Epson V750 vs Imacon 949

[scovell001]

Joanna,

Thanks for the PM reply.

I want to keep this test to 1 transparency if I can, however looking at the examples above, I'm thinking 2 might possibly be a consideration. I don't want to prove a point by giving out a hugely underexposed tran. I'm looking for an image which has been exposed normally. But, by combination of what it is, it has shadow detail to resolve along with fine detail resolution & highlights. In other words a well exposed but full dynamic range 'stress' type image. I've got a couple of trans that come to mind.

David Whistance doesn't know this yet. But, I'm thinking of asking him to scan the final prints with his betterlight back just to take any form of scanning prints de-resolution into the comparison.

I'm thinking out loud about what kind of printing stock would be the best to demonstrate any differences. Innova USG or Photo-rag would be normal exhibition choices. But, they have a texture (masking a little detail) would a full gloss be the better choice as this is more of a 'science experiment' ?

What does the floor think. a)exhibition papers you'd normally print with? or b)ultra gloss paper to show every detail?

The final results will be posted here as thumbnail crops, and I'll post the original files for download on my website so everyone can download and peep to their hearts content.

I'll put instructions in with the tran. But, just to summarise, what I want you to do is provide 2 versions of the tran. 1 with absolutely no CMS of any kind (a raw scan), and a 2nd with full CMS. No sharpening and at 2000dpi, don't worry about spotting.

[Dave Tolcher]

Not wanting to expand the scope but.... I have been playing with a Microtek F1 flatbed. I have to say that I am impressed by what I have seen so far. Most striking is the complete absence of any shadow noise and colour cast which I found impossible to control with the Epson. Sharpness and focus across the frame is excellent. I have seen some slighting stepping on highlight edges but very limited & am yet to determine why it occurs so little of the time and so randomly. Scans straight out of silverfast are clean and excellent colour fidelity - even though I havent calibrated the scanner yet. The big downside is lack of ICE on the transparency draw.

Reviews suggest that there isnt much difference for the price differential but I have to say that it is a significant step up from the V700.

If you wanted to include this in the test then I am happy to participate.

[DJ]

The F1 is the American model isn't it? I think the European model is the M1, which does have ICE, the disparity was something to do with licensing, otherwise they're identical.

Whilst these are flatbed scanners, the film scanning part is done *under* the glass, and this scanner actually focuses so it would be an interesting scanner to add to the comparison.

[timparkin]

Not wanting to expand the scope but.... I have been playing with a Microtek F1 flatbed. I have to say that I am impressed by what I have seen so far. Most striking is the complete absence of any shadow noise and colour cast which I found impossible to control with the Epson. Sharpness and focus across the frame is excellent. I have seen some slighting stepping on highlight edges but very limited & am yet to determine why it occurs so little of the time and so randomly. Scans straight out of silverfast are clean and excellent colour fidelity - even though I havent calibrated the scanner yet. The big downside is lack of ICE on the transparency draw.

Reviews suggest that there isnt much difference for the price differential but I have to say that it is a significant step up from the V700.

If you wanted to include this in the test then I am happy to participate.

Hi David - send me your address and it will go off to you today.. I'm just wet scanning the transparencies on my Epson..

Tim

[Dave Tolcher]

Hi, yes the M1 & F1 are the same except for ICE. This says F1 on the box and is European model, M1 was American and has no ICE as you say but other way around . I have just read the manual and it does indeed have ICE on the film drawer too, thought it was only on the flatbed as in the i900. Good news

[scovell001]

Tim,

Can we get the transparencies to Joanna next please.

[timparkin]

Tim,

Can we get the transparencies to Joanna next please.

Yep - can you or joanna send me contact details?

Tim

[scovell001]

For those that have been interested in this topic. We've finally got our results together & Tim Parkin is currently collating.

We'll be posting some comparisons very soon.

Best


Ian

[scovell001]

O.K. time for an update on our scanner testing! To bring you up to speed, I sent out 3 transparencies something dark, something light & something in-between. The most pertinent of which was the underexposed 5x4 a thumbnail of which is below:
The full size JPEG Imacon scan can be downloaded here: http://dl.dropbox.com/u/9599431/imacon9 ... covell.jpg
The full size JPEG Howtek 4500 scan can be downloaded here: http://dl.dropbox.com/u/9599431/Howtek4 ... Parkin.jpg
The full size JPEG Howtek 7500 scan can be downloaded here: http://dl.dropbox.com/u/9599431/Howtek7 ... mpbell.jpg
The full size JPEG Epson V750 scan by Tim Parkin can be downloaded here: http://dl.dropbox.com/u/9599431/epsonv7 ... parkin.jpg
The full size JPEG Epson V750 scan by Joanna Carter can be downloaded here: http://dl.dropbox.com/u/9599431/epsonv7 ... carter.jpg

[Dave Tolcher]

Very interesting, many thanks for doing this test and publishing the results.

The nett for me appears to be that all are good enough - very good in fact and that there is a variation in output based on operator. Is it possible that all scanners suffer disproportionally from the 'challenging slide' ? I have seen some stuff of Jon Brock's that Tim scanned (they were good slides) and were a cut above, esp in terms of colour, this isnt borne out in the test results shown.

Again many thanks, it took a lot of work.

Dave

[Charles Twist]

I agree all are very good and I rather suspect they would all print up fairly well (which was one of the important goals of this exercise, I thought - or is this still coming?).
I notice mainly that the desired information is present in all scans, but that the V700 suffers from greater noise. Joanna's scan has got the usual green rash problem which Tim's got a way of removing. I feel however he's over-compensated as the blacks are a bit red. Pushing the V700 shadows so that they give some visual presence in a print is difficult because of the green rash. Noise management can reduce it from green dots to a green veil, which can then be colour-balanced, but it's still not brilliant and does leave ghosting on the print without due care.
The drum scanner doesn't seem to suffer these issues.
Is there any evidence that the cut-off point at which the scanner decides the pixel is not dark but black, is different for the different scanners? Any ideas?
Regards,
Charles

[Patrick Dixon]

It's going to take me hours to download those to view them, but from the sections of the brh corner posted last night there seems to be more difference in colour than in anything else - with the howtek 7500 looking the most real.

[timparkin]

with the howtek 7500 looking the most real.

As it was on Velvia, this doesn't testify to the accuracy of the scan

[timparkin]

I agree all are very good and I rather suspect they would all print up fairly well (which was one of the important goals of this exercise, I thought - or is this still coming?).
I notice mainly that the desired information is present in all scans, but that the V700 suffers from greater noise. Joanna's scan has got the usual green rash problem which Tim's got a way of removing. I feel however he's over-compensated as the blacks are a bit red.

Strangeley, the very darkest parts of a transparency have a slight red cast (from what I can see holding transparencies up to the sun and to colour balanced lights).

Anyway - you can't push jpg's to see what is in the shadows, all sorts of strange stuff happens..

This is what the boosted shadows look like..
http://static.timparkin.co.uk/static/tr ... corner.jpg

The drum scanner doesn't seem to suffer these issues.
Is there any evidence that the cut-off point at which the scanner decides the pixel is not dark but black, is different for the different scanners? Any ideas?

Yep - Dmax is the scanners ability to read deep shadow and the howtek's photo multipliers just about get that far (at least on the transprencies I've been testing). It still gets a little noisy and push really hard and I do get some pattern noise. Also, I'm still amazed at how much detail velvia can manage. With a drum scan, I reckon you can get just about 7 stops of detail with reasonable colour. 2 above and 5 below

I should add that I've changed my scanning a little because I was using a small aperture to get the most detail out. However, this was adding more noise than was necessary and a larger aperture captured all of the same detail but removed a lot of noise (hence, my scans are now less noisy than these).

Tim

p.s. some of the effects you might be seeing are because the images are all in the Ektaspace colour space. This means unless you open them in Photoshop or Firefox 3+ or Safari then you won't be seeing the proper colours. I've converted the sample above to sRGB so should be displaying similiarly in all browsers..

p.p.s. Thanks to Ian for organising this, it's been quite enlightening and I'm hoping to write up my conclusions in the next few weeks..

[AbsolutelyN]

with the howtek 7500 looking the most real.

As it was on Velvia, this doesn't testify to the accuracy of the scan

Nor does it testify to the accuracy of Velvia. Glad you found using a larger aperture helpful.