Scanner Comparison 2019 [Epson Flatbed | Eversmart Flatbed | Drum Scanners]

I am going to go through a scanner comparison exercise to provide an updated resource on the capabilities of various scanners as of 2019. I will use this thread to document the results and encourage others to chime in and help guide the process but please refrain from taking things off-topic. I am also keeping my fingers are crossed that we will remain kind to one another.

Scanners and computer systems that I will be using:

• Epson V700 | Windows 7 64bit | Vuescan & Silverfast SE
• Creo/Scitex Eversmart Pro | Mac OS9 | G5 Mirror | Eversmart Scan
• Creo/Scitex Eversmart Supreme | Mac OS10.6.8 | G5 Mirror | OxygenScan 2.6.4
• Scanmate 5000 | Windows 7 64bit | Color Quartet 5
• Scanmate 11000 | Windows 7 64bit | Color Quartet 5
• Heidelberg Tango | Mac OS10.2.8 | G5 Mirror | Newcolor 2000

The scanning tests will include scans for the following:

• Wolf Faust IT8 Color Target
• USAF 1951 Resolution Target
• Black and White Negatives [8x10 | 4x5 | 6x7 MF | 35MM]
• Color Negatives [4x5 | 6x7 MF | 35MM]
• Color Slides [8x10 | 4x5 | 6x7 MF | 35MM]

Since each scanner/software setup is different my approach will be to follow these requirements in order:

• 5000 DPI Resolution [Espon V700 will be scanned at 6400 DPI and down-sampled to 5000 DPI in Photoshop
• RAW Scans [Note: RAW is not possible for Eversmart Pro and Tango but I will turn off all software settings to get as close to RAW as possible
• Positive conversions for negatives using ColorPerfect Plugin only and no additional Photoshop edits
• Photos will be posted in sRGB using convert to profile method in Photoshop

A little bit information on the USAF Resolution Target that I am using:
See this link for a little history about the USAF 1951 resolution target and how it can be used to test for resolution. The scans that I'll post will be able to give a very good estimate of actual resolution using the reference table below:

https://www.netsoft2k.com/Docs/Media...ion Target.jpg

Reserve 1

Reserve 2

Reserve 3

Reserve 4

Reserve 5

Reserve 6

Reserve 7

Reserve 8

Reserve 9

Reserve 10

Epson 700 | SilverFast SE | IT8 Colortarget RAW

LINK TO FILE

https://www.netsoft2k.com/Docs/Media...6400DPIWeb.jpg

https://www.netsoft2k.com/Docs/Media...6400DPIWeb.jpg

Epson 700 | SilverFast SE | USAF 1951 Resolution Target RAW

LINK TO FILE

Image below is cropped from a 6400 DPI scan but not resized. This is the best resolution my V700 can achieve. Vertical is ~2800 DPI and Horizontal is ~2200 DPI

https://www.netsoft2k.com/Docs/Media...6400DPIWeb.jpg

Below are links to my "focus calibration" scans

Focus Setting 1
Focus Setting 2 **Best Focus**
Focus Setting 3
Focus Setting 5

Pali, great thread and you are the right person to do it.

I ask that you quantify your target charts. Such as 1951 USAF resolution test chart and how to use it. I use a letter sized download Laser printed and know I am doing it wrong. Several times I have asked for help with that on this forum.

I notice that https://www.edmundoptics.com/c/test-targets/1099/ sells many, some are very expensive and tiny.

I plan to buy this one as it's big and 'only' $50 shipped. https://www.edmundoptics.com/f/resol...r-chart/11903/

Also a great idea to 'reserve' linear posts so it gets recorded in one string. Genius!

Quote:

---

Originally Posted by Pali K

Epson 700 | SilverFast SE | USAF 1951 Resolution Target RAW
Image below is cropped from a 6400 DPI scan but not resized. This is the best resolution my V700 can achieve. Vertical is ~2800 DPI and Horizontal is ~2200 DPI

---

Pali, yes, this is consistent with all tests posted around, and my own tests. Thanks for making this test.

Think that 2800 effective dpi is amazing for a single pass covering 5.9", any expensive pre-press or pro scanner has to stitch strips to beat that. Sadly the poor vertical driving quality degradates the Y performance.

From scans of practical shots it would be interesting to find at what enlargement a drum makes a difference. My bet is 1.5m for 8x10 and 1m for 4x5.

Still for 8x10, with area guide it has a lower resolving power because the other lens takes the full bed, instead 5.9", it would be also interesting to measure that.

Quote:

---

Originally Posted by Randy Moe

Pali, great thread and you are the right person to do it.

I ask that you quantify your target charts. Such as 1951 USAF resolution test chart and how to use it. I use a letter sized download Laser printed and know I am doing it wrong. Several times I have asked for help with that on this forum.

I notice that https://www.edmundoptics.com/c/test-targets/1099/ sells many, some are very expensive and tiny.

I plan to buy this one as it's big and 'only' $50 shipped. https://www.edmundoptics.com/f/resol...r-chart/11903/

Also a great idea to 'reserve' linear posts so it gets recorded in one string. Genius!

---

Thank you Randy! Good thought on qualifying the resolution target which I will do so now.

Quote:

---

Originally Posted by Pere Casals

Pali, yes, this is consistent with all tests posted around, and my own tests. Thanks for making this test.

Think that 2800 effective dpi is amazing for a single pass covering 5.9", any expensive pre-press or pro scanner has to stitch strips to beat that. Sadly the poor vertical driving quality degradates the Y performance.

From scans of practical shots it would be interesting to find at what enlargement a drum makes a difference. My bet is 1.5m for 8x10 and 1m for 4x5.

Still for 8x10, with area guide it has a lower resolving power because the other lens takes the full bed, instead 5.9", it would be also interesting to measure that.

---

Thanks Pere! Yes, I plan to post everything here step by step and will consolidate everything in the first 10 posts when all said and done.

Pali,

Have you considered doing a slanted edge test? You can a razor blade at 5degrees for the flat beds, but you may need to create a suitable target for the drums. For the color negative what are you planning on using?

Good going Pali.

Great job, Pali. I look forward to more of your results!

Quote:

---

Originally Posted by Pali K

Vertical is ~2800 DPI and Horizontal is ~2200 DPI

---

Pali, just I checked the calculations, in the vertitical bars we can see element 5.6 clearly. This is 57.0*25.4*2 = 2895.6, so it's way better rounded to 2900dpi.

With Hor bars (vert axis) we see 5.4 at least, this is nailing 2300 dpi. Always there is a subjective factor for the 1951, and important thing is having a uniform criteron, but to me a fair reading would be 2900x2300, in special the 2900 reading is clear to me.

Quote:

---

Originally Posted by Ted Baker

Pali,

Have you considered doing a slanted edge test? You can a razor blade at 5degrees for the flat beds, but you may need to create a suitable target for the drums. For the color negative what are you planning on using?

---

Ted, I had not and have never heard of this test but I can definitely look into it. I am curious to know what would this test help with? I am using this image for Color Negative testing because the leaves are subtle variations of colors that will be interesting to see across the scanners.

http://www.netsoft2k.com/Docs/Media/...gTestImage.jpg

http://www.netsoft2k.com/Docs/Media/...tImageCrop.jpg

Quote:

---

Originally Posted by Corran

Good going Pali.

---

Quote:

---

Originally Posted by Peter De Smidt

Great job, Pali. I look forward to more of your results!

---

Thank you Bryan and Peter!

Quote:

---

Originally Posted by Pere Casals

Pali, just I checked the calculations, in the vertitical bars we can see element 5.6 clearly. This is 57.0*25.4*2 = 2895.6, so it's way better rounded to 2900dpi.

With Hor bars (vert axis) we see 5.4 at least, this is nailing 2300 dpi. Always there is a subjective factor for the 1951, and important thing is having a uniform criteron, but to me a fair reading would be 2900x2300, in special the 2900 reading is clear to me.

---

Pere, I think my astigmatism may be coming into play here but I went with safer option to what my eye can see. I think that is why you and I are coming at slightly different conclusion. I agree with your assessment though and that is why I noted the resolution as approximate. I will update the numbers once we are done with all the tests and what everyone feels like is the right summary for each scanner.

Quote:

---

Originally Posted by Pali K

I think that is why you and I are coming at slightly different conclusion.

---

The 1951 had been criticized because that silght subjectiveness about what's contrast extintion. IMHO a good criterion is being able to be perceive for sure the bars' orientation, without considering the extremes of the bars.


[QUOTE=Pali K;1478033]I am curious to know what would this test[/IMG]

http://kilopixel.net/publications/EI5294-11.pdf

Epson 700 | SilverFast SE | Kodak Portra 160 Color Negative 16BIT RAW Scanned at 6400 DPI and down-sampled to 5000 DPI

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...6400DPIWeb.jpg

The best ColorPerfect conversion I could achieve without increasing saturation. LINK TO 5000 DPI FILE

http://www.netsoft2k.com/Docs/Media/...n6400DPICP.jpg

http://www.netsoft2k.com/Docs/Media/...00DPICPWeb.jpg

Creo/Scitex Eversmart PRO | Eversmart Scan 3.0 | IT8 Colortarget 8 BIT RGB - All Settings OFF

DISCLAIMER:

• Eversmart PRO does not have the ability to scan in RAW or save files in 16 BIT.
• 5000 DPI is interpolated by the software.

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...5000DPIWeb.jpg

Creo/Scitex Eversmart PRO | Eversmart Scan 3.0 | USAF 1951 Resolution Target 8 BIT RGB - All Settings OFF

DISCLAIMER:

• Eversmart PRO does not have the ability to scan in RAW or save files in 16 BIT.
• 5000 DPI is interpolated by the software.

This is the best resolution my Eversmart PRO can achieve. Both Vertical and Horizontal are either Group 5 - Element 5 or 6 (~2500 - 2800 DPI)

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...5000DPIWeb.jpg

UPDATE: I zoomed in the image and it seemed the resolution is greater than what is visible on the original image due to file size so I scanned it again at 8000 DPI max allowed and upscaled to 11000 DPI. Here is the result. I think the resolution is safely Group 5 Element 6 putting it at ~2896 for both Vertical and Horizontal.

http://www.netsoft2k.com/Docs/Media/...1000DPIWeb.jpg

Creo/Scitex Eversmart PRO | Eversmart Scan 3.0 | Kodak Portra 160 Color Negative 8 BIT RGB - All Settings OFF

DISCLAIMER:

• Eversmart PRO does not have the ability to scan in RAW or save files in 16 BIT.
• 5000 DPI is interpolated by the software.
• I converted the 8 BIT file to 16 BIT in Photoshop before using ColorPerfect

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...5000DPIWeb.jpg

The best ColorPerfect conversion I could achieve without increasing saturation. LINK TO 5000 DPI FILE

http://www.netsoft2k.com/Docs/Media/...O5000DPICP.jpg

http://www.netsoft2k.com/Docs/Media/...00DPICPWeb.jpg

Thanks for your hard work. SO what setting should I use on my Epson V600? I shoot medium format 6x7.

Great thread, thanks for doing this!

Quote:

---

Originally Posted by Pali K

Ted, I had not and have never heard of this test but I can definitely look into it.

---

It is one way to measure the MTF of the scanner, for example rather than trying to measure what can be resolved at around 5-10% contrast which is what the USAF target your are using does. You can measure what can be resolved with a contrast of say 50%, or any contrast level for that matter. It is a much better comparison method.

This blog explains https://harvestimaging.com/blog/?p=1328

There is some software in imageJ to the calcs (Its on my todo list to test)

Quote:

---

Originally Posted by Pali K

I am using this image for Color Negative testing because the leaves are subtle variations of colors that will be interesting to see across the scanners.

---

Yes that looks pretty good. I was wondering if you have a photograph taken of color checker chart, in daylight. I am interested in comparing with can be achieved by other scanners or a DSLR so having a target that anyone can create if they have same stock might be useful and for other selfish reasons but perhaps that will take this thread on a tangent.

Quote:

---

Originally Posted by Pali K

This is the best resolution my Eversmart PRO can achieve. Both Vertical and Horizontal are either Group 5 - Element 5 or 6 (~2500 - 2800 DPI)

UPDATE: I zoomed in the image and it seemed the resolution is greater than what is visible on the original image due to file size so I scanned it again at 8000 DPI max allowed and upscaled to 11000 DPI. Here is the result. I think the resolution is safely Group 5 Element 6 putting it at ~2896 for both Vertical and Horizontal.

---

Pali, in this case I'd also rate it better, in Element 6.0 (at bottom) bars orientation is perceived with absolutely no doubt. To me it's 3250dpi.

Thanks for doing this comparison. I appreciate that you have included the raw negative scans. I wouldn't have processed your leaf picture the same and being able to compare them using my normal workflow is nice. Like Ted Baker suggested it would be nice to include an example with a color checker, or at least some reference colors with known neutrals and skin tones.

Quote:

---

Originally Posted by Alan Klein

Thanks for your hard work. SO what setting should I use on my Epson V600? I shoot medium format 6x7.

---

Thank you! What software are you using? I would recommend scanning at the highest DPI setting possible in 16 BIT RAW and then using Color Perfect to convert to positive.

Quote:

---

Originally Posted by EH21

Great thread, thanks for doing this!

---

Thank you!

Quote:

---

Originally Posted by Ted Baker

It is one way to measure the MTF of the scanner, for example rather than trying to measure what can be resolved at around 5-10% contrast which is what the USAF target your are using does. You can measure what can be resolved with a contrast of say 50%, or any contrast level for that matter. It is a much better comparison method.

This blog explains https://harvestimaging.com/blog/?p=1328

There is some software in imageJ to the calcs (Its on my todo list to test)

---

Thank you Ted - learned something new. I will try to see if this is something I can incorporate towards the end of the other planned tests. Since the resolution is consistent across all scanners, I would say that we should at least get a reliable comparison even if it doesn't test the performance across all possible scenarios.

Quote:

---

Originally Posted by Ted Baker

Yes that looks pretty good. I was wondering if you have a photograph taken of color checker chart, in daylight. I am interested in comparing with can be achieved by other scanners or a DSLR so having a target that anyone can create if they have same stock might be useful and for other selfish reasons but perhaps that will take this thread on a tangent.

---

I can look into this for you Ted but likely something that will be done at the end, if time permits :)

Quote:

---

Originally Posted by Larry Gebhardt

Thanks for doing this comparison. I appreciate that you have included the raw negative scans. I wouldn't have processed your leaf picture the same and being able to compare them using my normal workflow is nice. Like Ted Baker suggested it would be nice to include an example with a color checker, or at least some reference colors with known neutrals and skin tones.

---

You're welcome Larry! I will be including a variety of other color negatives and will keep some normal scenes in mind as I select the right LF and 35MM film to scan. When all said and done, we should have a comparison across multiple scenes.

Quote:

---

Originally Posted by Pere Casals

Pali, in this case I'd also rate it better, in Element 6.0 (at bottom) bars orientation is perceived with absolutely no doubt. To me it's 3250dpi.

---

Pere, I think you are correct on both the Epson and Eversmart Pro. I'll update the DPI accordingly when I summarize the tables. Thank you.

Thinking about using this image for 35MM Color Negative scans. This image seems to have quite a variation of light and colors from high saturation to pastel. The scene itself is a common one so we should know what the true colors are. Any thoughts before I start scanning?

https://farm6.staticflickr.com/5446/...162a7872_b.jpg

Is it just my old eyes or is the image a bit on the soft side??? (Nothing wrong with a soft image - I like it.) But if one is going to compare scans wouldn't a crisper image be better?

Quote:

---

Originally Posted by Pali K

Thank you! What software are you using? I would recommend scanning at the highest DPI setting possible in 16 BIT RAW and then using Color Perfect to convert to positive.



Thank you!

---

I use Epsonscan with my V600 and usually scan at 2400. With medium format that goives me 200mb files. Highest DPI isn;t practical. It would give me 1tb + files. At what pointt does the scanner really not distinguishing pixels?

Quote:

---

Originally Posted by Alan Klein

I use Epsonscan with my V600 and usually scan at 2400. With medium format that goives me 200mb files. Highest DPI isn;t practical. It would give me 1tb + files. At what pointt does the scanner really not distinguishing pixels?

---

Alan, with the V600 scanning at 2400dpi or a bit higher is OK, as you noted, in practice you won't improve much from scanning at higher dpi. I had the similar V500.

https://www.filmscanner.info/en/Epso...V600Photo.html
https://web.archive.org/web/20180131...V600Photo.html

While a very basic film scanner, with MF film for sure it will deliver good enough images to be displayed in monitors and tht are also good for moderate size prints, if no high densities there having to be recovered.

A used V700 for some $250 would be a big improvement, if you can sell the V600 this would be a $150 investment, allowing also for LF scanning.

Maybe we should let this thread be strictly about Pali's comparisons, using other threads for advice on particular workflows.

Quote:

---

Originally Posted by Peter De Smidt

Maybe we should let this thread be strictly about Pali's comparisons, using other threads for advice on particular workflows.

---

I agree, this is a really interesting initiative and better if we don't go much off topic.

Quote:

---

Originally Posted by Alan Klein

I use Epsonscan with my V600 and usually scan at 2400. With medium format that goives me 200mb files. Highest DPI isn;t practical. It would give me 1tb + files. At what pointt does the scanner really not distinguishing pixels?

---

According to filmscanner.info:

"In order to achieve the maximum resolution of 1560ppi, one does not have only to scan the with the highest optical resolution but it is sufficient to digitalize the original with 3200ppi. The effectively achieved resolution does not differ, no matter if one scans with 3200ppi or with 6400ppi - in both cases, an effective resolution of 1560ppi will be achieved!"

I scan MF at 4300 (optical) on my IQsmart and 6 x 6 scans run around 500MB

I think you meant 1GB, not 1TB

By the way I often get 900MB files from 6 x 12. No issue working them.

Quote Originally Posted by Alan Klein View Post
I use Epsonscan with my V600 and usually scan at 2400. With medium format that goives me 200mb files. Highest DPI isn;t practical. It would give me 1tb + files. At what pointt does the scanner really not distinguishing pixels?

Quote:

---

Originally Posted by Jim Andrada

Is it just my old eyes or is the image a bit on the soft side??? (Nothing wrong with a soft image - I like it.) But if one is going to compare scans wouldn't a crisper image be better?

---

Good point - its too soft so I'll find another image to go by. Thanks Jim!

I'll be scanning and posting more results tonight. Stay tuned :)

It's OK to say that it's my fault because I'm getting old - my wife always says things like that so I'm used to it - except she'd blame my brain instead of my eyes. (just kidding of course)

Creo/Scitex Eversmart Supreme | OxygenScan 2.6.4 | IT8 Colortarget 16 BIT RAW (DT)

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...5000DPIWeb.jpg

Creo/Scitex Eversmart Surpreme | OxygenScan 2.6.4 | USAF 1951 Resolution Target 16 BIT RAW (DT)

This is the best resolution my Eversmart Supreme can achieve when scanned at 11000 DPI. I think resolution is around Group 6 Element 5 putting it around 5100 DPI.

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...1000DPIWeb.jpg

UPDATE: Rescanned the resolution target with the Black Frame Glass vs. Purple Frame Glass. I can see lines upto Group 7 Element 1 but lets go with Group 6 Element 6 to be safe which puts DPI resolution around 5700 DPI. Impressive :)

http://www.netsoft2k.com/Docs/Media/...ckGlassWeb.jpg

Creo/Scitex Supreme | OxygenScan 2.6.4 | Kodak Portra 160 Color Negative 16 BIT RAW (DT)

LINK TO FILE

http://www.netsoft2k.com/Docs/Media/...5000DPIWeb.jpg

The best ColorPerfect conversion I could achieve without increasing saturation. LINK TO 5000 DPI FILE

http://www.netsoft2k.com/Docs/Media/...P5000DPICP.jpg

http://www.netsoft2k.com/Docs/Media/...00DPICPWeb.jpg

One thing not mentioned here, is a wet scan on the v700. It does make a huge difference...enough that I will use my v700/wet over my dedicated 35mm scanner with 35mm film.

Although I have no dpi figures, it resolves the grain much better.

I think that Supreme might need a bit of cleaning up of dust. Vertical lines in USAF 1951 target get messed up well before horizontal lines. This is how it should look like.

Attachment 186459