Minolta Elite 5400 - Focus, GD, IR cleaning

Discussion in 'Digital Darkroom' started by og, Oct 17, 2003.

  1. og

    og

    After Exposure, I am now looking at Focus, Grain Dissolver and IR cleaning (ICE) in order to improve my technique with my scanner Elite 5400. FOCUS You should always use it. I let you check the results on the attachment (see 'Centre' - crop of a 5400dpi scan). This film is curved, and you will notice that with Auto-focus ON, the centre is very sharp whereas the corner is soft ('Auto-focus OFF' is sharp only in the corner - Its focus plan was here...). Please note that about 90% of the picture is like the center, and was really sharp with Auto-focus ON. It is probably very similar (or better) to what you may find in other scanners (Nikon LS4000...). So, if you have a curved film, you should carefully choose the focus point and check the results. If someone found a way to improve the handling of curved films on the Elite 5400, please let us know... For Low-Res scans, you may want to disable 'Auto-Focus on each Scan' and just keep Auto-focus on 1 Scan, in order to increase speed. I consider than Manual Focus is not usefull, except when 'Auto-Focus' can't find enouch contrast to focus properly, or if you want to de-focus precisely.
    006GHH-14910384.jpg
     
  2. og

    og

    GRAIN DISSOLVER The Grain Dissolver does it job : Grain is really less visible; it softens a bit the scan, but the details are still here. It also helps to reduce effects of scratches/dust... You get a better scan to work from, but, unfortunately, scanning time does increase a lot... GD should also reduce Pepper-Grain effect. IR CLEANING IR cleaning is a Must! You can see the results with the Minolta Software (GD+ICE) and with Vuescan (you can choose GD / 3 levels of IR cleaning). On this scan, ICE is very impressive! You don't see the dust anymore, and you can hardly guess where it was. Really Great! ICE does not soften the image. Vuescan's implementation of IR cleaning is efficient, but you can see the blurring effect afterwards. IR(light) does not soften the picture, but IR(medium) and IR(heavy) do really soften the whole image a lot. On another dusty scan, ICE was not able to remove completely a dust and left a white part, whereas Vuescan's IR(light) managed to correct it completely. Overall, I prefer Minolta+ICE over Vuescan's IR, because ICE corrections go almost unoticed. You may need to manually correct some heavy defects, but at least you don't have to work over what has already be done by the software. I will use regularly GD+ICE for my high-quality Scans. I may be interested by IR cleaning alone [Vuescan's IR(light) without GD], in order to increase speed for low-res Scans (indexing). Did you come with the same findings ? Olivier
    006GIw-14911184.jpg
     
  3. Olivier,

    Thanks for reviewing the 5400's features. I'm considering this scanner, and there's not a lot of reviews out there, it seems. While the brightness of the Minolta ICE samples was darker than the Vuescan, cranking up my monitor bright evened things out, and I'd agree, ICE is better than Vuescan cleaning, with the heavy scratch.

    It disturbs me that Minolta OEM software with ICE enabled does a more seamless job of removing defects than Vuescan, at any of it's 3 levels of cleaning. I much prefer Vuescan. The heavier Vuescan cleanings look to be increasingly indiscriminate in softening, without budging the big scratch remnants. I do like the fact that with Vuescan the cleaning can be used with or without GD, I believe this is not possible with Minolta OEM, the GD is always on if ICE is on. Perhaps Vuescan's use of the infrared cleaning channel will improve. I wonder if Minolta's "16 bit linear" output can have the ICE applied, and could it be used as Vuescan raw file? This could couple Minolta's superior ICE with Vuescan. Unfortunately, you would still be stuck with HAVING to use GD. Frustrating! For me, best would be if Vuescan could just improve it's infrared treatment. With new releases of Vuescan coming out about once a week, I hope this will be the case.

    Funny that implementing grain dissolver alone would increase scan times. I thought it was purely mechanical, making a frosted piece of glass intervene between light source and film, but maybe more is happening.

    Regarding focussing, Vuescan default focus point is 1/3 (or 2/3, no diff.) up vertical centerline of portrait oriented image, this seems a good compromise point when dealing with the dish (or dome) of film.
     
  4. og

    og

    Hi Mendel,

    I did the move : my old Scan-Dual II is on sell : ) FYI: some people are waiting for the New-York Expo Photo (October 30th) to check if Nikon proposes a new product.

    I have just tried what you proposed : scanning with Minolta in 16 bits linear with GD+ICE (and you are right: in Minolta, you CANNOT switch GD off when you use ICE... This is on my wishlist for the next release of the Minolta's Software). You can then use this file in Vuescan and process it as usual. I guess that it is very similar to the Raw file. Excellent results as well (actually, you can nearly put whatever image you want into Vuescan).

    The Grain Dissolver is a physical element only. Scan time increases because less light does reach the CCD (absorbtion/reflection/diffusion), thus increasing exposure time to maintain the same exposure target.

    After your interesting comment on how Vuescan does focusing (default), I made another test on the same picture with Vuescan: Centre is very sharp (exactly like Minolta's), and the corner is soft again (I found it slightly softer than Minolta's): I guess that Minolta used the same approach, but was able to take into account the film holder + lens designs to get a (very) slightly more usefull DOF.

    I get excellent results with Minolta's software (Exposition, Auto-focus, ICE...), as good as (or better than) what I can get with Vuescan. I like Vuescan's interface (kind of 'poweruser tool') and options, but at the end it is the Image file that matters. I am not a big fan of the 'Raw' files as well, because you already lost some important settings (exposure, focus...), and you are just post-processing the file. Vuescan keeps an edge for color correction (+ calibration of scanners, films...), but I prefer to re-interpret the image myself in an editing software.

    Olivier
     
  5. I've been wondering about Minolta's 16 bit linear for some time now, regarding it being similar to raw file. Minolta's (very brief) description seemed very sim. to Vuescan's description of raw file. Another piece of the puzzle worth knowing, thanks for the update.

    I've compared Vuescan raw files to Minolta 16 bit linear output, the brightness seems similar, the colour balance seems a little different. Vuescan raw files can be output during the scan, or after. If after, the infrared treatment will be incorporated, as well as grain reduction, sharpening etcetera. Also, while most Vuescan raw files are similar, if you have the media type set to colour negative, the percent of red, green and blue is shifted quite a bit.

    I have scan dual II, it's been mostly idle since Jan. last while I do dust and scratch correction directly on Vuescan raw files from a large backlog of b/w negative scans, where ICE is not effective. When I get through these (1800~) b/w images, I have a similar or larger backlog of color images to scan in, and with the majority ICE would be a great help, apart from some Kodachromes, which I've heard may or may not benefit from ICE.
     
  6. og

    og

    Small update on Focus: with GD on, corners get slightly sharper (difference in DOF / Focus ?). It is not a huge improvement, really, but I take it... (I would prefer an improved film holder, though).

    Olivier
     

Share This Page