Thank you.
Many Japanese people shared your files 
2 Likes
The Kamakura Daibutsu is simply amazing! I totally understand why, after the tidal wave, the people did not want to rebuild the temple around it. The Tōdai-ji Daibutsu in Nara is larger, but the Kamakura Daibutsu is open to the light, the breeze, the open sky. And openness is key to kenshō.
Sorry, I wandered pretty far from RawTherapee there.
1 Like
Hello Franz, as you have noticed, cleaning negatives and slides before scanning is primordial ! Take care with “aggresive cleaning” though, you can damage the emulsion, so the image.
I always store my negatives in those semi-transparant negative sleeves (if that’s the right word) and my slides in slide boxes (those that go into a projector) or in dedicated plastic slide holders.
When I take some of them to scan, I just use a simple dust blower without touching the surface of the negatives or slides. Idem for the scan surface on the scanner. Works well.
I promise to be gentle, Paul! I was a bit concerned about the aggressiveness of the cleaning I did on this slide, but I have a very similar slide, so I took a chance with this one. I have both compressed air and a classic dust blower like yours and I plan to use only those unless the film has large homogenous swaths and large amounts of dust.
This experience has convinced me to return my recently purchased Plustek film scanner. I have, instead, purchased a macro lens and a digitizing adapter that fits on an extension tube for my lens and comes with film and slide holders and its own LED and diffuser. This method should work quickly and accurately. It’s true that it will require color inversion on negatives to evaluate whether to do much work on them, but RT has that tool built in. I’ll be working with true Raw files, not TIFFs and it avoids scanning negatives quickly to evaluate them, then having to rescan my selects at high quality to preserve them.
I realize no method will be fast or easy, but this will work. Plus it gives me the cool macro lens I’ve wanted for ages!
Same procedure applies : clean the negatives and slides before scanning.
Plus: don’t forget to clean your sensor and objective ! 
Gosh, I don’t even recall the last time I cleaned my sensor. Oh, the shame! I cleaned my old Nikon D40 sensor religiously, and I still have the tools for doing so with my Olympus. The Nikon sensor got dirty quite easily, the Oly not so much. I reckon the auto-cleaning function has improved. But thank you for reminding me to do it manually; it is time, before my lens is delivered.
I have to admit, I’m kind of excited.
@buddhaauthor
I do not understand your problems using iSRD with SilverFast. Here is an example how well it works for E6 emulsions:
Upper left is the original, upper right the iRSD corrected version. Below you see the IR-channel.
So it works nicely. A problem is Kodachrome K-14 emulsion. Although SF does a pretty good job there also. VueScan has a different approach and it in my opinion fails for K-14, since it corrects all over the place.
If SF does not seem to work, there is probably an offset between IR and RGB image. Use the manual alignment then to correct this.
Hermann-Josef
Hermann, if that is your IR result, I totally understand your perplexity with me. I just have not come even close to your lovely mage. I assume the problem is my operator error. Still, on balance I think I prefer the DSLR mode for digitization. Maybe I’ll come back around to scanners, but I’m going to give the DSLR mode a chance.
Dry cleaning a negative or a slide which has no glass cover with a cloth is bound to cause scratches on the film emulsion. If the dust on the surface contains silica paricles (i.e. tiny sand particles), rubbing them along the surface will almost certainly give a scratch. I have found a marter hair brush sold for water color painting to be useful for removing dust particles which stick to the surface of a film negative. These brushes are very soft and you apply much less force to the emulsion surface compared to wiping with a cloth.
Please also keep in mind that brushing or wiping the emulsion surface will create electostatic charge on the surface which attracts further dust. Therefore, you should make sure to have as litte dust around as possible when you clean negatives or slides mechanically:
- Clean the floor by vacuuming or wet wiping to avoid stirring up dust
- Wipe the desk you work on wet to remove any dust
- Wear clothes that do not give off dust. I use a nylon windbreaker for that purpose
I use a Reflecta ProScan 7200 with SilverFast for negative scanning at 3600 dpi. I check for dust particles by looking at the negatives in the film holder at an angle and remove all visible dust particles wuth the marter hair brush. Then I run a pre scan, chekcm for any visible dust specks and repeat the cleaning until the pre scan is clean or the dust specks cannot be removed. The iSRD function will then remove the remaining small dust specks and scratches without loosing to much detail. As Jossie pointed out, it sometimes takes manual alignment to get good correction.
I propose you check whether the IR image obtained with your scanner actually shows the dust specks you see in the RGB image. If the dust specks do not show up in the IR image, the problem is most likely not an operator error but an erroneous setting of the scanner or the software.
Let me elaborate on this somewhat more. Some scanners make two passes, one for the RGB and one for the IR scan. If the mechanics is not perfect (which seems to be the case for some scanners) there will be a shift between the RGB- and the IR-image, since the stepper motor does not bring the scan head to the exact same position for the second scan. iSRD tries to find this shift, but quite often it fails dramatically. Then the defects found in the IR are offset from the real defects in RGB. There is a correction done, but at the wrong place and one sees no effect of the cleaning.
Therefore SF has changed the procedure for some scanners (depending on the driver), making the IR-scan at the same position of the stepper motor as the RGB-image. Then there will be no shift between RGB and IR, except for a small shift due to the different wavelengths and (possibly) the different light path. But this shift is constant for the scanner and quite small (0 or 1 pixel for my DigitDia6000).
In expert mode of iSRD one can set a manual offset to correct any misalignment.
Any misalignment is easily seen if one toggles the IR-channel with the RGB-image in the 1:1 display (original image displayed) by pressing CTRL+SHFT+left mouse button.
Hermann-Josef
Thanks, Jossie and schorschbey. I have spent another hour or so on this pursuit. I have used manual mode, including expert mode, to attempt to change the size, strength, and offsetting of the IR dust reduction. The IR scan finds plenty to change. And I have used the offsetting to center the marked changes directly above the visible dust. Toggling between IR view and RGB view, the marked changes are absolutely steady; they do not even move a pixel, so far as I can see. But I still have failed to make any significant changes to my final results.
As you can see in the attached images, the IR scan marks changes and the IR viewing mode shows a good deal of found blemishes. But the “Correct” result looks, with only the very tiniest differences, just like the “Original.” It is the same with the final scans following these prescans.
I truly cannot tell what I am doing wrong, but I feel dissuaded from further experimentation. I have purchased an electrostatic brush, an electrostatic cloth, and nitrile gloves, and I do intend to clean my digitizing room and clothing before my sessions. But I intend them to employ my DSLR and I hope my physical cleaning and, later, software cleaning will be a lot more effectual than IR cleaning has been for me.
@buddhaauthor
This is very strange. What I see is that the IR-channel is much more diffuse than the RGB image. Perhaps this is only in the screenshots but you should look into that to confirm that the IR has the same focus.
If you don’t mind, you could also make a “raw” scan available (64bit HDRiRAW setting in SF) as a TIF. This is a large file so you have to upload it somewhere for retrieval and post the link here, so I can take a look to perhaps see what is going on.
To my experience, especially with older slides, one cannot remove most of the dust because it is embedded in the emulsion. I have used an air compressor with moderate pressure and used a soft brush for those dust specks still remaining. Still there were a lot of defects in the scan. Correcting them manually is a lot of work. So as you seem to have a good scanner and the software to take care of this, I think it would be worthwhile to find out why iSRD is not working for you. But in the end it is your decision, of course.
Hermann-Josef
Jossie, thank you for the encouragement. I still have two weeks before my return window runs out on the scanner. From my research I am sure it is a good scanner, just one not proving itself right now.
I entirely agree that the blurry detail in the IR scan is highly suspicious. If that’s what the scanner is seeing, then those blurs are too diffuse to remove. The scanner is single-focused, though, so I cannot alter the focus point of either the RGB or the IR scan. The RGB looks sharp enough, so perhaps the IR scan works in a different way. I’ll contact Plustek about the issue.
Meanwhile, here is a link to a 64bit HDRiRAW TIFF of the scanned image. I hope scanned it in the correct manner. Please advise me if this was not correct.
I really do appreciate your kind attentions in what has become quite a saga! Thank you very much.
unfortunately I do not have a google-account. Could you put the scan somewhere, where no account / login is needed?
Hermann-Josef
Jossie, you don’t need a google account to download the file. Well, okay, that’s true now. When you followed the link I had forgotten to change the access settings. My apologies. Please try again.
Good morning,
download now worked. But as I said above, the IR-channel is extremely out of focus and there is a shift of 13 pixels between IR and RGB. The latter can be taken care of by the manual offset in iSRD. But there is no way to make iSRD sensitive enough to the extremely shallow IR-image.
Upper left corner of scan: Left is the R-band, right is the IR-channel.
And there are lots of defects:, which I doubt you want to edit manually:
Also the out-of-focus-IR is very evident in this enlargement (dougnat shape). This screenshot is not from SilverFast but from another tool to enhance the contrast. SilverFast detects not complete (coherent) dust specks, but only parts. That is why there is no successful correction, despite the fact, that defects seem to be marked:
original
IR
defects found
defects “corrected”
I would conclude, that there is definitively a problem with the scanner. So I would indeed contact Plustek.
Hermann-Josef
PS: There may also be a problem with SilverFast in that sense that the defect markings suggest coherent defects. Then these should also be corrected perfectly. But they are not. The slider “defect size” behaves in a non-intuitive manner for me.
Jossie et al.,
Thanks again Jossie, for your help. I have been in contact with Plustek customer service and they are very responsive. They guided me through a SilverFast software reset, then asked me to rescan using Vuescan, and finally concluded that my scanner’s IR channel is bad. Exactly what you concluded, beforehand.
Though I am leaning heavily toward going the DSLR route, I understand that scanners may be like lenses: some “copies” are just better than others. My current Plustek 8300i is clearly inferior, but my next one may be wonderful. I think that, with this many hours sunk into the process, I should replace my copy. Only if I can evaluate a better sample, can I evaluate whether I should go the scanner route or the DSLR route. Thank goodness for Amazon’s 30 day return policy!
Oh, one more question: How did you find that there was a 13 pixel correction to be made between the RGB and the IR scans? I experimented with 13 pixel shifts in four directions, but no adjustment made any thing look better on my end.
More soon, and always thanks.
I just experimented with the manual offset in iSRD. Again there is something strange in SF. It is not sufficient to turn off the manual offset but one has to set it explicitly to 0,0. Then the shift is evident to the value of (13,-1) found by the automatic mode.
Here is the more exact shift determined with imageJ:
On the left the RGB-, on the right the IR-image.
The table below shows the center XM,YM of the signal in the rectangles marked.
Shift in X = 14.5 pixels, in Y = 2 pixels.
Hermann-Josef
Dear Jossie et al.,
I have received my second Plustek scanner. I attach images for you, but I can quickly say that I am encountering the same problem. I can see that the there is a pixel shift to be made to get the IR to work efficiently. But the IR scan seems again entirely out of focus to me; I can’t imagibe any tweaking of pixels that would help much when they are as blurred as this.
I will ask Plustek for help again, but I imagine I am returning this scanner, paying the return cost myself, and not buying a third scanner from them. I really do understand and respect your wise advice to get a scanner that can quickly remove decades of dirt and other defects from my slides. It just seems that the most expensive of Plustek’s scanners cannot. Perhaps I will consider Pacific Imagining’s leading film scanner, but that is a lot more money, given it does not include high grade software.
Franz
In case anyone cares, I purchased a replacement scanner directly from Plustek (not Amazon). The IR scan was blurry (as with the previous unit I purchased). I contacted Plustek with specific details of my problems. They replied:
“Unfortunately we have tried all of the options and even the options from the developer. Hopefully if you are still within amazon’s 30 day return period so you can pursue that option if you see fit.”
It seems scanners are just dead. And/or customer service.
Jossie: I truly appreciate your efforts to get me going with IR dust/scratch removal. It just seems as if hardware issues are going to prevent that.