What is the state of the art for scanning 4x5 color transparency film?

Discussion in 'Digital Darkroom' started by kostak, May 16, 2022.

  1. After a long period of time, I've decided to scan in my 4x5 transparencies in hopes of having large prints made in the future (up to 40"x50").

    Now I am in Seattle, and in my area Panda Lab performs scans with an Imacon X5. Hasselblad Flextight X5 Scanner

    Judging by the OS supported listed, it is an older scanner. I'm curious as to what would be the latest. Or has digital pretty much killed R&D? I ask as I see Agx using the same scanner.
  2. digitaldog

    digitaldog Andrew Rodney

    I've owned (and trained others) on numerous models of Imacon's. It is a very good scanner and has many advantages. It's not a true drum scanner (and it's not a PMT scanner, even better) but the quality of the optics and the way it was designed produces very good sharp scans. Yes, it's kind of old and needs to be run on an older OS; not a big deal and to be expected these days. Yes, digital has pretty much killed the scanner market, certainly those high(er) end scanners like Imacon. There's simply no market for them. Short of a scan on something like a Tango or similar PMT drum scan, gel mounted (which some don't wish to subject their film to), you're not going to get much better quality.
  3. I would guess that the Imacon would do just fine--for only going to 40x50" prints I'd think much less would do.

    I've digitized hundreds of 4x5 negatives and Polaroid 4x5 prints on my old Canoscan 9000, but I'm only going up to 10 or 14" maximum print size.

    There are limits, you know, to the resolution and acuity of the originals depending on the lens, the film, and the gear holding the film in the camera. You don't gain anything by going to 'fictional' settings.

    You will also need to invest in multiple Terabyte storage-- the files at highest resolution will be truly enormous,
    Last edited: May 16, 2022
    kostak likes this.
  4. Not wishing to sidetrack the thread, nor be contentious, but what advantage does a Photo-multiplier have over a silicon 'blue' sensor these days Andrew? The light sensitivity, spectral response and response time of a silicon sensor can be made very similar nowadays with modern integrated low-noise amplifiers.

    As to resolution, that would appear to be strictly a function of the scanning spot size and hence dependant only on the light source and quality of the optics... assuming that the mechanical transport is equally, or more precise. In fact the more compact and defined sensitive area of a silicon photodiode might be advantageous in this respect.

    Just curious in what way a PMT adds anything to the rather outdated and mechanically dependant 'flying spot' design of a drum scanner.
    kostak likes this.
  5. digitaldog

    digitaldog Andrew Rodney

    Not all PMT and CCD scanners are created equally, indeed. I did specify a specific state of the art PMT drum I've got some experience along with a Scanmate and Howtek drums, gel mounted.

    Can CCD scanners produce the same quality scans as PMT scanners? They certainly can. In fact, a well-designed CCD scanner can actually surpass the quality of a poorly-designed PMT scanner.
    Drum Scanners: Perfecting the Film Photograph
    By drilling down into each pixel of a photograph, the scanner is able to reproduce the original photograph more precisely than CCD methods.

    Capturing Pixels CCD Versus PMT - Color Space - Tanguay Photo Mag
    The overwhelming majority of today's scanners share a common technology, illustrated in Figure 2.31- The film image is recorded onto a CCD (Charge Coupled Device) chip. Drum scanners, by contrast, rely on the older but optically superior method of PMT (Photo Multiplier Tube) capture
    In terms of comparative quality, a drum scanner run by an experienced operator will produce sharper images over a wider dynamic range and resolve noise-free shadow detail right down to the grain structure of the emulsion. For difficult-to-scan negatives destined for large output, a drum scan can be the best way to faithfully replicate all of the detail in the film.
    Last edited: May 19, 2022
    kostak likes this.
  6. Ah, but that's comparing old CCD technology with a Photo-multiplier. I was considering replacement of the PM tube with a photodiode (or 3, filtered RGB). Seems like an obvious update and does away with the need for a high voltage supply to drive the tube. Has nobody introduced such technology?

    And what does the phrase "drilling down into each pixel" actually mean? 'Cos pixels don't exist in the film, they're created during digitisation.
    kostak likes this.
  7. PMT scanners scan through each point in the film sequentially using the very center of the lens instead of scanning with a row detector (array of sensors). This gives them a slight edge in sharpness since the center is the sharpest part of just about any lens's image. I've looked through a Kodachrome 25 slide through a high-grade microscope and viewing through the microscope, there was sharp spider web which looks blurry through any CCD scanner I've tried. What's more the grain wasn't as apparent as it is in scans. I think quite a bit of image quality is lost by use of suboptimal scanning. Unfortunately, the best quality scanning technology was never inexpensive.
    Last edited: May 20, 2022
    kostak and digitaldog like this.
  8. digitaldog

    digitaldog Andrew Rodney

    Good question.
    "When someone says, That's a good question. You can be sure it's a lot better than the answer you're going to get." -Source Unknown
    kostak likes this.
  9. AFAIK, the lens in a drum scanner is on the light source side, and simply focuses a tiny spot of light on the film emulsion. The expanding cone of light is then picked up by the PM tube(s) on the other side of the film. So no actual imaging of the film is done at all in the conventional sense. And like any spot of light, its size is subject to aberration and diffraction effects, with diffraction being the main limiting factor.

    An interesting side phenomenon of the above arrangement is that if the sampling rate is made smaller than the spot size, it has much the same effect as a USM digital filter. I.e. No real detail is gained, but the apparent acuity of the image is increased.
    Last edited: May 20, 2022
    kostak likes this.
  10. digitaldog

    digitaldog Andrew Rodney

    Scanning Resolution | Professional Drum Scanning
    It should be remembered that even a 2,000dpi scan from a drum scanner will be sharper and have less noise than a 2,000dpi scan from an Epson, Imacon or equivalent non-drum scanners. This is because of the minimal use of optics in the scanning path (no long optical path mirrors) and the accuracy of construction of the scanner. It should be remembered that these drum scanners originally cost the price of a small house, and it shows in the engineering – we have nearly a ton of scanners in our studio!).
    And visual examples!

    And that lovely Tango, more visual examples:
    Drum Scanning - What exactly is it? - Michael Strickland Images
    The scanner actually scans in analog using RGB lights, collects that data with the use of old vacuum tubes, giant capacitors, resistors, etc., and then converts it to a digital file. Compared to an image scanned entirely electronically, the Tango produces a file that has far less issues with banding in areas of transitioning color and tonality.

    Sharpness of the film is at the grain level, therefore sharpness of a scan resides in the scanner's ability to focus on the actual grain of the film and most flatbed scanners don't have that much control. The ability to scan at the grain level ensures the photographer achieves every bit of detail out of his or her piece of film. Fortunately, the Heidelberg Tango focuses its lenses on the grain of each piece of film to ensure sharp grain in the final image.
    kostak likes this.
  11. Thank you everyone. This has been really educational. I appreciate everyone taking the time to contribute. I've been reading and rereading your comments to ensure I am grasping all the technical details/concepts/issues.
  12. Even in that case, the fact that only the center part of the microscope lens's image circle is used to focus the light on the film reduces aberrations and there is a smaller spot of light on the film. This creates the main difference in detail compared to scanners which have a row or several rows of detectors which rely on the off-axis part of the image circle to be used. The off axis parts suffer from aberrations to a greater degree than the center. In larger format scanners the difference is greater than in 35mm scanners which are a bit sharper (but in drum scanners there would be no difference between the formats provided the film fits in terms of degradation caused by the scanner). The other difference is that PMTs can have really low excess noise factors in the amplification cascade which can help with achieving good reproduction of the shadows in the slide.
    digitaldog likes this.
  13. But then so can a large area photodiode, and the shadow noise level is really just a matter of how bright the illumination can be made. Which is no problem for compact and highly collimated LED sources.
    I'm not making an argument for old CCD technology against even older photomultiplier tubes. Just wondering why drum scanner technology hasn't been dragged into the 21st century.
  14. digitaldog

    digitaldog Andrew Rodney

    For the same reason darkroom enlargers and 8x10 view cameras as two examples.
    Last edited: May 24, 2022
  15. Darkroom enlargers (especially for colour and VC) would greatly benefit from the addition of LED lighting. And I believe new LF cameras are being constructed using carbon-fibre and 3D printed parts.

    FWIW I'm in the process of replacing the hard-to-find, expensive and ridiculously hot 250 watt TH bulb in my Durst L1000 with a COB LED lamp.
    Last edited: May 25, 2022
  16. digitaldog

    digitaldog Andrew Rodney

    And horseshoes (and buggies) made from titanium would benefit from that alloy but your question was WHY such 21st century technology isn't being adopted and I ask you again: are you serious?
  17. OK. Now I get what you're saying, which is basically that the market for such advancements doesn't warrant the R&D effort - right?
    Or "If it ain't broke, don't fix it"?
    So the answer to the OP's question about the 'state of the art' of scanning 5"x4" film, is that it hasn't advanced much in the last 2 decades.
    Fair enough. But people still interested in digitising film are looking at alternatives to those obsolescent and/or expensive dedicated scanners. And maybe there are some decrepit or half-wrecked drum scanners out there that could benefit from some inventive renovation?
  18. digitaldog

    digitaldog Andrew Rodney

    I think you're getting it now Joe. :rolleyes: The OP summed it up well: digital pretty much killed R&D (for film scanners).
    Last edited: May 25, 2022
  19. There is a discussion somewhere else on whether scanning is, or isn't, photography.

    It seems that calling it not imaging is a related question.

    I suppose in the optics sense, of real or virtual image, it isn't one of those,
    but in the photography sense it is an image.

    The first time flying spot scanning was explained to me, over 50 years ago, it was
    imaging a CRT onto the film, and then through the film to a PMT. That allowed for
    generating a video (TV) image without electronic storage. That way, it is imaging
    but on the other side of the film.

    As for the PMT vs. silicon sensor, in the case of flying spot video, it needs to be
    done at the scan rate for the video signal. For storing a (usually digital) image,
    it can be done slower. That might matter.

Share This Page