question about using the 4x5 Epson scanner lens to scan 5.9" wide and stitch 8x10

I've been reviewing the scanning threads but have not found any posts from anyone who has actually done this--so if you have experience with it, I would appreciate hearing from you.

From what I've found, the actual resolving power of the 8x10 film area guide lens on the Epson 4990/V7xx/V8xx scanners is only around 1,600-1,800, but the dual higher resolution lens used for 4x5 and smaller on the V7xx/V8xx modesl is about 2,300. The higher rez lens only scans 5.9" wide, however.

A few have suggested the higher rez lens could be used to scan two sections of an 8x10 and then the two scans could be stitched together in PS to achieve a 2,300 8x10 scan.

I have also shot quite a few 5x8 and 4x10 images--two of each on one 8x10 sheet of film--and theoretically could scan these with the higher rez lens in one pass if it's possible to rig up a holder to position the film for the higher rez lens. Similarly, I could do a 60% crop of an 8x10 in one pass (5.9/9.8).

Anybody have experience using the higher rez lens with an 8x10 sheet of film? If so, would you please share your process? Is it even possible to "trick" the scanner into scanning part of an 8x10 sheet of film using the higher rez lens? Doug indicates in this post that it is possible simply to select the high rez lens with the software, but that still leaves open how to get the correct calibration: https://www.largeformatphotography.i...=1#post1487198

I have an older 4990 that only has the lower rez lens. I'm wondering if it would be worth it to upgrade to the V700 or later model. I would only want to upgrade if I can use the higher rez lens for scanning 8x10 in sections (assuming stitching).

Thanks!
Michael

I haven't tried to stitch 8 x 10 but I have scanned 5 x 7 in the high res area of the 750 (Wet mounted on the glass carrier) so 5 x 8 will work just fine.

IIRC in VueScan you can select the area that you want to scan by drawing a window around it or moving crop lines so you shouldn't have any problem just scanning part of the negative. If I'm remembering correctly

I've been using an IQsmart 2 for a few years and I've sort of forgotten some of the details of using VueScan but it's a nice package IMHO.

Hi Jim,
Thanks for your response. Is there a glass carrier for 5x7? That's news to me. Is this an Epson carrier? BetterScanning? Some other brand?

Ahh...I see that BetterScanning has a 5x7 available.

Still looking for 8x10 though....

The carrier is what Epson calls a fluid mounting station. It came with the 750 but not the 700. It sits on top of a plastic base with a grid on it and you use the grid to position the neg and then lift the carrier off the base and put it on the scanner bed. Munting on the carrier is pretty easy - I timed it and I could get a good wet mount of 4 x 5 or 5 x 7 in 23 seconds. 8 x 10 would be a bit trickier as you'd have to align it so part of the width hung over the edge of the carrier on the right with the left edge aligned with the marks for the left side of the 5.9" area, then scan and remount it with the other edge hanging out and the right side aligned with the marks for the right side (or vice versa of course) When I was doing 8 x 10 I just layed it flat on the scanner glass because even at 1800 effective dpi you could still make prints of pretty large size. I actually made some canvas prints that were about 40" by 56" and mounted them as a canvas wrap - came out pretty nicely.

Adding to what Jim pointed...

Quote:

---

Originally Posted by Michael Roberts

the Epson 4990/V7xx/V8xx scanners is only around 1,600-1,800, but the dual higher resolution lens used for 4x5 and smaller is about 2,300.

---

with the LR lens it performs 1700 for the vertical axis and 2300 for the horizontal axis. With the HR lens it is 2300 for the Y and 2900 for X (2900! amazing for a 5.9" scanning width).

Low res lens is focused o the bed glass itself, the hr lens if focused some mm higher:

Attachment 194276
("sharpness" vs height)

Quote:

---

Originally Posted by Michael Roberts

Is it even possible to "trick" the scanner into scanning part of an 8x10 sheet of film using the higher rez lens?

---

Some made 5x7" DIY holders, for 4x10" a possibility I'd consider is taking an spare 4x5 holder, sawing the plastic separation between the two frames and then adding a metallic frame (made with two overlaping frames) that would take the 410 and it would fit in the 5" wide hole.

A 8x10 holder would ber a bit more coplex as it should shift to take the crops... it can be done I guess.

...but for 810 I don't see much the need to improve the scanning with the high res lens, the LR lens delivers insane 300MPix effective!

This shows how it is in a 6m print: https://www.flickr.com/photos/125592...5/32535835184/

To me the area guide allows perfect 2m prints, you always can wet mount on bed glass, and many drum scanning services even don't offer scanning beyond 2000 dpi for 8x10. A way to get all possible quality for a monster print is an optical enlargement, scanning all that quality is a challenge, a way it would be building a very sound dslr scanner and stitching around 50 shots.

Quote:

---

Originally Posted by Michael Roberts

I'm wondering if it would be worth it to upgrade to the V700 or later model.

---

Upgrade to V800, sell the 4990 for some $150. The V800 has LED illumination so there is no heating time delay for pre-scanning.

Quote:

---

Originally Posted by Jim Andrada

The carrier is what Epson calls a fluid mounting station. It came with the 750 but not the 700. It sits on top of a plastic base with a grid on it and you use the grid to position the neg and then lift the carrier off the base and put it on the scanner bed. Munting on the carrier is pretty easy - I timed it and I could get a good wet mount of 4 x 5 or 5 x 7 in 23 seconds. 8 x 10 would be a bit trickier as you'd have to align it so part of the width hung over the edge of the carrier on the right with the left edge aligned with the marks for the left side of the 5.9" area, then scan and remount it with the other edge hanging out and the right side aligned with the marks for the right side (or vice versa of course) When I was doing 8 x 10 I just layed it flat on the scanner glass because even at 1800 effective dpi you could still make prints of pretty large size. I actually made some canvas prints that were about 40" by 56" and mounted them as a canvas wrap - came out pretty nicely.

---

Thanks, Jim. I appreciate the information. I've done wet mounting on a Howtek 4500, but not on an Epson using the Epson station. For context, I'm printing up to 71x120" from drum scans. I don't expect to print that big from an Epson scan, my thought is simply that if I am going to go to the trouble of scanning at home with an Epson, then why not wring the best resolution I can out of an Epson. Mine is an extreme position, I understand that.

Pere,
Thanks for your reply, it's very helpful.

A question: with my 4990, the 8x10 film area guide has the long dimension consistent with the long dimension of the scanner, but my 4x5 Epson holder has the 5" dimension at 90 degrees. If I am understanding your x/y axis info correctly, this means the 10" length of the 8x10 sheet is scanned effectively at 2300 and for 4x5, the 5" length is scanned effectively (with the higher resolution lens) at 2300 also. Is this correct?

Quote:

---

Originally Posted by Michael Roberts

this means the 10" length of the 8x10 sheet is scanned effectively at 2300 and for 4x5, the 5" length is scanned effectively (with the higher resolution lens) at 2300 also. Is this correct?

---

The counter...

The 10" of the 8x10 has effective 1700dpi, and the 8" is 2200dpi.

For 4x5" (sheet rotated 90º) the 4" is 2300dpi effective, and the 5" is 2900dpi.

While the Epson has a sound optics (it resolves 18400 pixels, [8" x 2300dpi] in the scanning width) the drive train looks not as good and in the motion direction we have a loss.

Quote:

---

Originally Posted by Pere Casals

...but for 810 I don't see much the need to improve the scanning with the high res lens, the LR lens delivers insane 300MPix effective!

This shows how it is in a 6m print: https://www.flickr.com/photos/125592...5/32535835184/

To me the area guide allows perfect 2m prints, you always can wet mount on bed glass, and many drum scanning services even don't offer scanning beyond 2000 dpi for 8x10. A way to get all possible quality for a monster print is an optical enlargement, scanning all that quality is a challenge, a way it would be building a very sound dslr scanner and stitching around 50 shots.

---

Pere,
This bell seems extraordinarily sharp for a 6m print! I guess my question is: at what viewing distance does this represent?

For example, I am using an Excel spreadsheet developed by another forum member that calculates a sharp viewing distance of 1.1' (13in) for a 120"L print from an 8x10 when scanned at 3200 and a viewing distance of 2.2' (26in) for a 120"L print from an 8x10 scanned at 1600.

So, for your hypothetical 6m print with the enlarged bell, what is the assumed viewing distance? Just curious, I have no problems agreeing that this would look sharp at a normal viewing distance for a 6m print (>3' for certain).

Unlike most folks (it seems--I agree it is highly unusual to request a drum scan of an 8x10 at >2000--I've spent the last two years trying to do this), I am very interested in how close I can walk up to a mural sized print and it look sharp. I can get the sharp viewing distance down to about 6" with an 8x10 drum scan at 3200--but only for a print that's 60" wide; for 120" wide I'm stuck with a 13" viewing distance at best. [don't worry, I know this is crazy]

Quote:

---

Originally Posted by Michael Roberts

Howtek 4500, I'm printing up to 71x120" from drum scans. I don't expect to print that big from an Epson scan

---

hmmmm.... this depends. Most of the times the Epson is not the limiting factor in the LF image quality. While there is no doubt that the howtek is a way superior system this may not have an effect with many LF shots. The Epson resolves 65 Lp/mm in the transversal direction and 45 lp/mm in the motion direction, at those cycles/mm the LF lens-film combination has also contrast extintion for usual shooting conditions.


Years ago I made a test, I downloaded crops from the Collaborative Large Format Scanner Comparison (https://www.largeformatphotography.i...an-comparison/) and I just edited the crops in PS to see the real effect after each crop was optimized.

Pro scanners make a very good digital image optimization, the Epsons are not Pro and you have to do it later, but with a few clicks an LF scan from an Epson is amazingly close to the Howtek:

https://www.flickr.com/photos/125592...5/28420386682/

There is no secret, it's about downloading the crops from that "side by side" and sharpening to optimum, and adjusting curves to match.

If it was a CMS20 film shot, with the lens focused at infinite for a very distant subject (no dof required), with a very sharp lens at optimal aperture, with a contrasty texture, and with no wind... then probably we would see some difference if the drum scans at 4000 dpi or beyond.

Of course for 35mm roll film it's easy to see a difference, if it is a very steady shot.

For two-up 4 x 10. why not just cut the 8 x 10 in two?

Quote:

---

Originally Posted by Michael Roberts

This bell seems extraordinarily sharp for a 6m print! I guess my question is: at what viewing distance does this represent?

---

To me a 6m print from that 8x10" negative scanned with the Epson is perfect at 2m distance, so a 3m print would be perfect at 1m, and a 1.5m print would be good at reading distance.

This would be for a demanding photographer, IMHO a regular viewer would not complain if distance was shorter.

Quote:

---

Originally Posted by Michael Roberts

Unlike most folks (it seems--I agree it is highly unusual to request a drum scan of an 8x10 at >2000--I've spent the last two years trying to do this), I am very interested in how close I can walk up to a mural sized print and it look sharp. I can get the sharp viewing distance down to about 6" with an 8x10 drum scan at 3200--but only for a print that's 60" wide; for 120" wide I'm stuck with a 13" viewing distance at best. [don't worry, I know this is crazy]

---

I've been also digging in that...

IMHO it's about digital post processing, this is very important.

But there is another answer: Sally Mann. She had been departeing from silver nitrate, br salt, glass sheets, pottery... to end in the most impressive big prints many have ever seen on a wall. Plenty of autenticity also.


We have to return back to the darkroom... this is the key. Making sound negatives that print amazingly well, and then projecting the negative on silver paper.

I made tests I found that even an old enlarging lens introduces no loss in the image quality, I found that my old, cheap and scratched "no letter" Rodagon takes insane 145 lp/mm from the negative and it crafts every bit on the paper:

Attachment 194305

Attachment 194306

Back to the darkroom !!!


It's cheaper, it's not reprography, it's authentic.

But one also has to obtain better tonal results than with hybrid. This looks a challenge, but's about technical preparation.

Quote:

---

Originally Posted by Pere Casals

Pro scanners make a very good digital image optimization, the Epsons are not Pro and you have to do it later, but with a few clicks an LF scan from an Epson is amazingly close to the Howtek:
There is no secret, it's about downloading the crops from that "side by side" and sharpening to optimum, and adjusting curves to match.

---

I really, really want to believe this--it would make my life so much easier! But I also understand the seductive power of wishful thinking and confirmation bias, so I really, really want to do this kind of comparison test myself--hence this inquiry into what/how to wring the most resolution out of an Epson....

Quote:

---

Originally Posted by Pere Casals

To me a 6m print from that 8x10" negative scanned with the Epson is perfect at 2m distance, so a 3m print would be perfect at 1m, and a 1.5m print would be good at reading distance.

This would be for a demanding photographer, IMHO a regular viewer would not complain if distance was shorter.

---

Agreed!!!

Quote:

---

Originally Posted by Pere Casals

Back to the darkroom !!!

---

I really, really do not want to believe this!

Oh, Reality--such a difficult mistress.

Quote:

---

Originally Posted by Jim Andrada

For two-up 4 x 10. why not just cut the 8 x 10 in two?

---

Yes, Jim, I agree--cutting the film after it's processed would not be a problem (same w/ two-up 5x8s). But...is there a 4x10 holder available? If not, how would this help? Sorry to be slow....

Quote:

---

Originally Posted by Michael Roberts

Oh, Reality--such a difficult mistress.

---

The lightjet/lambda/inkjet way is also demanding.

We need 12 to 20 sharp pixels per mm.

> We have to work 16bits per channel with an oversampled image (what is oversampled from a 8x10 scan? :) )

> We have to perform initial sharpening

> Edition

> After edition sharpening

> Image reduction (bicubic for reductions) to the printer matrix size

> A Pixel level sharpening

> A viewing distance intended additional sharpening

> The right raster, the right ink set, piezography, etc.

True dat!

Quote:

---

Originally Posted by Michael Roberts

I really, really want to believe this--it would make my life so much easier! But I also understand the seductive power of wishful thinking and confirmation bias, so I really, really want to do this kind of comparison test myself--hence this inquiry into what/how to wring the most resolution out of an Epson....

---

Whether people like it or not, it's safest to regard the best possible output from an Epson to be 1200ppi at not spectacular MTF performance. No amount of silly claims of magical sharpening routines can pretend that the fundamental optical performance is poor, or that manipulating low resolution web images somehow 'proves' it isn't. The A3 Epsons give you a fighting chance at getting better performance thanks to an actual focusing system.

Quote:

---

Originally Posted by interneg

Whether people like it or not, it's safest to regard the best possible output from an Epson to be 1200ppi at not spectacular MTF performance.

---

interneg, your statement is NOT ridiculous

But it also happens with the other scanners:

https://www.largeformatphotography.i...=1#post1509776
https://www.largeformatphotography.i...=1#post1479178


It is interesting to compare LF performance of the V700 vs your X1

Your hasselblad X1 is pure gold (+6000dpi effective) for 35mm, but this is a LF forum, IIRC, and in that context the V700 outresolves the X1.

in the hor axis:

> For 4x5" the V700 delivers 2900dpi effective, your X1 delivers 1800.

> For 5x7" the V700 delivers 2900dpi effective, your X1 delivers 1440 (the half).

> For 8x10" the V700 delivers 2300dpi effective, your X1 delivers 0.

(for the vert V:2300-2300-1700 vs X:1620-1300-0) (the x1 has a 10% resolving power loss in the Vert vs Hor, 20% loss in the Epson case)


If for 4x5" your X1 extinguishes contrast at 1800dpi... do you think that you will have "spectacular MTF" at 1200 dpi ? You have worse MTF than the Epson at 1200dpi !


Also you should consider that anyway the LF lenses-films extinguish contrast at similar (or lower) cycles/mm than the Epson, so a higher resolving power has a minor effect.

Quote:

---

Originally Posted by interneg

The A3 Epsons give you a fighting chance at getting better performance thanks to an actual focusing system.

---

With the V700 if you scan 8x10" on glass bed you have perfect focus, on any doubt you may even wet mount on bed.

Contrast Adaptive Sharpening (CAS)

Adaptive Sharpening with Overshoot Control.



Additionally I would like to remark that the digital optimization work many Pro scanners do is completely evident.

If we see the Howtek crop from the Collaborative Scanner Comparison it is clear that the edges have a deep digital work. If one often takes a microscope to inspect film those edges are from another planet, they are not natural to the lens-film work. Those scanned/enlarged edges are impossible with that low contrast.


If we inspect film with the microscope what we see (at that enlargement) is Epson look.

I theory "internal sharpening" is intended to correct the system pitfalls, not the taken image, as the taken image should be corrected in another level, but in certain conditions it would be difficult for the internal sharpening to guess if blur comes from the taking lens or from the scanning system.


What it's absolutely clear to me (from next crop) is that the howtek did a deep digital optimization work in those (another planet) edges that had to be blurred in the negative, for this reason the (rawer) Epson image can be matched to the Howtek one with Ps:


https://www.largeformatphotography.i...ullframe-u.jpg

https://live.staticflickr.com/8561/2...62c476d1_b.jpg


IMHO the reason why the Epson matches a fancy Howtek is simply because that LF negative has not much more than the Epson can take, and if the Howtek showed those alien edges it is because the edges were blurred on film and the internal digital sharpening did a job, while in the Epson case that job has to be done in Ps.

Quote:

---

Originally Posted by Pere Casals

To me the area guide allows perfect 2m prints, you always can wet mount on bed glass

Upgrade to V800, sell the 4990 for some $150. The V800 has LED illumination so there is no heating time delay for pre-scanning.

---

Pere,
A couple of follow up questions for you:

1. Others (e.g., Sandy King, Ben Horne) have commented the sharpest focus for the lens with Film Area Guide (i.e., the 8x10 lens) is 1mm above the glass. Have you tested this? In your experience is performance "good enough" with dry or wet mounting directly on the lower scanner glass? I am thinking about buying the Aztek kit and just mounting 8x10 directly on the lower scanner glass....but if I need to raise the film 1mm, then I need to figure out how to do that also.

2. Is there any other advantage to the 800 over my current 4990 if I'm just going to fluid mount directly on the scanner glass? My understanding is the resolution for the 8x10 lens (with Film Area Guide) is the same for the 4990 as for the 7xx/8xx series of scanners. Seems like I may as well keep the scanner I have than trade up unless there is some other advantage--unless the cooler leds make a big difference.

Thanks!
Michael

Those test scans were done many many years ago by operators who had no idea how to properly use the software. That was clearly evident in the comparison between Aztek's version of the same HR8000 Howtek where the maximum resolution was wildly different (lower) with the identical Howtek. Well, it turns out that Colorbyte designed the "auto" aperture feature of the software to NOT choose the 3 micron aperture when selecting 8000 ppi. That was because so often, that would result in grainier, noisier scans when the film itself didn't support that res. I'm the person who discovered that and that you have to manually set the aperture if you want that minimum aperture. All that is to illustrate the the scanner operator makes a huge difference and if you don't understand the software, you get skewed results. Since the default behavior of Trident IS to provide a certain amount of sharpening, that is most likely what you're looking at in the Howtek scans. The Trident sharpening module is very interesting when you get into it and it was genius for prepress purposes but has no place in fine art RGB scans. There is absolutely no hardware sharpening going on in the Howtek drum scanners. I've been through that for the last twenty years when I spent more of my time conversing with ColorByte about how their software worked. I'm going to go out on a very short limb and say that those scanner tests are flawed and unless I could see screenshots of all the windows in Trident, I'll stick by that opinion. Another thing that makes a big difference on these scanners is whether the optical system on the scanner has been manually custom aligned. The early off the shelf Howteks were more likely to not be perfectly aligned compared to the same machine sold by Aztek where that was one of the pre-sale operations done before delivery. Having compared the hi-res scans from my own HR8000 with the optical alignment done, those scans are better than any other scanner I've seen except perhaps the ICG's, which, of course, have no support left.

Quote:

---

Originally Posted by Michael Roberts

1. Others (e.g., Sandy King, Ben Horne) have commented the sharpest focus for the lens with Film Area Guide (i.e., the 8x10 lens) is 1mm above the glass. Have you tested this? In your experience is performance "good enough" with dry or wet mounting directly on the lower scanner glass? I am thinking about buying the Aztek kit and just mounting 8x10 directly on the lower scanner glass....but if I need to raise the film 1mm, then I need to figure out how to do that also.

---

That bell was scanned directly on glass https://www.flickr.com/photos/125592...5/32535835184/ with emulsion down to avoid NRs

The Low Res lens is also used to scan opaques, so it has no sense to focus 1mm higher as the opaque is pressed aganist glass. I tested performance with the glass slide directly on glass to check if gear was ok but I've not tested 1mm. Nor Sandy King or Ben Horne would be jocking about that, so it would be worth to investigate it.

The LR lens should have some greater DOF than the HR lens, probably there are assembly tolerances and perhaps focus is placed in a way that it ensures a good performance, and perhaps there is some variability in different units. I'll check it next time I scan.

Without a USAF 1951 glass slide, there is a very easy way to check what is perfect focus distance:

https://jrileystewart.com/blog/2017/...atbed-scanner/
https://web.archive.org/web/20190701...atbed-scanner/

Quote:

---

Originally Posted by Michael Roberts

2. Is there any other advantage to the 800 over my current 4990 if I'm just going to fluid mount directly on the scanner glass? My understanding is the resolution for the 8x10 lens (with Film Area Guide) is the same for the 4990 as for the 7xx/8xx series of scanners. Seems like I may as well keep the scanner I have than trade up unless there is some other advantage--unless the cooler leds make a big difference.

---

I've never tested the 4990, tests around indicate that the 4990 performs like the V850 with the Low Res lens, so my guess is that you won't notice much improvement for 8x10, but some report an slight improvement.

https://www.flickr.com/photos/mikeco...7594426285173/

Here they find that the drum at 3300 is slightly better than the 4990 at nomial 4800
https://web.archive.org/web/20171203...tbed-scanners/

And what I find is that by editing that crop in Ps I equal the drum result, in two or three clicks :)

Attachment 194338


Anyway the V800 has no heating delay, this is very convenient!

Quote:

---

Originally Posted by Sasquatchian

There is absolutely no hardware sharpening going on in the Howtek drum scanners. I've been through that for the last twenty years when I spent more of my time conversing with ColorByte about how their software worked

---

This does not happen in software, but in firmware.

Pere, thank you so much!

Hi Michael. I thought I had responded to your question of how to mount 4 x 10 on the Epson but it doesn't show u here, so here goes again

Just wet mount the 4 x 10 on the carrier glass (the glass carrier that's part of what they call the mounting station)

This thread seems to have gone off the rails a bit into the tech weeds - I think you were asking a pretty simple straightforward question.

I solved the question by getting a bigger scanner with focusing ability. I haven't used a larger Epson but...

Pere - How would you know, looking at those scans, whether something happened in software or firmware? Do you own a Howtek? Have you ever operated one? I have never seen anything that looked like anything other than film grain from the high res scans, either 4000 or 8000 ppi from either of the Howtek's I've owned. Now it is true that certain Hell Chromagraph scanners did do what you claim and even Imacon Flextight's sharpened the image somewhat when the USM was set to zero, but what evidence do you have that Howtek did anything like this in firmware? I'm not buying what you're saying without definitive proof.

Quote:

---

Originally Posted by Sasquatchian

Now it is true that certain Hell Chromagraph scanners did do what you claim and even Imacon Flextight's sharpened the image somewhat when the USM was set to zero, but what evidence do you have that Howtek did anything like this in firmware? I'm not buying what you're saying without definitive proof.

---

And it is important to note that most of these USM's can be bypassed or turned off - because it's usually just in the scanner software. The infamous -120 USM setting for the Imacon in TIFF mode switches off the USM - confusingly, in the 3F mode, 0 in the USM apparently equates to the same thing... I long ago ceased wondering about the eccentricity of high end scanner software.

Firmware sharpening suggests it has to happen inside the scanner before it even reaches the computer - and these sort of claims need exceptional evidence - which you won't find forthcoming because Pere has never operated or owned a Howtek/ Aztek, let alone looked inside one, or gone through the software as comprehensively as you have. He seems to believe he's one sharpening routine away from turning a V700 into a high end scanner...

Quote:

---

Originally Posted by interneg

He seems to believe he's one sharpening routine away from turning a V700 into a high end scanner...

---

:) LOL

Personal attack is your single resource.

V700 is not Pro gear, but it resolves 2900" x 5.9" = 17110 effective pixels in the scan width, no crappy lenses and folded optics inside. Your X1 resolves 1/3 of that.

While the X1 is stellar for 35mm rolls it's beaten by the cheapo V700 in LF, sorry, but your personal attacks won't change that. The X1 has a painful shorcomming: a sensor with only 8k, and no way to displace it to take crops that can be stitched.

Quote:

---

Originally Posted by Sasquatchian

but what evidence do you have that Howtek did anything like this in firmware?

---

With all user level sharpening disabled, we have found shots that are sharper in the howtek scan than in the microscope inspection of the film, me... I've no doubt, but it can be checked by producing a set of CMS20 negatives portraying edges at different contrasts, and with different CoC, and cheching when edges are improved.

I fact it would be stupid not implementing a certain degree of sharpening in firmware because this allows a huge speed vs quality benefit in the machine to computer communications. This allows an oversampling_the_medium+sharpening+ downsampling_to_required_size cycle that improves quality a lot and it does not not overload the computer and the communications.


Look, "in firmware" sharpening has huge technical benefits, and this is one of the strong points of very Pro hardware.


In the same way one has to understand that the not Pro V700 scans have to be optimized in Ps before they are compared to result of a Pro scanner, because Epson does not that process with that efficiency.

Quote:

---

Originally Posted by Michael Roberts

I've been reviewing the scanning threads but have not found any posts from anyone who has actually done this--so if you have experience with it, I would appreciate hearing from you.

From what I've found, the actual resolving power of the 8x10 film area guide lens on the Epson 4990/V7xx/V8xx scanners is only around 1,600-1,800, but the dual higher resolution lens used for 4x5 and smaller is about 2,300. The higher rez lens only scans 5.9" wide, however.

A few have suggested the higher rez lens could be used to scan two sections of an 8x10 and then the two scans could be stitched together in PS to achieve a 2,300 8x10 scan.

I have also shot quite a few 5x8 and 4x10 images--two of each on one 8x10 sheet of film--and theoretically could scan these with the higher rez lens in one pass if it's possible to rig up a holder to position the film for the higher rez lens. Similarly, I could do a 60% crop of an 8x10 in one pass (5.9/9.8).

Anybody have experience using the higher rez lens with an 8x10 sheet of film? If so, would you please share your process? Is it even possible to "trick" the scanner into scanning part of an 8x10 sheet of film using the higher rez lens? Doug indicates in this post that it is possible simply to select the high rez lens with the software, but that still leaves open how to get the correct calibration: https://www.largeformatphotography.i...=1#post1487198

I have an older 4990 that only has the lower rez lens. I'm wondering if it would be worth it to upgrade to the V700 or later model. I would only want to upgrade if I can use the higher rez lens for scanning 8x10 in sections (assuming stitching).

Thanks!
Michael

---

I have a 4990 I found the optimum focus is slightly above the glass. Mine was about 1.5m, I think, FWITW... With a wide lens, the depth of field is not as critical as the longer lens that the v700/800 have.

They had to achieve some compromise in any case because the film holders can not be at 0mm. The V700/800 have two lenses so you have two focus plans for different tasks to choose from. But you still have the same problems just more choices to help solve it, just not all choices to solve all the problems...

Quote:

---

Originally Posted by Sasquatchian

whether something happened in software or firmware?

---

Every development team is a different world. What is clear is that in those designs Image Quality would be a concern of different development departments and even of different companies, for example SilverFast is a 3rd party software interfacing scanners.

From this perspective it is very important that the machine itself makes the sharpening that can be optimally done at that firmware level and it would be difficult to perform optimally in the following processing layers (software development team).

For example the firmware in the Hasselblad X1/X5 can make a different optimal sharpening depending on the lens magnification, sure the Linos Rodagon inside has a different CoC depending on magnification, and it requires a USM mask with higher radius/factors in the Y direction than in the X direction.

Also a drum has motion blur that depends on the physics, so it's the machine itself that knows how to do it optimally. Delegating that job to next layers is a mess, not mentioning the chance we want 3rd party software interface the scanner.

A Pro machine not having a sound internal optimization is disappointing.

IMHO, the idea is that the firmware should do a sharpening that is related to the physics of the machine, and the user layer sharpening should do anything else.

Quote:

---

Originally Posted by Ted Baker

I have a 4990 I found the optimum focus is slightly above the glass. Mine was about 1.5m, I think, FWITW... With a wide lens, the depth of field is not as critical as the longer lens that the v700/800 have. They had to achieve some compromise in any case because the film holders can not be at 0mm.

---

Yes...

Quote:

---

Originally Posted by Jim Andrada

Hi Michael. I thought I had responded to your question of how to mount 4 x 10 on the Epson but it doesn't show u here, so here goes again

Just wet mount the 4 x 10 on the carrier glass (the glass carrier that's part of what they call the mounting station)

This thread seems to have gone off the rails a bit into the tech weeds - I think you were asking a pretty simple straightforward question.

I solved the question by getting a bigger scanner with focusing ability. I haven't used a larger Epson but...

---

Thanks Jim! What I am seeing is the carrier glass is 7.25 x 9.75in, so the film will need to fit inside this area. Seems like 4x10 could be trimmed slightly to fit w/o losing any of the image; could be an issue with fluid leakage, though.

Quote:

---

Originally Posted by Ted Baker

I have a 4990 I found the optimum focus is slightly above the glass. Mine was about 1.5m, I think, FWITW... With a wide lens, the depth of field is not as critical as the longer lens that the v700/800 have.

They had to achieve some compromise in any case because the film holders can not be at 0mm. The V700/800 have two lenses so you have two focus plans for different tasks to choose from. But you still have the same problems just more choices to help solve it, just not all choices to solve all the problems...

---

Thanks, Ted! Good to have more data on the optimal plane of focus for scanning directly on the lower glass.