Canon 9950F scanner review now up on Photo-i

[Roger Richards]

Hi all, like many of you I am watching carefully for reviews of the new Canon 9950F scanner. It seems that Photo-i has the beginning of one now up at http://www.photo-i.co.uk/Reviews/interactive/Scanners/Canon_9950F/page_1.htm (http://www.photo-i.co.uk/Reviews/interactive/Scanners/Canon_9950F/page_1.htm) . It seems to be a work in progress so far, only five pages, but the review has been positive so far. I was considering an Epson 4870 or Microtek 1800f but this new Canon might do the trick at a very reasonalble price.

[Bernard Languillier]

Thanks for the link, it seems that several pages were now added, including some first results of 35 mm negative scans.

IMHO, the comparison between the uncorrected images of the 9950 vs the Epson 4780 shows more sharpeness and dynamic range on the Epson.

Am I the only one to feel this way?

Besides, both scanners seem to be unable to resolve down to the grain level. These first results if confirmed would just confirm that such cheap flatbed scanners are still significantly behind a dedicated film scanner like a Nikon 9000 for instance.

The first test results of my new Epson F-3200 are unfortunately going in the same direction...

Best regards,
Bernard

[Ken Lee]

I don't see any of these comparison images at the link. I see the captions, but nothing more.

[Don Miller]

I can't figure out why he's scanning at 3200 dpi.

The problem with comparing tiny bits of film is that the different contrast and levels settings affect perceived sharpness, dmax and color. I'm convinced at this point that the best way to evaluate these tools is to make a subjective judgment based on large prints made from best practices for each scanner.

I will sell 20x24 landscapes from my 9950f, but I would not from my Epson 3200. When scanning Velvia 100f, the scans from my Epson look like they came from an Epson scanner. The scans from my Canon look like they come from Velvia 100f. My brain hurts from trying to rationalize the differences. I'm going to let the other half of my brain make an aesthetic judgment.

[Noshir Patel]

I think it's sad that manufacturers keep releasing scanners with higher resolution CCDs but without the optics to take advantage of them. It just wastes space and time (and money if you upgrade from the last generation in order to get the next higher resolution). It is clear that this scanner doesn't come even close to delivering 3200 dpi, as anyone who has seen a full resolution scan from a good dedicated film scanner can see.

Don, to answer your question... You scan at 3200 when doing a review because you want to test the manufacturer's claim (especially their wild claim that the flatbed can replace a dedicated film scanner for 35mm). True 3200 dpi should be able to make truly gargantuan prints from 4x5 (not just up to 20x24... maybe more like 64x80). Yes, there are some things that shouldn't be judged from a tiny crop (color balance being one). Other things become more obvious in a crop. Yours (in a different post), for instance show the Epson blowing out tiny highlights. Noise and sharpness are best judged in a crop. Of course if you are doing comparisons, you should try to have color and contrast adjusted as close as possible between two scanners.

[Brett Deacon]

Like many others, I was very hopeful that the Canon 9950F would be a significant improvement over my Epson 4870. Although Don's results are encouraging, the (incomplete) Photo-i review was a disappointment. From what I can tell, I agree with Bernard that if anything the Epson looked better than the Canon. Ever since the 9950F was announced, I haven't been able to help thinking that no matter how highy Canon touted this scanner, in the end you get what you pay for. I suppose it's possible that a $370 flatbed scanner could outperform a $1500 dedicated film scanner, but it sure isn't very likely.

[tim atherton]

"I can't figure out why he's scanning at 3200 dpi. "

"Don, to answer your question... You scan at 3200 when doing a review because you want to test the manufacturer's claim (especially their wild claim that the flatbed can replace a dedicated film scanner for 35mm). "

I think Don's question is - why test a 4800 dpi at 3200dpi...?

[tim atherton]

or rather:

I think Don's question is - why test a 4800 dpi scanner at 3200dpi...?

[Noshir Patel]

Ah, oops. Didn't realize it was 4800 dpi.

[Kirk Gittings]

Let me offer some observations at the other end of the quality spectrum, batch scanning 6x9 roll film on the 4870 vs. the 9950f for architectural magazine reproduction. This is what supports my family.

With sharpening, auto tone and exposure on in both:

At a 300 DPI 8.5 x 12.5 scan (what most magazines around here want unless it is a double truck spread). the 4870 scans are slightly sharper. This may be a function of film flatness, which is better in my "modified" 120 holders in the 4870. I am exploring this further.

The dust suppression is superior in the 4870 (Digital Ice), but adequate, even at the low setting, in the 9950f. The "remove dust and scratches" at the low setting also does not degrade scan sharpness.

Highlight detail is superior in the 9950f but mid-tone detail is superior in the 4870.

Film frame recognition, auto exposure and auto color balance is vastly superior in the 9950f.

The 9950f is considersbly faster than the 4870.

Unfortunately the 120 holders of the 9950f only hold 4 6x9s (I was hoping for 6-I will have to fashion my own holder for that), but that is still one more than the 4870.

All in all, for batch scanning commercial images, the 9950f has the edge in everything but slight sharpness. With sharpening in photoshop that difference is all but erased. Maybe the film holder can be modified to minimise that difference.

Just some thoughts from the other end of the spectrum.

[Roger Richards]

Kirk, thanks for the interesting comparison. Overall, which one of the two, the 4870 or 9950F, would you choose for 4x5 and medium format scanning? Thanks very much.

[Kirk Gittings]

Since my digital darkroom space is limited I would like to have just one scanner. Also even if you wanted to run two scanners at the same time your cpu would not handle it.

I scanned a 120 job for a client this afternoon and Fed Exed it by 6:00. It took me exactly half as much time with the 9950f as it would of with the 4870. A little additional sharpening seemed to get me where I wanted to be with sharpness. That is cost effective!

Tonight I am going to do a 4x5 b&w test and compare them.

Also I really agree with Don on his statement "the scans from my Epson look like they came from an Epson scanner. The scans from my Canon look like they come from Velvia 100f." Scans from the Canoscan hold the personality of the original film and not the scanner. In my case this afternoon it was NPS. I really prefered the "look" of the Canon scans over the Epson scans.

So I will let you know. All I can say right now-for production color work, I prefer the Canon.

[Kirk Gittings]

So here it is. A 4x5 Ilford FP5 negative scanned on the 4870 with both Silverfast AI and Epson Scan software and the 9950f with Canoscan. Each was done at 2400 DPI at 100% in 16 bit, about 188 MP files.

Why the scanning software would make such a difference I do not know. All of these were unsharpened with no auto-anything turned on.

The 4870 with Epsonscan is clearly the sharpest in the highlights with the best highlight separation and the worst shadow separation.

The 9950f with Canoscan is clearly not as sharp, but has the best shadow separation by far.

The 4870 with Silverfast is clearly not a sharp as the Epsonscan but Sharper than the 9950f. It also has the best midtone separation. It also appears to have the least visible grain.

Does anyone have any thoughts to illuminate these results?

The 9950f is also decidedly faster.

I'm on my way out of town for a shoot so I don't have time to post ant images.

[Roger Richards]

Kirk, thanks very much for testing and for your comments. If you have a chance in a few days I would love to see an image comparison.

[Craig Schroeder]

FWIW.... Vuescan's newest version, dated 11/13 (11.1.8 I think...I'm not where my scanner is as I write this) now seems to work with the transparency adapter in the 9950F. I just downloaded it late last night and didn't get a chance to really get a feel for it. Is there a method for batch-scanning? I couldn't quickly find the way to accomplish this.

[Kirk Gittings]

http://www.photo-i.co.uk/Reviews/interactive/Scanners/Canon_9950F/page_1.htm

The above reveiwer has officially given the boot to this scanner as it is his third one to malfunction.

Mine has worked flawlessly through about 100 scans varing from 50 to 230 mb, but I may be the lucky one.

I would not buy it yet if you were thinking about it.