Scanner for 6x7 color negatives

[barry_r]

<p>Now that I've decided on a format and camera, I'm turning my attention to the scanner. I've read through numerous threads about this subject, to the point that I realized I was starting to read the same posts over and over, and still don't think I have decisive information. Since photography is a hobby for me and this whole medium format film thing is somewhat of an experiment, I am limited to reasonably priced scanners for maybe making up to 12x16 prints and since I have experience with the Epson V750 during my working days, I have limited my choice to the Epson V600 or V800, with a cost difference of about $450.<br>

I've tried to make sense out of the specs, but have also read that scanner specs are basically fairy tales, so I am asking for personal experience with these scanners or their similar predecessors.<br>

1. Is the alleged Dmax of the V600 (3.4) enough for color negatives, as far as being able to capture gradations in dark areas, especially considering that the final result might be a JPEG image, which is what the print labs I've used have asked for, and supposedly 256 grey levels are more than can be discerned visually.<br>

2. Does the V600 come with film profiles for the more common films, or will I have to purchase either Silverfast or VueScan to convert the negatives to positives? If Epson Scan does convert, would those other programs do a better job?<br>

3. Some websites purport to show the advantage of one scanner over another at 100% image magnification, but what small differences there were could easily be attributed to operator input. In general, given two prints side by side made from scans of a 6x7 color negative, how obvious would the difference, if any, be for these two scanners in a 12x16 print?<br>

Thank you, and if there are other relevant things to consider please let me know, but I'd appreciate if the discussion can be limited to these two scanners.</p>

<p> </p>

[thirteenthumbs]

<p>I have a V500 and V700. I do not buy the consensus that manufacturer specs are fairy tales. Peoples eyes are what determines many review opinions not high tech test techniques.<br>

There are currently V700, V750-M, and V800 in the Epson USA clearance center at reasonable prices. Any of these will will provide scans of your negatives more than adequate for your needs provided you know how to scan. <br>

I do not think Epsonscan has profiles for types of film, I work in B&W mostly and it does not. I turn off all auto stuff, save to Tiff at the highest resolution the Optical spec states then edit in PS and resize copies as needed.<br>

You probably won't see the difference between a 4800 dpi scan and a 6400 dpi scan unless you are scanning 110 film or similar. If Microtek was still marketing in the U.S. I would have a M1 or M2 in place of the Epson V700 which I picked up from the clearance center a year or so ago, light usage, no problems. It came with film holders and PSE 11.</p>

[peterbcarter]

<p>DMAX is much more important than most people think. The greater the amount of light coming from your scanning source and the the ability for the CCD not to saturate determines the DMAX of the scanner. The higher the DMAX means the more detail you can pull out of that film, thick or thin. It's not just about the shadows.</p>

<p>There are huge differences between the V600 and the V800. The first is micro-focus on the V800 (or V700). programs like Vuescan take advantage of this feature and compensate when the holder is not perfectly in plane. The second is the ability to support single pass over sampling. On the V600 you have to make 2 scans, which alignment errors will occur. This technique is used to increase scanner accuracy. I use 16 times over sample for my important scans.</p>

<p>For software, I tend to lean towards Vuescan. It can be as simple as you need it to be or give you the control over everything you need. I recommend the pro version, as the constant updates are invaluable. I've been burned by the people at Silverfast and would not recommend them for anything. They also lock their software into a specific scanner serial. Vuescan will use anything you have plugged in....long list of supported models.</p>

 

[peterbcarter]

<p>And I am not saying you can't get a good scan from a V600. You can. The V700/V800 are just superior scanners.</p>

[AlanKlein]

<p>I have a V600 that I use for 6x7 medium format film. From what I've seen from others, the V700 or V800 produces better scans. </p>

<p>Also note that color negative film like Portra scans better than slide film like Velvia although I regularly use Velvia because I like the colors and the fact slide film shows if the shot was on target just by looking at it. Slide film is harder to penetrate but I'm not obsessive about shadows. I like contrast. I use Epsonscan and usually adjust in post. I don't know why you would need film profiles. You can see my scans on FLickr, link below. Good luck on whatever you decide.</p>

[AlanKlein]

<p>Sorry, here's the link. https://www.flickr.com/photos/alanklein2000/albums</p>

[barry_r]

<p>Peter: Can you elaborate on "single pass over sampling." I couldn't find reference to this in the Epson V800 literature. Is this something Epson Scan can do or do you need VueScan or Silverfast? The version of Silverfast that comes with the V850 apparently does some kind of double scan to get first the highlights and then the shadows (like using HDR with a digital camera), but the version with the V800 doesn't do that. Is that what you are referring to?</p>

<p>Also, you say it's possible to get a good scan from the V600, but the V700/V800 are superior scanners. What I am trying to figure out is what visually discernible difference would result from using either one. If the V600 scan is of lesser quality such that the advantage of using film in the first place is compromised, then I would get the V800. If the differences are not visually discernible in the end product (which is usually a high-quality JPEG sent to a print lab), then the V600 might better suit my needs for this whole experiment in using medium format film.</p>

<p>Alan: Re. the film profiles, I'm a complete newbie in this area, but my understanding is that if you scan a color negative you need to "invert" it to make it a positive image. Apparently VueScan and Silverfast have profiles for each film type in order to do this automatically and presumably in an optimal way. I was wondering if EpsonScan had these (as the V600 doesn't come with either of those programs).</p>

 

[c_watson1]

<p><em>I've read through numerous threads about this subject, to the point that I realized I was starting to read the same posts over and over, and still don't think I have decisive information.</em><br>

<em> </em><br>

<em><br /></em>That's because there's been nothing new in scanner technology for years. Epson's servicing a residual market with its film scanners, i.e., "why bother?"</p>

<p>You might check out DSLR scanning.<br>

https://luminous-landscape.com/scannerless-digital-capture-and-processing-of-negative-film-photographs/</p>

 

[AlanKlein]

Epsonscan inverts negatives. Gets rid of the orange

mask on auto. There are always final adjustments on

past however. If you scan flat, you have to process

more in post. I don't know if other programs do a

better job as i never used them.

[rodeo_joe1]

<p>A colour negative's density range should be nowhere near to 3.4. If it is, then the negative has been grossly overdeveloped. I've only had one case where a colour negative film wasn't scannable, and that was due to some careless lab overdeveloping the film. That film wasn't conventionally printable either BTW.<br /> From then on I did my own C-41 processing, using Kodak's pre-exposed calibration strips for process control. No scanning issues after that!</p>

<p>Likewise with B&W negs. They should rarely exceed a Dmax of 2.2 by much of a margin. If they do then you're overdeveloping. The real issue with scanning negs is that there's too little contrast if anything, especially in the shadow region (low density areas). Digital discrimination isn't good there, and not enough "bits" are allocated to achieve a smooth tonal representation. So no, 256 levels of grey are <strong>not</strong> enough, and you should scan at 16 bits or whatever the highest depth the scanner will allow. Only convert to 8 bits after all inversion and tonal correction has been done. And not even then if you want an archival copy of the file.</p>

<p>But really, after all the investment of time (scanning is never a fast process) and money into cameras, scanners, film & processing (not cheap these days), and then ending up with a digital file? My conclusion was that a 1st generation digital capture from a high pixel-count DSLR was far superior to anything that MF film and 2nd generation scanning could achieve. Not without spending thousands on a drum scanner at least.</p>

<p>P.S. I found Silverfast software to be inflexible and non-intuitive, with mediocre results. ViewScan beats it hands down IMHO.</p>

[david_henderson]

<blockquote>

<p>and still don't think I have decisive information</p>

 

</blockquote>

<p>I think you said something very similar about the 67 cameras. Of course its unlikely that you're going to get a unanimous view on any forum worth reading. The people who contribute have different experiences and different views. The more threads there are and the more contributions- and Photo.net used to have a much bigger active base than it does now- the greater the opportunity for views to be at variance. Even within the thread you started there was a difference of opinion on the relative merits of 645 and 67 and how image quality can be measured. The OP needs to assess the credibility of the posters , the plausibility of their answers and the extent to which their experience/practice mirrors your own. If you really want a single vview you're likely to find it more easily from a book or article </p>

<p>Anyway. I have an Epson V700, the predecessor to the the V800 you're thinking about now. I use it for making scans for online use, for Blurb books where I print to about 10" sq from a 6x6 original, & the occasional print up to 12" across. Anything bigger than that and I'd prefer to use a scan made on a Coolscan or an Imacon. I use a Betterscanning 3rd party holder with adjustable height , and have followed the testing regime its supplier specifies to determine optimum height. I use AN glass to keep the film flat. I'm happy with the results I get for the purposes I've stated . And I think that despite that I prefer to print that size from film scanners rather than a flatbed, that a 16" x 12" from a quality 67 original is tenable once you've got a little way up the learning curve. <br>

As someone else mentioned the weakness you'll find is as much Dmax as resolution. So as well as using a tripod & cable release and making sure that you focus your new Bronica perfectly and take into account that depth of field is more of a challenge than with 35.mm/dslrs, you might like to avoid shots where the ability to see a long way into the shadows is vital. </p>

[peterbcarter]

<blockquote>

<p>Peter: Can you elaborate on "single pass over sampling." I couldn't find reference to this in the Epson V800 literature. Is this something Epson Scan can do or do you need VueScan or Silverfast? The version of Silverfast that comes with the V850 apparently does some kind of double scan to get first the highlights and then the shadows (like using HDR with a digital camera), but the version with the V800 doesn't do that. Is that what you are referring to?</p>

 

</blockquote>

<p>Multi pass scanning is a technique to combine several scans together to reduce scanning noise. The problem is most scanners have repositioning errors that limit it's effectiveness. This scanner (by whatever black magic) has the ability to take several samples while the scanning head is in the same position.</p>

<p>What Silverfast is doing is combining several scans of different exposures. It's closer to HDR than the above. The version (software) that comes with the V800 is just a lower version of the software. The hardware is the same. Physically the only difference between the V800 and the V850 is the latter comes with a wet mounting kit. That kit will help you get the most out of the scanner, IMO.</p>

<p> </p>

[wogears]

<p>+1 Rodeo. Chrome films require a DR of about 3.6, but negatives are a different story.</p>

[nick_sandin]

I recently went through a similar scanner search. I ended up with the v850 and I am very happy with it, stated simply. I'll

attach a scan from it below.<div></div>

[thirteenthumbs]

<p>The main difference between a V700 and V800 is the light source. V700 cold cathode fluorescent, V800 LED.</p>

[barry_r]

<p>"Of course its unlikely that you're going to get a unanimous view on any forum worth reading."<br>

You want answers?<br>

I want the truth.<br>

You can't handle the truth.<br>

David: It was worth a try. What I've decided is that I will shoot a roll of film along with my digital camera, trying to match the field of view, f/stop, effective focal length, etc. as closely as possible. I'll then have the lab do the scans and compare to the camera images. I know the scanning will be somewhat of a "black box" but it will give me an idea of what direction to take. Then, if I decide to continue down that road, I've read enough reviews to convince myself it would be the V800.</p>

 

[ondebanks]

<blockquote>

<p>Multi pass scanning is a technique to combine several scans together to reduce scanning noise. The problem is most scanners have repositioning errors that limit it's effectiveness. This scanner (by whatever black magic) has the ability to take several samples while the scanning head is in the same position.</p>

</blockquote>

<p>The 'black magic' of <em>multi-sampling</em> is simply that the scan CCD dwells on each position for a while, taking multiple independent exposures, before being stepped forwards to the next line to scan. At each pixel, the scan software averages the data from this stack of exposures, reducing random noise in the CCD (readout noise) and in the nature of measuring light itself (Poisson noise).</p>

<p>It is a smarter technique than <em>multi-pass scanning</em>, where the scan CCD is passed over the entire image repeatedly. They should in theory generate identical results for an identical number of samples/passes, but when revisiting the same piece of film, the stepper motors never quite line up the scan CCD with the sub-pixel precision required.</p>

<p>Vuescan supports both modes in several scanners, such as my old Epson 4990 Photo. I can't remember if the EpsonScan software supplied with the 4990 can do multi-pass scanning, but it definitely cannot do multi-sampling.</p>

[AlanKlein]

Multi sampling and multi scanning give less sharp

details because of combining samples so i understand

from others who have tried them. There are others

who claim that these techniques bring out the

shadows better but I have yet to see a posted example

that proves it. The dMax of the machine is set by the

manufacturer to give the max light to penetrate on the

first scan. Subsequent scans or samplings won't

penetrate deeper. The light isn't a shovel. These are

just selling points contrived by Vuescan.

[peterbcarter]

<p>DMAX and Dynamic range are 2 different fishes. DMAX refers to how much light penetration you can achieve, but in now way dictates the dynamic range of the sensor. You can easily have a strip of film with high dynamic range (like a slide) where you would normally only be able to scan for the highlights or the shadows. Multi-pass scanning does multiple passes with different exposures and combine them with higher dynamic range.</p>

<p>Vuescan, as scanning software goes, is a late comer. They didn't invent much, except leave as much user control as possible. That, they wrote the book on.</p>

[AlanKlein]

<blockquote>

<p>Multi-pass scanning does multiple passes with different exposures and combine them with higher dynamic range.</p>

</blockquote>

<p> The full range of a negative or positive film image cannot exceed 0-255 unlike when you initially shoot a picture where it can exceed the range of the digital sensor or film. When shooting, you are concerned about DR and the sensor or film DR range is limited. That's when graduated neutrsl density filters or HDR with multiple shots are applicable.</p>

<p>But nothing can be done once the film is exposed. It goes from black to clear. 0-255. That's it. You cannot be beyond that range. Now what happens (at least on my v600) is that the results are squished into a shallower range. You have to adjust in post or during the scan to spread out the range to 0-255 to get the correct range and contrast. But the point is, there is nothing you can do to extend the DR range through the scan process. </p>

<p>Additionally, the maximum light output is max output of the scanner. You can't increase it. Even if you reduced the light, the clear portion would still be at max. </p>

[AlanKlein]

<p>Slide film has 5 stops and negative color film has 7 stops. All these highlights and shadow areas are picked up with one scan because as I said the film already reduced the range of the actual light in the field to the 5 or 7 stops. If you look at a histogram of the scan, you will see with a single scan that all of the black to clear will fall in the range of the scan sensor. That's unlike when you look at a shot through the camera sensor. Then, the histogram of the light reading does exceed the range of the camera sensor. But once the scene is recorded, the film is black to clear - that's it. 0-255 equivalent. The 15 stops or more in actual light photographed would be condensed to 5 or 7 stops.</p>

[rodeo_joe1]

<blockquote>

<p>"The full range of a negative or positive film image cannot exceed 0-255"</p>

</blockquote>

<p>Sorry, but that's just a nonsense. True a JPEG can only store 8 bits per channel colour depth, but any decent scanner will have at least a 12 bit A/D converter, allowing 4095 levels (+ pure black) greyscale to be scanned and stored (as a 16 bit TIFF). Most modern scanners will have a 14 or 16 bit A/D, allowing 16,383 or 65,535 levels. That's without any gamma curve being applied. The sRGB or Adobe RGB gamma of 2.2 theoretically allows far more than 12, 14 or 16 stops of dynamic range to be encompassed and stored.<br /> Even the 8 bits/channel of an sRGB JPEG can easily encompass a dynamic range of 10 stops, with some posterisation of shadows admittedly. Adobe RGB's more sensible implementation of gamma can store around 12 stops dynamic range within an 8bit/channel file.</p>

<blockquote>

<p>"...the film already reduced the range of the actual light in the field to the 5 or 7 stops."</p>

</blockquote>

<p>No, it didn't. The gamma (contrast curve) of slide film actually increases its Dmax-Dmin range to more than the Subject Brightness Range. The two things are completely different. And negative film can encompass more like a 10 stop SBR, but reduces it to around 7 stops density range on film.</p>

<blockquote>

<p>"Vuescan, as scanning software goes, is a late comer."</p>

</blockquote>

<p>I remember using Vuescan at least 10 years ago, and it wasn't new then. Ed Hamrick has been developing that software for a good part of his life I'd guess, and would have a lot to say at being called a "late comer". From my correspondence with him he's one of the most knowledgeable people about scanning and scanner hardware on the planet.</p>

<p> </p>

[AlanKlein]

<p>I wasn't referring to the A/D converter. I was using 0-255 as a general histogram reference scale you see on the histogram of the scanner. The film must range between 0-255. It can't be below 0 or higher than 255. The range is from clear to black on the celluloid. One scan can pick up the full range. More scans won't help what you capture in D/R or dMax.</p>

[rodeo_joe1]

<p>Alan, the histogram scale of 0-255 is simply an arbitrary scale based on an outdated 8 bit/channel digital capture. Of course it can't show anything whiter than white or blacker than zero luminance by its very nature.</p>

<p>"The film must range between 0-255" - A non-sequiter argument based on insufficient knowledge of how a digital scan is processed after capture and before being represented on screen or in print.</p>

<p>Properly processed slide film has a density range of between 3.4D and 3.6D, with 0.3D approximately representing one stop of "dynamic range". Meaning that most slide films have a viewed brightness range in the region of 11.5 to 12 stops, and this is a linear bit range of 12 bits or 4095:1. So we need a minimum A/D bit depth of 12 to attempt to capture the full brightness range of reversal film. Way more than 255:1.</p>

<p>Once captured in a linear fashion by the A/D conversion, a gamma or tone curve is applied such that the visual representation of the digital image on a computer monitor looks close to the original slide, even though most monitors can realistically only show around a 300 to 500:1 brightness range. (Claims of TVs and monitors having a contrast ratio of several thousand to one are exaggerated or in plain language "lies".)</p>

<p>Anyhow, leaving that aside. A gamma curve is a way of squeezing a greater dynamic range into a bit space that would otherwise not accommodate it were a linear tone curve employed. The most commonly used gamma is a value of 2.2, meaning that the fractional representation of brightness is raised to the power of 1/2.2. So a brightness level of 0.75 (three quarters fully white) becomes a level of 0.877; 0.5 becomes 0.73 and so on. Those fractional brightness values are then multiplied by 255 - or whatever bit depth is being used - and the new digital value substituted in the tone curve. If you calculate all the gamma-corrected values down to where there's only a 1 bit level difference, you'll find that a gamma of 2.2 can accommodate, or represent, around 12 stops of brightness in a single 8 bit digital byte. Quart into pint pot time!</p>

<p>"One scan can pick up the full range. More scans won't help what you capture in D/R or dMax."<br>

Not true either. If the exposure - brightness of scanning light source or acquisition time - is increased between consecutive scans, then obviously a greater density can be penetrated and hence a greater dynamic range can be captured.</p>

 

[barry_r]

<p>I was pretty much set on eventually getting the V800 over the V600 but started to read the numerous web postings about using a DSLR instead, as some people have mentioned in this thread. It seems to be unanimously favored over using a scanner that's in my price range, at least when used with appropriate software to process the image capture. I don't think I found a single website where the author concluded the V600/V800-type scanner was a better alternative. As I already have a decent macro lens, DSLR and flash unit, I think I will try to assemble a setup to do that.</p>