Achievable resolution with Epson V700/V750/V800/V850

alanrockwood · Dec 4, 2017

OK, let me first acknowledge that this is no a new topic, but let me explain why I think a new discussion is worthwhile.

Generally speaking, most reviews say that these scanners are capable of about 2200 to 2400 effective pixels per inch. However, what I would be interested in is what is the best one can do if the film holders are very carefully adjusted to the optimum height (possibly with aftermarket film holders) and if and appropriate level of sharpening is applied and if one scans at a setting 6400 pixels per inch. Ideally the sharpening would be done using a true deconvolution algorithm and the best point spread function for the deconvolution process, but even unsharp mask results would be of interest.

I haven't read any reviews or evaluations or discussions that address those optimal conditions. I'm going to take a wild uninformed guess that one might be able to get into the range of 3000 effective pixels per inch, or maybe a bit more. Does anyone know of real tests that try for the best possible results when varying the parameters that are under user control to their optimal values? I did see one web page where the results seemed to be visually as good as a 4000 dpi scanner (possibly even a bit better), but the work was not described well enough to go beyond a visual impression.

Ted Baker · Dec 4, 2017

The method of testing for scanners is not the same as is used by most reviewers, that is why the value is much lower. If you used the ISO method for scanners you would get 6400 dpi, but you wound not be able to see it

It also slightly more complicated because the scanner actually scans at its native resolution all the time, it just re samples the data give a lower resolution. I believe the native optical resolution of the sensor lens combination is 3200dpi, but it uses a technique called pixel shifting to improve the resolution.

It seems reasonable to consider using re sampling algorithm that is optimised for pixel shifting, this is on my list to experiment with.

Down Under · Dec 4, 2017

Interesting indeed...

Could this thread possibly be expanded to include the two 'lesser' models of Epson scanners, the V500 and V600? For a full comparison's sake and for the greater edification of those of us who are making do with those two budget models, and to give us the full picture on the resolutions and other differences between the entire range of six models.

I know of at least six photographers who still scan with Epson 4990s, but it may be asking too much to also include this venerable veteran in the list...

Alan Johnson · Dec 4, 2017

My test agreed with Scandig.com and Sandy King:
IMO what you suggest won't make much difference.
https://www.photrio.com/forum/threads/resolution-of-the-epson-v700.150649/
The pic at the foot of this page confirms that for 35mm the Plustek has higher resolution, in agreement with Scandig.
http://www.photo-i.co.uk/Reviews/interactive/Epson V700/page_6.htm

Ted Baker · Dec 4, 2017

ozmoose said:
Could this thread possibly be expanded to include the two 'lesser' models of Epson scanners

I believe they are all essentially the same, i.e. they all use pixel shifting, they scan at their respective native resolution and resample, so the concepts are the same. I have the 4990.

Frank53 · Dec 4, 2017

It is something that has puzzled me for years.
I’ve been using a V700 since this model exists. The results are useable. What always puzzled me, was that as soon as you start “pixelpeeping”, you see it is not sharp, the film grain is never really sharp. I use the original Epson filmholders, with anti newton glass on the hegatives, so the film is flat.
Next to the V700 I always used a Minolta 5400. Now those scans were sharp! Sometimes even a bit too sharp.
Because the Minolta only did 35mm, I got an Imacon Precision II. It scans 35mm at 6300 and 120 at “only” 3200.
But at 3200 the scans are so much better than the V700 at 3200 or 6400.
So my unscientific conclusion is that the resolution of the V700 is way below 3200.
Regards,
Frank

removed account4 · Dec 4, 2017

can the untrained eye really tell if a pixel has been shifted ?

i mean, there have been treatises written about how 35mm shouldn't
be enlarged past 8x10, and i've spoken with art fair folks about
how digital files ( from a camera ) should never
be cropped or manipulated ... because the image becomes degraded..

Ted Baker · Dec 4, 2017

jnanian said:
can the untrained eye really tell if a pixel has been shifted ?

I think that's the point, it gives a measurable bump in resolution, that's all. It is used in the many of the epson scanners, it makes sense to consider it during resampling, others have even if they have not understood what really happens under the covers just by doing simple visual tests.

removed account4 · Dec 4, 2017

thanks ted

Les Sarile · Dec 4, 2017

alanrockwood said:
I'm going to take a wild uninformed guess that one might be able to get into the range of 3000 effective pixels per inch, or maybe a bit more.

Why guess when it sounds like you have a basis for a good test. Can't you test this yourself?

Patrick Robert James · Dec 4, 2017

I think you already have your answer. 2300. The problems with flatbeds are the pedestrian optics including the mirrors, and the scan is made through glass. If you removed the glass you would probably see an increase in resolution and clarity since you are reducing flare, but it might not be that much. Replacing the lens is a no go. If you are only going to use the scanner for film, removing the glass might be worth your while.

alanrockwood · Dec 4, 2017

Les Sarile said:
Why guess when it sounds like you have a basis for a good test. Can't you test this yourself?

It is a worthwhile thought, but no, I am not in a position to do a rigorous test, especially not one under truly optimal conditions. For example, I don't have film holders that provide continuous height adjustment, so I can't do a test at the optimal film height. I also don't have software to sharpen using deconvolution. I would also need to obtain some appropriate resolution targets and learn how to use them expertly.

It would be better to see results from someone who is truly an expert in these kind of measurements in order to take the human element out of it as much as possible, and why re-invent the wheel if someone else has already done it?

Alan Johnson · Dec 4, 2017

Scandig.com have resolution targets:
https://www.scandig.com/accessories/targets/usaf-1951-test-target.html

faberryman · Dec 4, 2017

You can optimize the V700, V750, V800, V850, and maybe increase resolution somewhat, but you are never going improve them enough to get the same resolution of a dedicated 35/120 film scanner.

Prof_Pixel · Dec 4, 2017

jnanian said:
i've spoken with art fair folks about
how digital files ( from a camera ) should never be cropped or manipulated ... because the image becomes degraded..

That's because of the way JPEG compression works by breaking the image down into small multipixel areas. If you crop a JPG image and then recompress it using JPEG compression, the small areas won't match the original ones and image quality can be lost.

faberryman · Dec 4, 2017

A good reason not to use jpgs in your workflow (except for the web).

Bob Carnie · Dec 4, 2017

faberryman said:
A good reason not to use jpgs in your workflow (except for the web).

only reason I would convert to jpeg is for web purposes.

ParkerSmithPhoto · Dec 4, 2017

Getting past all of this can be confusing. It was for me. I finally tested a 6x7 negative, shot on a tripod with mirror lock up. One silver print on 16x20 paper. One negative scan at 2400ppi and matching print size. The prints were essentially identical except for the matt/glossy surfaces. I placed them under glass and it was almost impossible to distinguish one from the other.

It's always better to scan "clean" data, and then res up from there in Photoshop, rather than trying to extract detail that just isn't in the negative. For medium format B&W films I always use 2400ppi. It's plenty large.

MattKing · Dec 4, 2017

There is only one reason to pay any attention whatsoever to "ppi" settings when one is scanning.
If your entire reason for scanning is to create a file that will be printed to a single size, using the "ppi" setting can help you work backwards to an appropriate scanning resolution for a file that will not need any re-sizing prior to printing.
For 99.999% of the time, the "ppi" setting should be ignored while scanning.

faberryman · Dec 4, 2017

MattKing said:
There is only one reason to pay any attention whatsoever to "ppi" settings when one is scanning.
If your entire reason for scanning is to create a file that will be printed to a single size, using the "ppi" setting can help you work backwards to an appropriate scanning resolution for a file that will not need any re-sizing prior to printing.
For 99.999% of the time, the "ppi" setting should be ignored while scanning.

Tough to ignore when it is one of the software choices, and the selection has an impact on the resolution of the scan.

MattKing · Dec 4, 2017

faberryman said:
Tough to ignore when it is one of the software choices, and the selection has an impact on the resolution of the scan.

I'm not saying to ignore it on software that forces you to choose something - not all scanning software does. I'm saying that you need to let it fall where it may, when you control the resolution choice that does matter - the scanning resolution.

faberryman · Dec 4, 2017

MattKing said:
I'm not saying to ignore it on software that forces you to choose something - not all scanning software does. I'm saying that you need to let it fall where it may, when you control the resolution choice that does matter - the scanning resolution.

What software doesn't give you a choice of scanning resolution, except perhaps the top end models that always scan at their maximum resolution? I don't understand the laissez-faire attitude of "let it fall where it may", whatever that even means, but I'm open to learning if that means better scans.

removed account4 · Dec 4, 2017

Prof_Pixel said:
That's because of the way JPEG compression works by breaking the image down into small multipixel areas. If you crop a JPG image and then recompress it using JPEG compression, the small areas won't match the original ones and image quality can be lost.

thanks !

i appreciate knowing how things work...
i typically work off of TIFF .. but
i know people ...... " i only shoot jpg "
what can you do ..

Alan Edward Klein · Dec 4, 2017

ParkerSmithPhoto said:
Getting past all of this can be confusing. It was for me. I finally tested a 6x7 negative, shot on a tripod with mirror lock up. One silver print on 16x20 paper. One negative scan at 2400ppi and matching print size. The prints were essentially identical except for the matt/glossy surfaces. I placed them under glass and it was almost impossible to distinguish one from the other.

It's always better to scan "clean" data, and then res up from there in Photoshop, rather than trying to extract detail that just isn't in the negative. For medium format B&W films I always use 2400ppi. It's plenty large.

What scanner were you using? How did you process the scan (sharpen, etc.)? what dod you mean "except for matt/glossy surfaces"? Thanks.

jim10219 · Dec 4, 2017

Remember, some scanners will apply their own sharpening without telling you. Also, the upscaling or sharpening algorithm used will change the results you get. And they will add information that wasn’t there in the initial scan. And how well all of that works can be effected by the photo itself.

My point being, don’t trust the numbers. They won’t help you much. Put your trust in your own experience. I was looking to replace my Epson 4990 until I learned how to use it properly. These days, user error is my biggest complaint. And when I scan 135 film, I use a DSLR because it’s quicker and has a higher resolution.

Achievable resolution with Epson V700/V750/V800/V850

alanrockwood

Ted Baker

Down Under

Alan Johnson

Ted Baker

Frank53

removed account4

Ted Baker

removed account4

Les Sarile

Patrick Robert James

alanrockwood

Alan Johnson

faberryman

Prof_Pixel

faberryman

Bob Carnie

ParkerSmithPhoto

MattKing

Moderator

faberryman

MattKing

Moderator

faberryman

removed account4

Alan Edward Klein

jim10219

Achievable resolution with Epson V700/V750/V800/V850

Moderator

Moderator

Privacy & Transparency

Privacy & Transparency