One of my Palmer Park photos :>

[martian bachelor]

Taken from the overlook at the top of the bluff and park, looking out to the WSW

and Pikes Peak (right of center), this is nothing special... Actually it's a

test shot to get the filter factor for my center filter tweaked, and it's also

the first B&W neg I've tried scanning with my new Canon 9950F.<P>

 

Agfapan APX 100 film shot at ASA 64 w/K2 filter, 1/8th sec exposure. Original

scan at 1200 dpi (47.25/mm) was resampled down to 9% (0.81% by area!) for

posting here; simple crop/brightness/contrast/gamma corrections to scan, but no

sharpening.<P>

 

<A href="http://www.endlessrapture.info/506c_caf.jpg">Click here</a> to see the

unshrunk 5408 x 4224 (2.7Mb) version, which is about a 10 sec download @

broadband rates.<P><div></div>

[wmwhee]

Home sweet home.

[pete_andrews10]

Nice! But the 9950F needs some sharpening to make the best of it IME.

 

This is what I've found works best for me:

 

No USM applied in the scanner software, and 16 bit RAW scans imported to Photoshop, then inverted and curves adjusted. Most B&W negs need some manual reduction of the exposure setting in Canoscan to avoid posterisation of the shadows.

 

Scans done at 4800ppi will take two shots of the "sharpen more" filter in PS to get close to the sharpness of the original neg.

 

Scans at 2400ppi need one dose of "sharpen more", plus maybe some USM with a small radius (1.2 - 3 pixels)

 

I never use 1200ppi, it seems a waste of a decent scanner to me, 'cos if you've got the resolution - why not use it?

 

BTW, using Vuescan will overcome the stupid 10,000 pixel barrier in the acquisition software. As will scanning in portrait format.

[mick has already gone]

Watcha Bill,

lovely pic, but, what worries me is, I had a look at the big pic, right clicked it and saved it!!! Don't worry, I'm not a thief, I had a look at it in PS, fooled around wiv it and then deleted it all. However, it does show that some nasties could get into this site and steal pics for their own end. The copying that I did was executed before I'd even subscribed to this site, bit worrying.

Anyway, I've just started LF this year, I had a brief affair (about 4 months) with digital after my AV1 eventually gave up the ghost, got fed up wiv it and bought a Canon EOS33v, film yes, but all this computer stuff got me down, so, I bought a Shen Hao 5x4 in March. I've finally learned how to slow down and get some decent pics :-)

best

Mick Saunders

[mick has already gone]

Watcha All,

Well, I've screwed that, this pic belongs to Chris not Bill, sorry :-)

best

Mick

[martian bachelor]

Mick - I thought briefly about your concern and only can say I'd be flattered if someone thought enough of it to take it and do more than just look at it, like make money from it. What are they going to do when they're rich and famous and I can produce the original neg and point to my post where they obviously got it from? Not that I'd be likely to run across such a thief, but the risk seemed minimal and I wasn't going to paste my name across the middle of it and ruin it (anything inconspicuous down in a corner could easily be cropped out). Besides, I got plenty more where that one came from that are a whole lot better; it really was just an exposure test shot. But thanks for your concern. Feel free to keep a copy for personal use if you like.

 

Pete - thanks for the tips based on your experience. I tried some 4800 dpi scans first on an Ektachrome slide to see what the thing could do and noticed the same slight mushiness there as I did at 1200 dpi on this neg. Reviews have suggested 2400 dpi is about all one can really get out of the scanner, but it's something I'd like to do rigorous tests on to convince myself that it's true. But I thought it was pretty cool that the power lines about two miles out were easily visible. I had all the extra controls (like sharpening and grain reduction) OFF for the scan, in case anyone's interested.

 

I'm not particularly a fan of Vuescan based on trying it with some 35mm slides on another scanner, but it did strike me as strange that Canon would make a product capable of doing 20+ kilopixel scans in one dimension and then limit you to 10k... even a signed 16-bit integer should get one up to almost 33k, so it's something which required intention and extra lines of code in the program to put in place for whatever reason they did so.

[mick has already gone]

Watcha Chris,

 

know what you mean about being flattered if they nicked it and made money on it, I get that feeling most of the time :-) The point I was really making was, I just wandered in here, brought up a big pic, copied it and went off with it, that is a hole in the security of the site.

 

Anyway, I reckon it's a guddun for a test shot, all my pics are test shots :-)

 

Best Mick

[pete_andrews10]

Belated reply to Chris:

I'm fortunate enough to possess a glass resolution test plate that goes down to 2 micron bars (500 lppmm!). I scanned this with the 4490, and it resolved 8 micron bars (just), and 10 micron bars easily. This converts to an optical resolution of just over 3000ppi, which is more than enough for Medium and Large format scans, and even for most 35mm scanning. In fact, on an A3 size print from a 35mm slide, it's difficult to see any difference between scans done with the 4490, and with a 3600ppi dedicated filmscanner.

 

However, film positioning is quite critical with the 4490, since it has no focusing facility, and I must admit that the filmholders could be a lot better made.

[pete_andrews10]

Sorry! That (500 lppmm!) should have been (250 lppmm).

[pete_andrews10]

Sorry again! Where's my head these days?

 

I kept referring to a 4490 in the post above, when I meant the 9950. To make it clear; all that I said above should have been with referrence to Canon's Canoscan 9950F scanner, which I own and use. I did have the use of an Epson 4490 for a while, and found it to have a much inferior resolution to the 9950.

 

I can only think that Epson have been brainwashing me with subliminal messages!

[martian bachelor]

Pete - gotchya. Again thanks for the great quantitative info. Most useful.

 

I wish santa had brought me a 2 micron test slide... :>

 

I've made up a big hardcopy test card which I shoot with my actual lenses and film, but this tests the entire system and is clearly different from testing just the scanner. In another forum I made the point once that if one can get just 25-30 lines/mm, which is mediocre to poor resolution by modern standards, that the very hugeness of large format can still produce awesome large prints - the math works out to this being 7-10 million independent pixels.

 

And I agree, the film holding arrangement is less than ideal, but it actually works better (less sag) than I had first feared. Has the idea occured to you to turn the machine 90 degrees so the film hangs vertically? It's not easy to do because of the sleak, rounded-edge styling of the machine; and keeping the lid shut tight could be a problem. I have some Tech Pan 2415, which is of course on a very thin substrate, that I may eventually try to scan in a vertical configuration.