Re: LS30 and jaggies was RE:Hello
photoscientia [EMAIL PROTECTED] wrote: Their fix was ingenious. More a stroke of inspiration than a logical engineering solution. They fire a reverse pulse of very short duration into the motor at the end of every step. This acts as electronic damping, and it's very controllable. Maybe Ed could try it? But presumably this would be a firmware level thing, not SCSI level commands. Rob The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: LS30 and jaggies was RE:Hello
ILyons [EMAIL PROTECTED] wrote: No, you need a Mac, which doesn't suffer the problem, even if CMS in On. Or you need for Nikon to use the same scsi timing as the Mac and then the PC wouldn't suffer the problem when CMS is turned ON. [snip] The problem has Sweet FA to do with resonance, if it was resonance Nikon could fix it with ease. I don't see what possible difference there could be between SCSI timing on a Mac and SCSI timing on a PC. SCSI is SCSI. If the Mac coders sent different commands to the scanner, or sent the same commands with different timings, that is the answer. It has NOTHING to do with platform. Nikon have made no public response to the issue that I'm aware of, so I've seen no evidence that they've attempted to fix it - but on the other hand I've heard at least one report that upgrading to the latest Nikonscan resolved the issue. Personally I don't really care whether it's resonance, timing, or anything else - but I *do* know that the commands Vuescan sends to the scanner resolve the problem. If Ed can do it, Nikon ought to be able to. Rob The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: LS30 and jaggies was RE:Hello
Roman Kielich® [EMAIL PROTECTED] wrote: 64K blocks that are causing the resonance that results in jaggies. My guess in your case was that switching on colour mangement increased the computer's CPU overhead to the point that the actual scanning process was slowed down. if it is CPU problem, then what is required CPU speed to work flawlessly? I don't know. I don't have the means to find out. In fact, I suspect that whether you have DMA enabled or not is probably a much more critical issue as the hard drive can steal 80% of the CPU cycles otherwise. My Celeron 300A and 256 MB of RAM seems to be sufficient, not to mention if overclocked. Meaning you have never experienced jaggies? The RAM would help for caching. What hard drives do you have running in what DMA mode? Do we need a solid base for a scanner to prevent the resonance? I don't think it makes any difference. Whether the scanner is vertical or horizontal doesn't seem to matter either. Rob The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: Help- 48 bit vs 24 bit RGB
Adjusting levels or curves does not lose data. What you are doing is re-mapping the data. When you decimate (reduce in size) you lose data, but changing a value of 187 to 192 does not lose data. The 'loss' occurs in rounding errors. Eg 187 becomes 192.4567, which gets rounded to 192 in an 8 bit scheme. Then when you do the reverse calculation 192 becomes 186.45678 = 186. By the time all bin values have been through this process, and a few iterations, you end up with datapoints which are some way different from the original values. And combed histograms are the result. Regards Tony Sleep http://www.halftone.co.uk - Online portfolio exhibit; + film scanner info comparisons The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: LS30 and jaggies was RE:Hello
Personally I don't really care whether it's resonance, timing, or anything else - but I *do* know that the commands Vuescan sends to the scanner resolve the problem. From what Ed has said previously, NS reads data in 64k chunks, and he modified later versions of Vuescan to instead read one scan line at a time which introduces a slight delay between each line. Regards Tony Sleep http://www.halftone.co.uk - Online portfolio exhibit; + film scanner info comparisons The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: Resolution revisited, or scanner resolves 2 microns!
Rob, I'm in Bucks, UK and have an LS30. If Pete wants to mail me off list, we may be able to arrange this. Regards Bob Armstrong On Saturday, October 28, 2000 Rob wrote: photoscientia wrote: Any chance the same slide could be scanned on an LS30 and/or a SS400? I don't have access ..to those .., but if anyone's interested, I'll lend them the mask. I'd love to .. but the postal trip to Australia and back would be awfully long. Anyone in the UK with an LS30 who wants to try? Rob The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Kodak RFS3600 Film Scanner
Any reviews on the Kodak RFS3600 film scanner? thanks, Ron The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: LS30 and jaggies was RE:Hello
Ian, I can see you're upset about this, but I think you're letting your anger get in the way of logical analysis. I'm not in the least upset with the "problem". I don't have it anymore :-) Users of Ed's more recent version of VueScan don't have it any more because he uses the same buffer size in the PC version as Nikon use in the Mac version of NikonScan. SilverFast doesn't have the problem because LaserSoft use the same buffer size as the Mac version of NikonScan. This change in buffer capacity effectively changes the timing of data being sent. the CMS probably adds CPU overhead in processing the data. Which would in theory mean that a low MHz would take a bigger hit than one of say 1000Mhz. Sorry, but the CPU overhead theory dies a death as soon as someone with a 200 Pentium using ICE says they don't have the problem and someone with the faster CPU does. Nevertheless, I was of the understanding that Scsi had the advantage of not (or only very slightly) impacting on CPU overhead. If the problem was only to do with Nikon CMS, it would NEVER have occurred with Vuescan. But it did until Ed fixed VueScan :-) I seem to recall a lot of folk singing praises for doing so. I also recall that he spent some time explaining the nature of the fix he applied. Ian Lyons http://www.ian.lyons.btinternet.co.uk The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Vuescan and color theory (hello again!)
Hello list, Greetings again. After absence from March I decided to re-join and I'm happy to let you know that reason for that is an FS2710 I recently acquired . I downloaded latest Vuescan and have immediately put it to my common test: it's an underexposed (-1EV, someone forgot to push the roll) slide of high contrast scene (white aircraft front-lit against dark sky and dark green mountains) to see how FS2710 Vuescan will cope with it as CanoCraft failed this test - have to admit that it works fine with good originals, both negs and slides. I was amazed with the improvement over Canon original software, and I set the go to match scan of same slide done on LS-2000 some time ago that eventually finished on the cover of colour magazine. OK, did not do it, but came quite near, far nearer than with CanoCraft. However, I have to admit that I got used to levels sliders and histograms in PShop and I really miss such both in VueScan and Canon software. So my basic questions are (Ed, anyone else?): any chance that such things will be available at future VueScans and other is until then, Can anyone point out good resource (preferably on the net) on colour mambo-jambo i.e.. black/white points and all other values one inputs in VueScan. OK, all for now, Hope to hear from you, Best regards, Igor Salinger The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: Help- 48 bit vs 24 bit RGB
WALTER writes ... The SprintScan 4000 allows me to select "raw color positive" (or negative) and scan at 4000 dpi. I can save that scan as a TIFF file at 12 bit with no gamma or other correction, or, I can make gamma corrections in the scanner software and save the file at 8 bit. Is it better practice to save it uncorrected at the higher bit depth and make all changes in PS or to make the gammma corrections in the scanner software and save as an 8 bit file? ... Either choice would depend on the "quality" of the software ... the tools provided, their accuracy, and how easy they are to use. That is, if the scanner software and post-scan software (e.g., Photoshop) were equal in quality, then there'd be no difference ... and you could make all adjustments with the scanner software, scan to a 24bit RGB file, and send it straight to the printer (... and save your self quite a bit of hardrive space ...). However, scanner software is hardly ever as good as softwares like Photoshop, and if PS is available to you, then you are probably better off with its high-bit tools and applying adjustments to 12bit scans (... but those 100Mb files!? ...). The highbits method will also put your CPU and RAM and harddrive I/O to work ... better be ready. shAf :o) The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: LS30 and jaggies was RE:Hello
ILyons [EMAIL PROTECTED] wrote: of say 1000Mhz. Sorry, but the CPU overhead theory dies a death as soon as someone with a 200 Pentium using ICE says they don't have the problem and someone with the faster CPU does. That depends on what else might be occupying the CPU - as in my case, the IDE hard drive. Nevertheless, I was of the understanding that Scsi had the advantage of not (or only very slightly) impacting on CPU overhead. Correct, but that's not my point. Colour management is not a function (AFAIK) of the scanner hardware itself. It's a software function. Anything which the Nikonscan software has to do to process the data between getting 64K chunks off the SCSI bus potentially could result in handshaking faults, and introduce pauses in the physical scanning process. From what Ed has said, the scanning needs to happen at a certain speed or problems will occur. In the case of my computer, the main thing causing a headache is the fact that I can't enable DMA on the hard drive, and that steals clock cycles from Nikonscan - possibly 80% of them. If the problem was only to do with Nikon CMS, it would NEVER have occurred with Vuescan. But it did until Ed fixed VueScan :-) I seem to recall a lot of folk singing praises for doing so. I also recall that he spent some time explaining the nature of the fix he applied. Um, so you're agreeing with me here in this respect? Rob The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: Resolution revisited, or scanner resolves 2 microns!
Bob Armstrong [EMAIL PROTECTED] wrote: I'm in Bucks, UK and have an LS30. If Pete wants to mail me off list, we may be able to arrange this. Excellent, Bob! I think you'd need to email Pete with your snail address. If possible, I'd like a raw scan of a section which shows the finest resolution - a PNG no more than 500KB so I can add it to the material about the Nikon scanners on my site. But either way, let us know how it goes! Rob The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: Vuescan and color theory (hello again!)
Salinger Igor writes ... ... I have to admit that I got used to levels sliders and histograms in PShop and I really miss such both in VueScan and Canon software. So my basic questions are (Ed, anyone else?): any chance that such things will be available at future VueScans and other is until then, Paraphrasing previous statements by Ed ... "not likely". Ed's philosophy (sofar) has been to deliver an accurate scan with a minimum set of very basic settings ... afterall, he cannot offer better tools before the scan, and in fact, he creates a "raw" scan first and applies software "cropping" tools to the "raw" scan. Ed would tell you, if you're more familiar and comfortable with the PS toools, then use them instaed of his cropping tools ... "highbit software tools" are "highbit software tools" ... they differ only in their ease of use. Can anyone point out good resource (preferably on the net) on colour mambo-jambo i.e.. black/white points and all other values one inputs in VueScan. I don't believe these properties for an image are defined any differently for Vuescan than they are in Photoshop's help files ... 'cept Ed's setting for whitepoint and blackpoint might be percent, while a similar PS tool would show you the histogram and allow you to choose some setting a bit more intuitively. I would simply play with the VS settings to get a feel for what is happening. First, have Vuescan create an accurate "raw" scan (use VS defaults and your best example of exposure and subject, it should include true white and black). Now, turn off "create raw scan file", either scan "memory" or change the "device" to 'disk' and point VS at the raw file you created. For experimentation, have VS write 1/8 size JPEGs as crop files ... and experiment with the crop settings provided by Vuescan. When the resulting image is just right, then have VS write a fullsize TIFF file. Granted, Vuescan is not easy to experiment with, ... but then Ed has kept the settings simple and to a minimum. Good luck ... shAf :o) The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
UNSUBSCRIBE FILMSCANNERS
UNSUBSCRIBE FILMSCANNERS The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Re: LS30 and jaggies was RE:Hello
praises for doing so. I also recall that he spent some time explaining the nature of the fix he applied. Um, so you're agreeing with me here in this respect? Nope, I'm saying that Ed apparently fixed the problem on the PC by effectively adjusting the timing for the data, nothing more. Ian Lyons http://www.ian.lyons.btinternet.co.uk The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
Density increase w/SilverFast and SS4000?
I seem to remember somewhere, on this or another list, that there was an increase of the Dmax from 3.4 to 3.? when using the multi-scanning capability with the Silverfast software package available with the SS4000. It would seem to me that one could average out noise this way, but not increase the density range. A lack of noise would be desireable, although truly better definition of shadow details would seem to be a hardware function, in my thinking anyway. I am wanting only to scan 35mm and am trying to decide if I should go for the SS4000 or wait for the SS120 with an advertised Dmax of 3.6 over the SS4000's 3.4. I realize that the 120 will be for medium format as well, though the Dmax increase may make it worth the wait and expense. I'm currently using a SS35+ and looking for an upgrade. Any comments from SS4000/Silverfast users out there? Thanks, Jake The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
RE: Help- 48 bit vs 24 bit RGB
Is it better practice to save it uncorrected at the higher bit depth and make all changes in PS or to make the gammma corrections in the scanner software and save as an 8 bit file? It depends on which you prefer. Both methods should give you equivalent results, since you are applying the gamma correction, in both cases, to the 12 bit data, not to the 8 bit data. The results should be identical only if the scanner software does it's manipulation on the 12 bit data. Not all software does. For instance, the Minolta Elite software allows 8 or 16 bit data output. However, the curves, levels tools etc in the software only seem to work on 8 bits, even if 16 bit output has been selected. As soon as any significant adjustment has been made, the result is a 16 bit output file with gaps in the histogram (and ultimately posterisation)! Outputting the file with no adjustment and then making the same adjustments in Photoshop gives a much smoother histogram and colour. Unfortunately, unprocessed can mean different things, as the raw data may or may not be subject to the setpoints. It depends on the scanner and scanner software. Indeed. The Elite software allows 16 bit output (which already has gamma correction and colour space conversions carried out) so may only need the black/white points set and levels/curves tweaked in PS to get to the end result. It also allows 16 bit linear output, which is gamma 1 in the scanners own colour space. Even this isn't an absolute. Ed Hamrick mailed me off-list on the differences between the Minolta software and Vuescan: "I think I found the difference between the Minolta software and VueScan's controlling of the scanner. The Minolta software exposes the CCD colors with RGB times of 1.00:1.05:1.07. VueScan uses RGB exposure times of 1.70:1.62:1.00. This uses more of the dynamic range of the CCD for most films, but it significantly changes the colors that are captured." And of course, just to make it even more confusing, all the manufacturers do different things! The bottom line is don't take anything for granted :-) Al Bond The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
UNSUBSCRIBE FILMSCANNERS
_ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. Share information about yourself, create your own public profile at http://profiles.msn.com. The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
RE: Help- 48 bit vs 24 bit RGB
Walt, When you select 12 bit "raw" scan you may also find it useful to check the box below to embed the scaer profile. This embeds the profile for the scanner and gets saves a lot of work you would have to do without it. David -Original Message- From: WALTER MESS [mailto:[EMAIL PROTECTED]] Sent: Saturday, October 28, 2000 12:03 PM To: [EMAIL PROTECTED] Subject: Re: Help- 48 bit vs 24 bit RGB Austin, this and your earlier posting on this subject are most helpful but I still am a little confused about the term "raw scan". The SprintScan 4000 allows me to select "raw color positive" (or negative) and scan at 4000 dpi. I can save that scan as a TIFF file at 12 bit with no gamma or other correction, or, I can make gamma corrections in the scanner software and save the file at 8 bit. Is it better practice to save it uncorrected at the higher bit depth and make all changes in PS or to make the gammma corrections in the scanner software and save as an 8 bit file? Thanks Walt on 10/27/00 1:32 PM, Austin Franklin at [EMAIL PROTECTED] wrote: In some cases true, but change the gamma of an 8 bit file and then examine the histogram again and you will see what I mean. I know exactly what you mean. That's not the way I would recommend doing gamma, since it will cause this problem. Either apply your gamma curve in the scanner 'software' (which may actually do it in the hardware), or scan at full res and then (in PS) apply your setpoints and gamma and then convert to 8 bit for printing. Applying gamma to an 8 bit data file and getting an 8 bit result is sure to replicate codes (values of the data), but won't happen to a 16 bit (or 12 bit) source file, providing your scan provides a range of at least 256 codes. The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest. The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest. The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.
RE: Density increase w/SilverFast and SS4000?
the pass-after-pass registration not being perfect. I have found that to be true on my SS4k...but not only is the overall image registration off, the line to line registration is off just enough to make it virtually unusable. It works great for a single pass though... The Leafscan 45 does three passes for color, and I find absolutely no issue with the registration... But that was a $16k scanner, as opposed to the $1350 SS4k... The filmscanners mailing list is hosted by http://www.halftone.co.uk To resign, mailto:[EMAIL PROTECTED] with UNSUBSCRIBE FILMSCANNERS in the title, or UNSUBSCRIBE FILMSCANNERS_DIGEST if you are reading the Digest.