[hinman]

Much thanks for the detailed reply. This has given me an alternative to my current flatbed scanner. It will be a while before I can start following your method for my home use. I will try to get something similar to the light-box as in the first link that I post. I actually need some gadget to preview film and this will come in handy for my home use. And I will look for the software on Mac. And I want to thank you for your time and I find your suggestions intriguing and informative.

Thanks,
Hin

[fotoreporter1975]

Yu're welcome!

About the software, remember that aside PanoramaMaker you can also use:

- if you have Photoshop its function "photomerge" (works perfectly 90% of the time)

- the free, multiplatform, Hugin (a bit intimidating at first, but really powerful)

- if you have a digital Canon camera (dslr, compact or whatever) it comes with a copy of Photostitch; it sports really good results, the only limitations being the lack of manual corrections (the same of photomerge, though) and a propensity to crash if used with files too large

and I think even the free Gimp used to have some panoramic plugin, but for the life of me I can't remember its name.

[jt_cph_dk]

I have been looking at this thread and your link: Best Film scanner.. many times over the last days. Now here is an example of mine. The scan took minutes. I think you are doing something wrong when scanning? Regards

JPG from original: Pentax 6X7@1:20s (handheld, no MLU), Super-Takumar 6X7 4/200@4, 81C, TMY400, Epson V750, Silverfast 6.6@2400ppi, USM: 70%, 1.7, 5X5

Please check the ‘Original‘ (JPG developed for web, not the scanned 105mb, 8bit, grey .tif):
http://www.flickr.com/photos/32108329@N08/8216193789/sizes/o/in/photostream/

100% crop of the .jpg in large (1024px) format


This is a 100% crop of the scanned tif file. A print of the whole file would be about 120X170cm. at 300ppi


Note: The last crop was automaticly reduced in size. See the actual one here: http://www.flickr.com/photos/32108329@N08/8223474853/sizes/o/in/photostream/

[tuco]

With positive film dynamic range is not a problem for a digital camera. But what about BW film that captured more DR than a digital camera can? I mean, do you lose that in this conversion?

[tuco]

With positive film dynamic range is not a problem for a digital camera. But what about BW film that captured more DR than a digital camera can? I mean, do you lose that in this conversion?

You're doing a single sample, right? If your scanning software and scanner supports multi-sampling, it takes much longer. With enough sampling you can reduce the scanner noise as well as improve your dynamic range.

[jt_cph_dk]

You're doing a single sample, right? If your scanning software and scanner supports multi-sampling, it takes much longer. With enough sampling you can reduce the scanner noise as well as improve your dynamic range.

Me? I´m doing a single sample (I tried a multiple 4800ppi scan for noise reduction, also including ICE once—it took six hours! ). The scan here doesn´even have ICE, the negative came straight from a very clean pro lab and out the plastic strip and there isn´t much dust on it. By the way: ICE does impare the files and leave jpg or lightning like marks, than can ce seen when pixel peeping. Regards

[tuco]

The scan here doesn´even have ICE, the negative came straight from a very clean pro lab and out the plastic strip and there isn´t much dust on it. By the way: ICE does impare the files and leave jpg or lightning like marks, than can ce seen when pixel peeping. Regards

Good thing because you can't use ICE on BW film.

Also, you usually can set the level that ICE will clean. Set it to small and it will look the same as if you had to go in and spot heal it yourself. Large scratches and dust are better cleaned up by hand anyway.

[jt_cph_dk]

You are right. I was thinking of other scans, that I have been doing the last days (just went through 10 rolls!). The SilverFast 6.6ai that I´m using has two ICE settings. One fast and one best.

..and back to the interesting point: Can you optain more quality (sharpness, less noise, more dynamics etc.) with the sugested rig, than on a scanner?

[fotoreporter1975]

I have been looking at this thread and your link: Best Film scanner.. many times over the last days. Now here is an example of mine. The scan took minutes. I think you are doing something wrong when scanning? Regards

JPG from original: Pentax 6X7@1:20s (handheld, no MLU), Super-Takumar 6X7 4/200@4, 81C, TMY400, Epson V750, Silverfast 6.6@2400ppi, USM: 70%, 1.7, 5X5

Please check the ‘Original‘ (JPG developed for web, not the scanned 105mb, 8bit, grey .tif):
All sizes | Gustav | Flickr - Photo Sharing!

100% crop of the .jpg in large (1024px) format


This is a 100% crop of the scanned tif file. A print of the whole file would be about 120X170cm. at 300ppi


Note: The last crop was automaticly reduced in size. See the actual one here: All sizes | detail from original tif scan file of ?Gustav? | Flickr - Photo Sharing!

Hi, you've got some sharp shots here!

Keeping in mind sample variation - i.e. my scanner can be not the best of the bunch or yours can be an exception - have you used the technique discussed here with this same frame or observed the film with a powerful loupe, like 22x? I was kinda content with the results of my Epson, until I started to look at how much more detail there was actually on film.

To answer the second question: my scans took a lot longer, because I noticed that - at least with my own scanner - I was able to extract more details from the film scanning at 6400, sharpening a bit the results, than resizing to 3200.

And, regarding the quality, it depends on how much close you go to your film, aka the reproduction ratio. I normally use 1:2 for the bulk of the shots, and 2:1 or 3:1 for the very best.

Here there are another example, to view what I mean. It has been shot on Hasselblad with the normal 80mm Planar, camera on tripod, soft shutter release 1m long etc. on Ektar 100 iso (the new kind). Don't take the tonality into consideration because I'm still struggling to find the perfect curve (being a color negative).

First the whole image



Then the 100% crops:

[tuco]

...
To answer the second question: my scans took a lot longer, because I noticed that - at least with my own scanner - I was able to extract more details from the film scanning at 6400, sharpening a bit the results, than resizing to 3200.

Your scanner is a V700 and you are scanning at 6400dpi? Commodity flatbed scanners can't do much more than 2400dpi.

Your results look good

[fotoreporter1975]

Your scanner is a V700 and you are scanning at 6400dpi? Commodity flatbed scanners can't do much more than 2400dpi.

Your results look good

Yes, indeed 2400 seems to be the real resolution of the Epson, not only by the many tests I saw online but also according my own.

The problem is that the Epson does not reach this resolution at the "2400" setting…

To be able to reach 2400 "real" dpi you, than, have to scan @ 6400 and downsample to a more manageable - and realistic - size. It was extremely time consuming, but the results were worth the effort, at least for the best images and at least before I found this other way

[tuco]

...
To be able to reach 2400 "real" dpi you, than, have to scan @ 6400 and downsample to a more manageable - and realistic - size.

Yes, scanning is slow and a laborious task indeed. It is something you do for either the love of film or old cameras not for the practicality of its workflow.

I've pretty much only used VueScan but it has a feature called "Size Reduction". My 9000ED can pretty much do 4000dpi. So I set it to that and set size reduction to 2 and it will average a 4000dpi scan to create a 2000dpi file size. It works pretty good. I'll repost a sample I did sometime ago when I first tried the New Portra 400.

2000dpi Scan

[jt_cph_dk]

Yes, scanning is slow and a laborious task indeed. It is something you do for either the love of film or old cameras not for the practicality of its workflow.

I've pretty much only used VueScan but it has a feature called "Size Reduction". My 9000ED can pretty much do 4000dpi. So I set it to that and set size reduction to 2 and it will average a 4000dpi scan to create a 2000dpi file size. It works pretty good. I'll repost a sample I did sometime ago when I first tried the New Portra 400.

2000dpi Scan

That´s a wonderful shot and the scan looks really fine. Perhaps there is even more detail to scan into?

[jt_cph_dk]

4) I was too concerned about color reproduction, even if, like I said, I'm mostly a b/w aficionado. But then I "scanned" this way an old project that consisted of photos with almost out of gamut (adobe rgb gamut) bright reds, oranges and yellows. I scanned this before with various film scanners (some problem mostly with the yellows) and with the Epson (the reds were mostly gone); the camera nailed the tonalities. And this was the first setup that almost (it required a little pull to the blue curve) nailed the peculiar shade of Velvia green of some of my slides. The paramount thing to avoid headaches is to shot with a custom white balance tuned to the light source you use; if you do so than the files require little or none post processing color-wise.

You really awake my interest here. Could you post this?

[tuco]

That´s a wonderful shot and the scan looks really fine. Perhaps there is even more detail to scan into?

Thanks. It was a full moon that night.

Might be more detail. But for my images destine for web display, it becomes a point of diminishing return for me. At 2000dpi I'm already scaling down a ~5000 pixel image to a 1000px image. I prefer the smaller file for these. And lets hope the strength of my image does not depend on seeing detail to the Nth degree. Photography is more than that.