Minolta 5400 Mark 1 - ZigZag artefacts

[scarbantia]

This implies that it's difficult or even impossible to verify the suspicion that the noise problem appeared suddenly. The scanner may have performed the same, always, but you only noticed it at a certain point.

I think the difference is quite obvious. I have looked at many of my previous scans and none of them show the defect.
In every scan, when Silverfast scans the first layer without Multi-Exposure, the defect is very noticeable(1:1 view mode). When the software adds the Multi-Exposure layer, the zigzag pattern is significantly reduced.


So I see the following, but I can reduce it when ME is on:

 

[scarbantia]

@scarbantia, that's with grain dissolver enabled, right?

Do you feel that you need to enable it on slide film? Unless it is Fuji Provia 100F and a scene with a lot of areas where infamous "pepper grain" is visible I would never enable grain dissolver. You loose a tiny bit of sharpness and scan time basically doubles (depending on the film this can cause the film to curl during scanning).

Grain Dissolver is not supported by Silverfast. I'm scanning RVP50 slides.

 

[koraks]

I think the difference is quite obvious.

There's an obvious difference between the samples, but as you have also stated, there are other differences apart from the noise pattern. Apparently, the scanning parameters were different for both samples.

Mind you, I'm not implying that the problem is somehow not there, or that it can vary from one scanner to another (we have evidence of that just as well) - I just wonder about the supposed suddenness at which it appears.

As said before, I expect that trying to improve the situation would require fairly thorough work on the electronics of the scanner, starting with the power supply - which is not really an isolated unit, but rather a function that's distributed across a large portion of the circuit. On the scanner of @Archiloque I was very hesitant to do this work because the problem only manifested itself under highly adverse conditions (straightforward scans of well-exposed and -processed slides showed no noticeable problems) and there's always an inherent risk to undertaken such repairs (esp. without a clear diagnosis of the problem).

 

[scarbantia]

I found the latest scan, which is flawless. I have to admit that, there was a few months between the two scans, but the scanner was out of service during that time.
Unfortunately, it also shows up a little in well-exposed photos.

 

[koraks]

Alright, well, either way, it's annoying for sure, I can imagine that. I guess if it were my scanner, I would probably undertake some probing around with a scope if the problem bothered me enough.

 

[scarbantia]

Maybe I will buy another one and try to replace the components one by one.

 

[brbo]

Grain Dissolver is not supported by Silverfast. I'm scanning RVP50 slides.

Interesting. I only have SF 6.6.x where GD is supported. I don't use SF with Minolta 5400, since Vuescan does a better job (which is not always the case with other scanners).

Here is my RVP50:



200% crop:

Lifted gamma +1:

As said, with GD it's considerably worse:

And using a profile made from IT8 target would make it worse as well. All scans with 6x multisample, single sample is also much worse.

It's interesting that R, G, B noise patterns run independently from each other. G channel is the dominant one on my scanner.

 

[Archiloque]

Maybe I will buy another one and try to replace the components one by one.

Good luck then, because I have 3 of them, and they all show the same symptoms more or less.

 

[scarbantia]

Interesting. I only have SF 6.6.x where GD is supported. I don't use SF with Minolta 5400, since Vuescan does a better job (which is not always the case with other scanners).
[...]
As said, with GD it's considerably worse:

View attachment 388870

And using a profile made from IT8 target would make it worse as well. All scans with 6x multisample, single sample is also much worse.

It's interesting that R, G, B noise patterns run independently from each other. G channel is the dominant one on my scanner.

Wow, that's very interesting! When you use GD, you have less light, and the scanner sets a longer exposure time or a higher signal gain.

I did a quick experiment:
I removed the backlight from the scanner and used an external light source to illuminate the film. (I propped up the grain dissolver).

Observations:
Silverfast measures the light and sets the white balance on startup, but does not correct for lighting irregularities. The Dimage Scan Utility software homogenize the brightness differences!

However! If I use stronger light, the scanner sets a shorter exposure time and scans faster. Stronger light = less noise, much more shadow detail and faster scanning! It's amazing with Multi-Exposure.

The zigzag pattern is not gone, but it can be significantly reduced.
I need to make a separately controllable RGB+IR LED backlight.

Yes, the Silverfast 6.6 supports the Grain Dissolver, and the backlight homogenity corrections. The operation is more similar to the Scan Utility software.

 

[koraks]

Hmmm, OK, so does this mean that maybe the backlight is simply going on these scanners? Would make good sense; those CFL's don't last forever. They get dimmer.

 

[brbo]

I found a ten year old scan of Velvia 50 and it shows similar pattern if I heavily lift the shadows. Unfortunately, I only have jpg and scan is from another 5400mk1 that I don't longer have.

I have a spare bulb for 5400mk (I think, it seems excessively long for a scanner that needs to provide 25mm wide light beam, so it might be for Minolta Multi Pro), I might try swapping it and making comparison scans. But, there is no guarantee that the spare bulb is indeed new/unused (could've been pulled from a dead scanner).

edit: checked the spare bulb size, definitely not for 5400

 

[brbo]

We can't compare the state of our CCFL bulbs directly, but we could compare them indirectly.

Vuescan always calibrates the scanner so that Exposure Lock value of 1 will be just bellow the clipping for Dmin of 0 (no film in the holder). If one's light bulb is particularly bad, then scanning of entire empty frame will be considerably longer than others.

What I did is:

- turn on the scanner
- opened Vuescan
- selected 'Scanner - Calibrate' exactly 2min after turning on the scanner
- after calibration, set the lock Exposure at 1, 5400dpi, 1x multisample, manual focus, 48bits (no IR)), no raw (since that brings disk speed into question with folks with slower computers)...
- and record the duration of scan time for frame 1 of the negative strip holder (FH-M10)

My time is 47s.

The other test would be to compare the "black" noise. Scanning totally opaque material at different exposure locks. Higher time (higher exp. lock) should produce more noise, but it would be interesting to compare samples we get.

 

[albireo]

Good luck then, because I have 3 of them, and they all show the same symptoms more or less.

I have one and it performs flawlessly. I have never, ever seen this particular issue. I use mine all the time.

I use it with Vuescan FWIW, and always calibrate before usage. I never use the grain dissolver and use multisample only very very rarely set at 3X maximum.

 

[brbo]

@albireo can you post a 5400dpi tiff of a totally opaque material, scanned at exp. value of 1 and no multisampling? Either unprocessed raw of tiff with Colour balance set to None, 0.25 and 0.75 for low/high curve points and brightness 1.

A small 100x100px crop would be enough...

 

[albireo]

@albireo can you post a 5400dpi tiff of a totally opaque material, scanned at exp. value of 1 and no multisampling? Either unprocessed raw of tiff with Colour balance set to None, 0.25 and 0.75 for low/high curve points and brightness 1.

A small 100x100px crop would be enough...

Will have to generate one I guess. What's the best way to do so? Glue a bit of paper onto a negative?

Or do you mean an actual 'saturated' exposure? Positive or negative?

 

[brbo]

Something that would totally block the light. I used two thin black plastic sheets one on top of the other. One would probably do it just as well...

 

[albireo]

So I tried this with a bit of brown drawing paper

Levels lifted:

Grayscale scan - green channel selected

Grayscale scan - red channel selected

 

[albireo]

Actually - I think I was wrong and my unit is showing the issue you're discussing, as well.

I went looking around the 5400dpi scan searching for patterns like those you are showing in this thread and I found a horizontal band about 100 pixels high, running through the whole larger dimension of the frame, towards the bottom of the frame, that might be the same as what you're seeing.

See bottom half of this image:

Notice the unevenness of the pattern. Is this what you're seeing? Does this suggest the line sensor is on its way out?

EDIT - just read the entire thread. Would be interested in any leads. I agree with post #30 by @koraks in that I had never noticed any patterns in my scans of normally exposed+developed negatives (I have no slides around here to test) so it seems to be a non issue for my work. However, if this is a symptom of decaying capacitors or the beginning of the end of the device, I'd be interested to know..

 

[destroya]

I read this thread a few days ago and wanted to see my results. turned on my 5400 scanner and nothing. its stuck with the blinking light forever and the computers (tested on 3 machines) no longer see the scanner. ugh.

wish i could help out, but think my scanner is finally dead.

john

 

[brbo]

Notice the unevenness of the pattern. Is this what you're seeing? Does this suggest the line sensor is on its way out?

EDIT - just read the entire thread. Would be interested in any leads. I agree with post #30 by @koraks in that I had never noticed any patterns in my scans of normally exposed+developed negatives (I have no slides around here to test) so it seems to be a non issue for my work. However, if this is a symptom of decaying capacitors or the beginning of the end of the device, I'd be interested to know..

I doubt something is wrong with your scanner. It's in line with what others are seeing and I don't think all the scanners have rotten by now. This site (Italian) has the best test of Minolta 5400 (I and II) ever made and also has Nikon 8000 test. If you check the samples and lift the shadows heavily you can see similar noise and scanners were relatively new at the time of test.

 

[Archiloque]

I doubt something is wrong with your scanner. It's in line with what others are seeing and I don't think all the scanners have rotten by now. This site (Italian) has the best test of Minolta 5400 (I and II) ever made and also has Nikon 8000 test. If you check the samples and lift the shadows heavily you can see similar noise and scanners were relatively new at the time of test.

I am not sure I see the same thing. With multisampling their test images show almost no noise whereas the zigzag effect is even more pronounced with multisample enabled.

Also, and as stated before I've scanned the same image with the same scanner at different stages of its life, and the scans from 2015 do not show the same patterns. We are running circles here.

 

[MattKing]

I wouldn't exclude the possibility that current problems are related to changes in operating systems and driver software.

 

[Crysist]

Does this issue have any relevance to the artifacts that occur when you enable multisampling, GD, and ICE? I saw that cited in this Italian review, but I see some of the examples everyone is showing are not with each enabled.

EDIT - oh @brbo already mentioned this review

Wow, that's very interesting! When you use GD, you have less light, and the scanner sets a longer exposure time or a higher signal gain.

I did a quick experiment:
I removed the backlight from the scanner and used an external light source to illuminate the film. (I propped up the grain dissolver).

Observations:
Silverfast measures the light and sets the white balance on startup, but does not correct for lighting irregularities. The Dimage Scan Utility software homogenize the brightness differences!

However! If I use stronger light, the scanner sets a shorter exposure time and scans faster. Stronger light = less noise, much more shadow detail and faster scanning! It's amazing with Multi-Exposure.

The zigzag pattern is not gone, but it can be significantly reduced.
I need to make a separately controllable RGB+IR LED backlight.

Yes, the Silverfast 6.6 supports the Grain Dissolver, and the backlight homogenity corrections. The operation is more similar to the Scan Utility software.

Wait, you can do that?? I was thinking that everything was calibrated to each other's level/color/voltages and that changing any component would require a huge amount of work to "balance" the differences from the stock component.

@koraks I have no EE expertise but I gotta say, the lengths you've go towards pinning down this problem are impressive... from my standpoint, not knowing anything! If I may interrupt the more technical-minded people in this thread, is it possible to replace the ADC and gain (rimshot) better noise performance? I imagine the ADC is half the battle, the line sensor might not be up to par, but I was under the impression that ADCs in cameras and other imaging devices were rather bad in the early 2000s...

I actually recently bought a 5400 Mk I - because it was better than the Mk II in every way except for speed. I love it except I'm still a novice at color balancing to original slides and the holder's ability to keep the film flat as well as the effort it takes to get the focus correct - especially some of my curly Velvia, despite sitting under weight for a few days - leave a lot to be desired. But I haven't seen this issue in any of my scans, I just checked one of the tiff's I scanned last week.

 

[koraks]

I wouldn't exclude the possibility that current problems are related to changes in operating systems and driver software.

I think that's a red herring.

is it possible to replace the ADC and gain (rimshot) better noise performance?

Practical answer: no. It would involve basically a complete redesign of the electronics. Might as well replace the light source with something modern as well while you're at it. And perhaps have a look at the optics. Heck, why not dump the whole think and build a new one? That's what you'd end up doing.
One of the key bottlenecks is that it's unlikely to find a pin-compatible, instruction-compatible drop-in replacement for this ADC with better noise performance. So you'd be re-designing the PCB (or making some kind of Frankenstein piggy-back device) as well as implanting a new processor with newly made software into the thing at the very least. Which is a gargantuan task in itself.

As to the problem at hand, the more I think about it, the more plausible it seems that gradual degradation of the light output of the CFL is the culprit of increasing noise. I also think it's good to keep in mind that the device has never been entirely noise-free - which is true for EVERY analog circuit on the planet, by definition.

 

[Archiloque]

Might as well replace the light source with something modern as well while you're at it.

Actually, it's probably the most sensible solution. I've been looking at small led light source with hight CRI. Less heat, stable light, possibly less noise. It shouldn't be a difficult mod and it might bring additionnal benefits.