 mega bytes. Ok. So I need a 324 mb file to obtain the best result. Got it!! Thank you so much!!

This is... if you send an smaller file the printer will extrapolate, if you send a larger file then printer will interpolate. Good printers today have very good resizing algorithms in the software drivers, so not big problem.

Anyway if you are the kind of guy that uses a different sharpening algorithm or setting for eyes than for the cheek you may want to control sharpness it in PS and not in the printer.

In reality this may be noticed by a close inspection, if a big print is viewed from a "normal" distance then the pixel level sharpening is not seen, so it is the same if it is made by the printer or by PS, then just it is necessary to understand how much information you have to send to not limit the printer capability.

I would make clear that when delivering an image pixel for each "printer pixel" (that can have multiple dots) we are talking about the pixel level sharpness, that can be important for the ultimate look in a close view.

[Pere Casals]

The bicubic is one of the oldest routines, but there are newer methods such as S-spline and lanzos.

Normally we will be downsizing from an LF scan to the printer, if not we should scan denser.

If we are talking about downsizing then you will see little difference bicubic vs S-spline, the important thing is that the algorithm is optimized for reductions, this is using "bicubic, optimal for reductions" (and not the other bicubic) at the bottom of the PS size dialog.


If we are to enlarge the image then you may notice S-spline vs bicubic benefit beyond 200% linear size increase.
You can judge here:

http://ronbigelow.com/articles/interpolation2/interpolation2.htm

[Luis-F-S]

I use an enlarger so don't have to worry about file size

[Johnny LaRue]

I use an enlarger so don't have to worry about file size

Another helpful gem; I'm sure happy there's an ignore button, I'm putting it to use, pronto!

[Luis-F-S]

Another helpful gem; I'm sure happy there's an ignore button, I'm putting it to use, pronto!

There's more than one way to skin a cat........You can ignore this one also. L

[Pere Casals]

I use an enlarger so don't have to worry about file size

I was in the digital side, but I'm going to the enlarger side. I still lack an LF enlarger, but 810 can be contacted. Now I think in scanners and printers as useful tools to print contrast masks in the Alan Ross way... digital is amazing for that ! http://phototechmag.com/selective-masking-part-iii-computer-techniques-for-the-traditional-darkroom/

[interneg]

An X1 will deliver less than effective 2000 dpi for 4x5, as sensor has 8000 pixels and you have 4", so divide...

The X1/ X5 runs at 2040ppi for the 4x5 setting - it's covering 99.6mm width (the actual specified width of 4x5 film!) at 2040ppi vis-a-vis 25mm at the 8000ppi setting on the X5. The X1 omits the Hi-Res mode & adds 5x7 at 1600ppi. The optical system tests out in the 6900ppi range I recall - so the scanner lens is doing 135 lines/mm - & that means that at 2040ppi, it should be easily able to match the manufacturers' spec - and this has been borne out by most of the reputable testing.

Having been running a Canon Pro-2000 for the last few months, there's definitely a case (with the right substrate & a file of suitable quality) to go well over 300ppi when making smaller prints - while there are limits (probably somewhere over 400 & under 600), it does seem to make the prints look better in certain aspects - much like Ctein argued was the case with Epson printers. QTR/ Piezography folk seem to be able to go a little further still. Whether the end audience will actually notice/ care is a different question...

[Pere Casals]

Having been running a Canon Pro-2000 for the last few months, there's definitely a case (with the right substrate & a file of suitable quality) to go well over 300ppi when making smaller prints - while there are limits (probably somewhere over 400 & under 600), it does seem to make the prints look better in certain aspects - much like Ctein argued was the case with Epson printers. QTR/ Piezography folk seem to be able to go a little further still. Whether the end audience will actually notice/ care is a different question...

IMHO PPI from a Lambda or Lightjet can be more "effective" than from an inkjet. A Pixel from an inkjet is usually made from a cluster of dots, and that cluster can be more or less fuzzy, depending on the particular printer. While 300 PPI from a Lightjet may be "the perfection" an inkjet may need higher PPI to match the laser on silver job.

As we are talking about PPI (one PPI may include several dots "dpi") with continous tone, 150ppi are quasi 6 pixels per mm that can print 3 line pairs in a mm. IMHO if we have a very strong microcontrast area (with complete black and white) we may be able to notice and improvement at 300 ppi from 150ppi, but this is not the general case in photography, most real scenes are continous tone with much lower than 1:100 on print micrrocontrast from one pixel to the next, and in those conditions human vision is not as good to see the flaw.

Clearly it also may depend on the viewer, one can have 80% or 130% sight score, and a near sighted person may see the print closer... So I agree, not much improvement beyond 300ppi, it has to be a well sighted viewer and a print with particular lack of continous tone from one pixel to the next to see improvements beyond 300, or as you say beyond 400-500 from an inkjet.

Recently I made a test of 150 vs 300 ppi from an Epson P9000... I also showed the man managing the printer a 8x10 contact print with a 8x magnifier :) this is some 1500 ppi quality. He was amazed with what he saw with the magnifier :), but in reality without the magnifier the 300ppi print was looking as sharp than contact print, IMHO perceived sharpness has strong factors other than resolving power... and digital process adds powerful sharpening tools...




The X1/ X5 runs at 2040ppi for the 4x5 setting - it's covering 99.6mm width (the actual specified width of 4x5 film!) at 2040ppi vis-a-vis 25mm at the 8000ppi setting on the X5. The X1 omits the Hi-Res mode & adds 5x7 at 1600ppi. The optical system tests out in the 6900ppi range I recall - so the scanner lens is doing 135 lines/mm - & that means that at 2040ppi, it should be easily able to match the manufacturers' spec - and this has been borne out by most of the reputable testing.



The optical system testing 6900ppi is when scanning 35mm film, but as zoom is applied to take 4" then resolving power "on film" of the lens will decrease dramatically in terms of effective on film ppi. The hardware 2040ppi of the X1 (for 4x5") may be resolving under 35 lppmm in USAF 1951 terms, perhaps 1800 effective ppi. Also an scanner resolving x lppmm will degradate the IQ of a media also resolving the same x lppmm.

What I suggest is that while the lens resolves 135 line pair per mm on film when taking 24mm (of 35mm film) it will resolve much less on film when zoomed out to take those 99.6mm of 45 sheets.


IMHO the X1/X5 is more intended for MF or 35mm, Hassy is in the MF business... for this reason even a cheapo V850 can (slightly) beat it with LF in resolving power terms. Still the amazing DMax capacity of X1/X5 may be much more important for some slide shots !!! So at the end the X1/X5 is great for 45 because few print or negatives may require more pixels.

Well, IMHO we can say that a true drum is a true drum, problem is that mounting task may not be worth for a particular job.

[Sweep]

"Well, IMHO we can say that a true drum is a true drum, problem is that mounting task may not be worth for a particular job"

But mounting does get easier and quicker, and the inclusion of dust and bubbles reduces, the more films you mount and the practice you get.
I have had my drum scanner for around 2 months and scan pretty much every 10x8 I take.
The first mount, which was a strip of three 6x7, took probably 3/4 of an hour and had bubbles, dust, tape on the image area, and was on and off the drum more than once before I got it anything like.
The 10x8s are harder to mount but I'm slowly improving my technique and getting cleaner mounts every time. I reckon I can get one mounted in about 15 to 20 minutes which I don't expect to reduce as long I improve the quality of the mount

[Pere Casals]

But mounting does get easier and quicker, and the inclusion of dust and bubbles reduces, the more films you mount and the practice you get...

Well, I don't have a drum, and you know what one has to pay for an 8x10 scan...

People that has a drum and a fatbed does 90% of the scans with the flatbed.

Sometimes the negative has less IQ than the flatbed is capable, not all shots are technically perfect. Sometimes the final output does not require effective 5000 dpi from a 8x10" sheet. And a lot of sometimes the negative is 1.8D only, making the PMT little difference...

In those conditions the important thing is Photoshop cooking. Of course there are jobs where a drum performs amazingly better, but only a share of the times, IMHO.

[Sweep]

My comment was only in reply to your post that the "mounting task may not be worth for a particular job" and you will notice that I did not extol the virtues of one system over another. Frankly I would have no idea what I was talking about!
I defer to your expertise on the merits of both drum and flatbed scanners.
I do thank you, however, for the time you have taken to draft your replies to this thread. It is really easy to fire questions into the forum but the difficulty is finding someone, like you and others, with the knowledge and inclination to draft a considered response.

...Sweep

[interneg]

As we are talking about PPI (one PPI may include several dots "dpi") with continous tone, 150ppi are quasi 6 pixels per mm that can print 3 line pairs in a mm.

The Lambdas etc are essentially filmsetting machines running RGB rather than single colour laser/ LED etc output. Essentially they use a stochastic screen (FM screening), which if done right will pretty closely resemble what you might perceive as 'film grain'. Inkjets also use a stochastic screen but can vary the droplet size - which essentially hybridises in elements of AM screening.


perceived sharpness has strong factors other than resolving power... and digital process adds powerful sharpening tools...

Artificial sharpening becomes very obvious, very fast - far too many people grossly oversharpen.


The optical system testing 6900ppi is when scanning 35mm film, but as zoom is applied to take 4" then resolving power "on film" of the lens will decrease dramatically in terms of effective on film ppi. The hardware 2040ppi of the X1 (for 4x5") may be resolving under 35 lppmm in USAF 1951 terms, perhaps 1800 effective ppi. Also an scanner resolving x lppmm will degradate the IQ of a media also resolving the same x lppmm.

What I suggest is that while the lens resolves 135 line pair per mm on film when taking 24mm (of 35mm film) it will resolve much less on film when zoomed out to take those 99.6mm of 45 sheets.

Nope. It doesn't 'zoom'. The whole thing moves internally on leadscrews to change magnification ranges. That's why it's so tall. More to the point, there's plenty of lenses out there that were designed to meet a spec demanding consistent performance between a 1:1 and 5:1 reduction ratio. No evidence exists to suggest that the Hasselblad is any different. All the available tests show that it delivers the resolution it claims at both the 120 and 4x5 settings & comes incredibly close for even the highest resolution setting (and that may have as much to do with limitations in the filmholders etc than anything else).


the X1/X5 is more intended for MF or 35mm, Hassy is in the MF business... for this reason even a cheapo V850 can (slightly) beat it with LF in resolving power terms. Still the amazing DMax capacity of X1/X5 may be much more important for some slide shots !!! So at the end the X1/X5 is great for 45 because few print or negatives may require more pixels.

Wrong. The Epson might slightly outresolve the X5 on an ultra high contrast test chart, but it's a soft aberrated mess otherwise. I suspect that the MTF graph for the Epson's optical system is horrifically awful. No amount of BS in PS or voluble sophistry will solve that. The filmscanner.info test of the X5 compared it directly to an A3 Epson which had been tested to nominally outresolve the X5 on a test chart, but even when the Epson had benefit of being downsampled to 2040ppi, the X5 blew it away in sharpness, colour, dmax. And that's against the only Epson that actually has a focusing system.

You should also ask yourself this: why do people seem to prefer the Heidelberg Tango to the Scanmate 11000 (to go from a recent thread) if the Heidelberg is outresolved by the Scanmate?


Well, IMHO we can say that a true drum is a true drum, problem is that mounting task may not be worth for a particular job.

For speed & quality trade-offs, the Hasselblad & some high end flatbeds can be unbeatable - until you have to clean the dust off the file - then a good fluid mount becomes a suddenly much more attractive proposition.

[bob carnie]

Quote Originally Posted by Pere Casals View Post
As we are talking about PPI (one PPI may include several dots "dpi") with continous tone, 150ppi are quasi 6 pixels per mm that can print 3 line pairs in a mm.
The Lambdas etc are essentially filmsetting machines running RGB rather than single colour laser/ LED etc output. Essentially they use a stochastic screen (FM screening), which if done right will pretty closely resemble what you might perceive as 'film grain'. Inkjets also use a stochastic screen but can vary the droplet size - which essentially hybridises in elements of AM screening.

No Lambdas use three lasers that combine the light into a single beam of energy to expose any emulsion you put in front of it.. . There is no Stochastic Screen involved with the Lambda Process.

If for example you print a high resolution scanned image what one sees on the print is the original film grain , not pixels or dots. Print resolution is determined by file quality and I have seen problems with small , poorly processed files .

[Pere Casals]

until you have to clean the dust off the file - then a good fluid mount becomes a suddenly much more attractive proposition.

I solved that completely with $60 HEPA filtering system combined with a suitable "clean room". No dust at no stage. Zero dust...



Artificial sharpening becomes very obvious, very fast - far too many people grossly oversharpen.

I completely agree... Anyway Ps sharpening tools are amazing, a lot of sharpening tools/settings are there and careful selective usage may solve problems in images. It is a matter of dose and smart local application.



Nope. It doesn't 'zoom'. The whole thing moves internally on leadscrews to change magnification ranges. That's why it's so tall. More to the point, there's plenty of lenses out there that were designed to meet a spec demanding consistent performance between a 1:1 and 5:1 reduction ratio. No evidence exists to suggest that the Hasselblad is any different. All the available tests show that it delivers the resolution it claims at both the 120 and 4x5 settings & comes incredibly close for even the highest resolution setting (and that may have as much to do with limitations in the filmholders etc than anything else).


Variable magnification system in X1/5 has a nice Linos (Rodenstock) lens inside (V850 is not as luxurious), but it is very easy to understand that while lp/mm on sensor can be the same for 45 than for 35mm it happens that the "on film" resolved lp/mm decrease dramatically to 1/4 of the nominal when magnification is 1/4 od the nominal, this is field changes from 1" to 4".

Look, if X5 (for 4x5) has 2040 pixels for each inch of film it wont be able to see well 2040 lines (or 1020 line pairs) in that inch, because in those condirions you have a very severe aliasing, as the pixel can be on a line or just in taking half black and white lines and delivering a perfect grey: this is it has 40 pixel pairs per mm, and a pixel pair per mm is not able to scan a line pair without severe aliasing.

V850 has a +36000 pix rgb sensor, while the X5 is 8000, this is what makes a difference for LF and makes X5 oriented to MF and 35mm. Why do hasselblad not use a 36k pix sensor ??? because they are not much interested in the LF market !!! IMHO.








Wrong. The Epson might slightly outresolve the X5 on an ultra high contrast test chart, but it's a soft aberrated mess otherwise. I suspect that the MTF graph for the Epson's optical system is horrifically awful. No amount of BS in PS or voluble sophistry will solve that. The filmscanner.info test of the X5 compared it directly to an A3 Epson which had been tested to nominally outresolve the X5 on a test chart, but even when the Epson had benefit of being downsampled to 2040ppi, the X5 blew it away in sharpness, colour, dmax. And that's against the only Epson that actually has a focusing system.

You should also ask yourself this: why do people seem to prefer the Heidelberg Tango to the Scanmate 11000 (to go from a recent thread) if the Heidelberg is outresolved by the Scanmate?

You overlook V850 capabilities... Sure that in very dense slides PMT and flextight makes a difference, I've seen that with velvia. But what's for BW and LF... V850 rocks. I fully agree with this Petapixel test:

https://petapixel.com/2017/05/01/16000-photo-scanner-vs-500-scanner/

https://petapixel.com/assets/uploads/2017/05/v700_1250_sharpening_before_resize-792x800.jpg

https://petapixel.com/assets/uploads/2017/05/sidebyside-800x549.jpg

Note this is MF, in LF V850 is slightly better if no important 3.2D areas

[interneg]

Bob - you're absolutely correct, for whatever reason I remembered that Lambdas & imagesetters are remarkably similar internally in the way they use the optical laser system for exposure, then forgot that they run it in completely different ways! Have been spending too much time variously dealing with offset pre-press & doing battle with Durst flatbed printers & their screening options...

Pere - it's pretty easy to skew a test to 'prove' that your cheap scanner 'outperforms' a professional workhorse or that a drumscanned 8x10 is 'outperformed' by the Phase One de jour. It makes for easy clickbait, & the hard work necessary to dismantle such nonsense takes time, money & a lot of effort. I've handled plenty of images scanned from 120 & 4x5 on Epsons & the sad truth is that they pale next to a competent scan from a Hasselblad, or an iQSmart or a Heidelberg or anything like that. It doesn't matter if the film is B&W, C41 or E6, the difference is visible, especially in prints. I've seen a fair few bad scans off high end kit too, but it's often bad sharpening decisions or poor colour correction (ie operator error) or a poorly maintained scanner rather than fundamental shortcomings with the scanner on a technological level. Rooting around in the weeds of aliasing isn't going to help your case either - yes the Hasselblad will alias grain, but that's largely a problem with 35mm at the highest resolution settings.

[Pere Casals]

Pere - it's pretty easy to skew a test to 'prove' that your cheap scanner 'outperforms' a professional workhorse or that a drumscanned 8x10 is 'outperformed' by the Phase One de jour. It makes for easy clickbait, & the hard work necessary to dismantle such nonsense takes time, money & a lot of effort. I've handled plenty of images scanned from 120 & 4x5 on Epsons & the sad truth is that they pale next to a competent scan from a Hasselblad, or an iQSmart or a Heidelberg or anything like that. It doesn't matter if the film is B&W, C41 or E6, the difference is visible, especially in prints. I've seen a fair few bad scans off high end kit too, but it's often bad sharpening decisions or poor colour correction (ie operator error) or a poorly maintained scanner rather than fundamental shortcomings with the scanner on a technological level. Rooting around in the weeds of aliasing isn't going to help your case either - yes the Hasselblad will alias grain, but that's largely a problem with 35mm at the highest resolution settings.


The test I skewed demonstrates nothing, but it is very consistent with what I found. Just it is the way I say what kind of practical difference I see with MF/LF when comparing X vs V devices.

For 35mm X beats V clearly, but then we can compare the X to a (also cheap) 35mm Plustek 8xxx delivering 3800 effective dpi, not as high as V but I'm not sure there is any practical benefit beyond those 3800.


Let me say my practical view when I made my V850 choice for LF: Under $1000, new with full warranty and official service, Windows 10 drivers. More effective resolving power than X5 for 4x5, makes 8x10, perfect for BW common densities, also perfect for negative color film densities, good even for most Velvia shots with multi-exposure feature, weights, occupies and costs little. And I still retain a budged for drum service for those little shots clearly requiring a drum job some because +3.2D in important subject areas.

It is true that pro scanners pre-cook the image better, with V you always need to adjust better in Ps...

But see the alternatives... and 150Lb old and beaten pre-press device with little service and spares or a drum monster ??

And for what ? to see an slight difference in 2m prints when viewed with nose on it ?


Understand me, if I was all day long (or 3 hours dayly) scanning or selling scanning services I would not consider V850 as the workhorse of my business. At best V850 is semi Pro.

But we can be careful about analyzing practical results.

[bob carnie]

I was making Mural prints 72 inch x 8ft on kodak paper from 8x10 colour film in 1980-83 at Jones and Morris Mural Enlarging in Toronto... I just finished a similar size print job using a 100mp Phase One Camera.
So we are talking about a 30 year gap in time.... For the first time I can say that when I looked at the current print project and walked around these monster prints I felt that the image matched what we could do back then with film. It was a very surreal moment for me as I have been vested in this long campaign of original file to print.. vs film to print and I think I was completely happy that at 65 years of age(2018) I will still have time to work with people that are using these cameras and be able to confidently make prints from these files.
I think the very first revelation to me was the Leica Monochrome (latest) that surprised me with its quality.

Just practical observation here, but I think the horse has left the barn so to speak on this particular issue and I am not sure if the camera makers will continue the constant sensor capability chase they have been on since the early 2000. Its kind of like the Lambda Issue- I am still using 1995 technology to make prints and once the manufactures hit this platform, they concentrated on smaller less expensive workarounds (Chromira) as the bar had been set. We humans can only appreciate certain resolution and then its all a matter of MATH equations.

I think 2018 is the year where people will comprehend that the resolution is at its practical limits, stitching programs, focus shift programs, capability for 3d technology , and of course more economic cameras at the 100mp range are going to be the norm and I do see a 100mp camera under 10k coming soon, its only logical as Spock would say.

I have seen this type of pixel peeping going on here for years when it comes to scanning technology, I have tested most devices and am quite happy with the results I get from my Imocan and Eversmart Supreme, as well my glass carrier enlarger may be old but they deliver great results that I have been able to sell over the years.
Almost in all cases the human at the controls will make the exceptional images, and a great photographer could put a coke bottle as a lens and make magic happen.

[Pere Casals]

...

Bob, I completely agree with you. At the end human vision see no IQ enhacement beyond effective 60 Mpix, this is related to the amount and type of "photosites" our retine/fovea has. A 4k TV has 8 Mpix... and a 100MPix IQ3 may end in at least 60 Mpix effective.

Single reason to need more is for very large prints and wanting to view that with nose on it and also see no flaw.

But most important thing you said is that a great photographer may do all the magics with a coke bottle. To me Sally Mann is a great example.

Only I would like to add that LF is much more than resolving power, as Sally (for example) proved.