[stevebrot]

This is all very interesting since I was considering a scanner. Not so sure now. How does the Epson do for color transparencies (35 & 120)?

I would expect about the same as for negative film as far as resolution. Color fidelity is another matter entirely and may depend on the transparency medium. I have not tried slides on my V700, so cannot say.

...On 35mm, how close can a DSLR (K-7) with 1:1 macro lens/bellows/slide copier get to the Epson?

Good question. I know that a couple of forum members have used this option in the past for mounted slides. I don't have a slide copier attachment for my bellows or I would do the comparison. I guess that leaves it to you to do after you buy your scanner


Steve

[Jimfear]

I would expect about the same as for negative film as far as resolution. Color fidelity is another matter entirely and may depend on the transparency medium. I have not tried slides on my V700, so cannot say.Steve

The 700 has some problems with slides, I scan a lot of Velvia and that is quite dense. The problem is usually that highlights and shadows goes beyond the histogram edges, not much but still. A dedicated scanner with better Dmax and D range will probably do this better. I still am very happy with the results though.

[artobest]

The V750 does an OK job with 35mm slides. Scans need to be heavily post-sharpened to bring out the best in them. Depending on the subject matter - high-key images are worst - there is often some blooming and colour fringing, which can only be partially rectified in PP. Here's a sample - Velvia 100, no colour correction (it was a reddish, sunset scene), minimal capture sharpening with Photokit Sharpener. The crops are at 200%(!) just to give a clearer idea of how the pixels look.

My feeling about the Epson scans is that they lack the texture and tactile feeling that visible grain lends to a really good scan. They also do struggle to control highlight and shadow, but this shortcoming can be largely offset using the superb functionality of Silverfast Ai (comes free with the V750). With careful preparation, they can produce excellent prints from 35mm up to about A3 size - especially on matte papers - and acceptable ones larger than that, depending on the intended purpose. Obviously, the less enlargement one needs to perform, the better, and that's why the Epson is popular amongst medium- and large-format photographers.

[k0og]

Steve,

I've been following this thread because I have an Epson V500 (since Christmas), and am VERY happy with its performance. As long as I am careful to keep the negatives flat as you rightly emphasize, it does well.

I was curious about your comparison, so I tried something. I did a little 1.5 pixel/200% + 12 pixel/20% USM on the right-hand side (the V700 image), and it looks MUCH closer to the resolution of the Nikon. Is it possible we are dealing at least in part with how the scanning software sharpens the data?

Interesting thread!
-Joe-

For completeness and also because I can be a little obsessive, here is a comparison between the two scanners at 4000 dpi. That is the max for the Nikon.


Nikon 5000 ED vs. Epson V700 @ 4000 dpi

[IMGWIDELEFT]http://farm5.static.flickr.com/4062/4653365304_5c1fd9f871_b.jpg[/IMGWIDELEFT]

The downsampling for the Epson was done at scan time using the Epson Scan software.



Steve

Attached Images

 

[Ben_Edict]

I don't know what downsampling algorithm the Epson software uses, but it doesn't look great. So soft after downsampling? You might get better results downsampling in PS using Bicubic Sharper, or better still trying one of the other, more esoteric algorithms out there, such as the ones available in Qimage.

Downsampling the Epson scan from 4800 to 4000dpi won't improve anything. After all the real scan resolution is around 2200-2300dpi and approximately comparable to the older Nikon LS30. I have the V700 and had the LS30 and sold the LS30 off after some initial comparissons. I usually do not scan 35mm but for larger formats the 2300dpi resolution of the V700 is enough for my purposes.

In my experience the software used with the scanner makes a lot of difference. I abandoned Epson Scan immediately and use VueScan and Silverfast. VueScan is a very easy to use solution and works fast and easy. Silverfast needs some tweaking, but for photogs its workflow is quite logical and it gives visibly (colour rendition and contrast) better results. Especially the multi-sampling feature is much better implemented than in VueScan, where it sometimes produces very visible haloing.

Adjusting the height of the film holders is very critical with the Epson scanners and has a large influence on the result.

Ben

[stevebrot]

Steve,

I've been following this thread because I have an Epson V500 (since Christmas), and am VERY happy with its performance. As long as I am careful to keep the negatives flat as you rightly emphasize, it does well.

I was curious about your comparison, so I tried something. I did a little 1.5 pixel/200% + 12 pixel/20% USM on the right-hand side (the V700 image), and it looks MUCH closer to the resolution of the Nikon. Is it possible we are dealing at least in part with how the scanning software sharpens the data?

Interesting thread!
-Joe-

I did my scans with scanner USM turned off.

I did the same experiment as you with similar results. Really cranking on the USM can increase the apparent sharpness substantially, but there is a price to pay in grain artifact and loss of border tonality. Notice that I said apparent sharpness. Ignore the railing and look at the details in the group of people on the left. The before and after are about the same for both scanners. While the railing is less fuzzy, the actual data remains unchanged.

Part of the "sharpness" from the Nikon has to do with a strongly focused light source. The Epson's is very diffuse. The situation is much the same as a diffusion vs. condenser head on an enlarger. Resolution remains the same between the two, but local contrast is higher with condenser heads.


Steve

[k0og]

Steve,

Thanks for all the testing you did, the explanations, and the followup by others. After all this, I feel confident I have an excellent solution with my V500, especially for the price (well, sort of - since it was a gift to me by my wife!).

My opinion is that considering the extreme pixel peeping, and practically speaking the difference is not all that great between the Epson and the Nikon at 4000 dpi, the V500 should give me excellent results for anything I would use it for (on-line posting, and occasional prints up to 8x10).

I may eventually get some ANR glass to help keep the negatives flat to take full advantage of the scanner's maximum capability. It seems that is the major shortcoming of the flatbed solution.

-Joe-

I did my scans with scanner USM turned off.

I did the same experiment as you with similar results. Really cranking on the USM can increase the apparent sharpness substantially, but there is a price to pay in grain artifact and loss of border tonality. Notice that I said apparent sharpness. Ignore the railing and look at the details in the group of people on the left. The before and after are about the same for both scanners. While the railing is less fuzzy, the actual data remains unchanged.

Part of the "sharpness" from the Nikon has to do with a strongly focused light source. The Epson's is very diffuse. The situation is much the same as a diffusion vs. condenser head on an enlarger. Resolution remains the same between the two, but local contrast is higher with condenser heads.


Steve

[goddo31]

Hi guys,
Well after buying an Epson V700 recently it's about time I had a go at providing some sample images for comparison. The following examples are both from a roll of Fuji 160C I shot earlier this year. Something to note that's most likely irrelevant, is that this film was shot while travelling; so it would have been x-rayed multiple times...

I will post up 2 images, with the following identical procedure:
1. Scan with Epson scan at 6400 dpi
- no USM, auto exposure, no colour correction or anything else enabled
2. Resize to 7MP in lightroom, perform basic colour corrections
- I was trying to resize to roughly 2400dpi, probably didn't do it the best way
- Basic colour corrections to restore colour vibrancy to to what I saw
3. On the sharpened versions, perform significant USM and about 20 clarity

Note that on the complete images, I have resized smaller again for web. But the 100% crops will probably be the most useful. It's also worth noting that the film I scanned could be a bit flatter!

Image 1 - complete and non-sharpened


Image 1 - complete and sharpened


Image 1 100% crop - non-sharpened


Image 1 100% crop - sharpened

[goddo31]

Ok and in the same format, here is my second image...

Image 2 - complete non-sharpened


Image 2 - complete with sharpening


Image 2 - non-sharpened 100% crop


Image 2 - sharpened 100% crop

[artobest]

Those scans are looking good - in particular, you seem to be hitting the sharpening sweet spot.

The only thing I can add is to suggest you look at doing some noise reduction - in both instances the grain has been sharpened, although it's really only the chroma noise in the second crop that is objectionable. This is very easy to mitigate in Lightroom if that's where you do your sharpening.

If you're using USM in PS, you should probably apply an edge mask. Also, you could try using a stand-alone noise-reduction program such as Neat Image before you sharpen.

[goddo31]

Thanks for the suggestions and tips artobest. I was worried that I had oversharpened a little bit. I'm adding about "60" on the sharpening slider in Lightroom here. I will have a play with the NR controls in Lightroom - I guess I am a sucker for trying to get as much detail as I can, normally I go for no NR (in digital shots too).
I'll have to at edge masking too, only just got photoshop.

cheers

[Vertex Ninja]

Goddo31,
I agree, these look nice. If they were oversharpened you'd see halos at 100%. If you are using lightroom it has edge masking built in; just under the detail slider in the sharpening area. Zoom to 100%+ and hold down 'alt' as you move the slider to see the mask. If you're doing the sharpening in PS, I created some actions a few years ago, and put them in the post processing forum. The links may be broken now, not sure. I'll upload them here for you if you'd like.

[lbenac]

Well I just received the quote from Nikon repair center. 1K with taxes to fix the LS-8000. The circuit board itself is $560!
So I learned my lesson about buying from eBay. The machine worked for a little bit and then came the dreaded message about holder not been recognized.
I am not going to make the repair as there was report on the web that even with the repair the scanner was still failing.
So now this V700 start to look better and better. I will have to sacrifice scan quality to having a working scanner.

Cheers,

Luc

[artobest]

I will have to sacrifice scan quality to having a working scanner.

A little ... but not too much. You could shrink the gap still further by upgrading to a Silverfast edition that has Multi-Exposure, to extend the V700's dynamic range - one area where I feel most film scanners have the advantage. I don't use M-E, and I hear that it's not useful in every instance, but it might be nice to have for those difficult images.

[stevebrot]

Well I just received the quote from Nikon repair center. 1K with taxes to fix the LS-8000. The circuit board itself is $560!
So I learned my lesson about buying from eBay. The machine worked for a little bit and then came the dreaded message about holder not been recognized.
I am not going to make the repair as there was report on the web that even with the repair the scanner was still failing.
So now this V700 start to look better and better. I will have to sacrifice scan quality to having a working scanner.

Cheers,

Luc

Sorry to read that Luc. Is there any chance you can just purchase the board and do the replacement yourself?


Steve