View Full Version : Slightly O/T: Q about Nikon Coolscan 9000
OK, this weekend I've been trying to scan as fast as I can with my new Nikon Coolscan 9000 (it was a collaborative effort present to me for my b/d a few weeks ago, my gf and I). I thought I was going to gain time over my Minolta Dual Scan IV.
The image quality is superb; the hype is real. The speed is...disappointing. I feel I've taken three steps back (in that respect). I'm really thinking about exchanging it for the Coolscan 5000, but the film roll feeder seems to be discontinued.
Using VueScan Pro (latest version), in batch mode 16-bit mode, 4 samples, if I scan one strip of five frames, with the Coolscan 9000 at 4000 dpi it takes me a little over one hour, whereas with the Dual Scan IV with the same settings (but at 3200 dpi) it takes me just under 20 minutes.
I've found that if I only do one sample with the Coolscan, same settings as before, the scan time is drastically reduced to about 20-30 minutes (interestingly enough, the time is the same if I turn on the Infrared clean filter to "medium", or I leave turn it off --and I do see the difference in the image), but same settings with the Dual Scan IV, it's a whopping 5 minutes. This is a clear 4-5 time ratio between both scanners.
For those of you who use the Coolscan 9000 and the Coolscan 5000, what is your experience?
i've got the 5000 model and use Vuescan (lastest version 8.3.57 released a few days ago with a fix to speed up the 9000) i use multi pass set at 12 and the time is short no more than 2-3 mins i think ( i can double check this 2moro) I usually work with trix b/w Imput set as a 48 Bit RGB at 4000 dpi and Output saving as a 48 bit RGB tiff file. all scanned as single frames and no batch scanning so maybe that's why my time is short
i'm not using any ice or infra red cleaning (dosn't work on silver film) or any grain filter either. You can also make RAW files from it using Vuecan and save Tiffs and Jpegs at the same time.
hope this might be of help :)
What's your take on the grain produced by the Nikon?
i use multi pass set at 12 and the time is short no more than 2-3 mins
Holy s----! 12 passes in 2-3 minutes?!? OK, I must compose myself, I really want to call and exchange this but I should now then see what the fix in Vuescan is. You said it addresses a speed issue with the LS-9000? Hmmm...
Per the grain...it is interesting. The 9000, according to the literature, has an LED disperser-rod (read: a pseudo light diffuser). I did notice an incredible difference on how this grain shows up. The grain actually is grain, not sensor noise, which is what I mostly see with films of ISO 400 and up with my Dual Scan IV (sensor noise + film grain ... must use LaserSoft's noise reducing algorithm, which works like a charm, but I don't like using software to filter out anything from film --other than crud or scratches, of course)
I didn't batch scan with NikonScan because, well, for my needs, it really is lacking. It doesn't have (or if it does, it's hidden) a way to scan a raw image, like SilverFast or VueScan can. Not to be confused with the ability to save a NEF file.
Thank you very much Simon, for the info. As soon as I get home I shall test it out.
12 passes in 3 minutes. Wow.
I also forgot to ask: do you get to scan the whole area of a frame, or does it get cropped using the Coolscan 5000? Nobody seems to mention this.
If your camera has no spacing issues, properly mounting on the strip holder for the 9000 you get pretty the whole frame, but you don't really have the option to "crop" outside the frame.
Is it correct to assume your long time frame is for 35mm? Because, obviously, a 3-4 pass will take far longer on MF than 35mm....
Simon's multi-pass speed is quite amazing as 12x.. it takes 2-3 minutes to do a single pass with my CSV. Wow.
Yes, Jano, it's 35mm I've been talking about (haven't even tried MF yet until I get the 35mm scanning speed issue resolved), although it sounds like the times would be for 6x7, doesn't it?
And, yeah, 8x with my Dual Scan IV is I think closer to 4 minutes. If I get those scan times, I'd be set, no need for that full-frame digital rangefinder in a few years...
sorry for the late reply i'm on a different time zone in Thailand so now it's early morning - i'll double check my scan times and get back to you - re other questions and this is only my personal take on this i find the Nikon Scan software produces noticeably more grainy and contrasty scans compared to Vuescan's.
The difference equating in darkroom terms between a condensor enlarger head (more contrast) and a cold cathode (softer contrast) . Supposedly this is due to some clipping in the Nikon software but i'm not 100% sure on this. On the Vuescan you can alter these clipping values to ensure all the information is read.
If you use the motorised film strip holder that comes with the 5000 there is some cropping on all edges and it's harder with the Vuesan to find each frame edge unlike Nikon Scan software which makes very precise job of finding each frame they both read the film strip in a different way. Using the optional film strip F-H-3
holder (that allows for shorter lengths of film) i've got better results and i think it holds the film flatter and you can get closer to the edges of the film rebate but not full frame on all four sides. Filing the holder edges is not really possible as it's plastic!
check Ed Hamrick's site for the lastest version of Vuescan with listed changes
Film Scan times on my Coolscan 5000
1 pass 45 secs
12 pass 4mins 20 secs
12 pass scan does contain more detail the highlights
TX 400 scanned as 48 BIT RGB which interestingly opens as a 16 bit gray in CS2
and not RGB ?
I'll try scans in 16 BIT Gray and see if the scan times are faster?
actually the scan times seem the same when set as 16 bit Gray input and output.
Thanks, Simon. Looks like we're the only ones with a Coolscan on RF (?)
OK, I downloaded the latest version of VueScan (I did have the previous two back), but no improvement in the speed arena.
Right after I tried it, I remembered what Mark had said about "getting around" the banding issue found in the Coolscan 8000, when he said that instead of using the three line sensors, he only uses one. Now, that is exactly what Super CCD scanning mode is under Nikon Scan, which Nikon says will make scanning times significantly faster? It seems he's decided to play it safe with the 9000 and use the same strategy. SilverFast scans significantly faster (4x faster than VueScan), but their cheapest offering for the 9000 is $400+ :eek:
And guess what? I turn that off (I had it on) in Nikon Scan, and now, for a 4x multisample, 16-bit (48 RGB) with ICE on (Normal or Fine), the scan time is now between 3-4 minutes for each frame! Of course, that's a sneeze compared to your 12x multisample scan time of 4.25 minutes.
So now I'm back in the game! But the 35mm frame holders are disappointing. And I'm afraid to file them down because I think the software uses the shadow of the guides to find out where each frame is.
I tried to scan a strip of 120 film...now, what a disaster. I have to find out more how to do it right (if I'm doing something wrong), but neither Nikon Scan or VueScan can position the frames correctly, and setting it to "Positive" gives you wildly different results than "Negative". But that's out of your scope.
If I could find the SA-30 roll film adapter for the 5000, then I'd exchange this (I have until Thursday to make up my mind), but it seems that they're discontinued and I can't find them anywhere (except some shop in England, where they're twice as much what they would sell here).
Thanks again, Simon.
vBulletin® v3.6.8, Copyright ©2000-2013, Jelsoft Enterprises Ltd.