Minolta 5400 Banding Caused by Radio?

[seth_prince]

I was scanning a b&w Scala slide last night, 16 bits, 16x multisampling, while

listening to the radio on my walkman. I was switching between AM and FM. When the

scan was done I opened the image in Photoshop and nearly fell out of my chair. I had

developed such a horrific banding problem. I scanned the picture again, but this time

I went to take a shower instead of sitting in front of the computer listening to the

radio. Came back, opened the image and magically, the banding had disappeared.

 

The walkman was laying right next to the Firewire cable the first time, when I

developed the banding issue.

 

Don't know if this helps anyone out there, but I thought I would post my findings

[jack paradise]

If you can replicate this at will, then you have found something really interesting. (Don't need to take a shower each time though...)

[david_hansen8]

Very interesting. I had a scan "jump" on me this weekend, and it may have happened when I turned on my PC speakers (one is right next to the scanner's cable). I'll have to try that again.

[johann_fuller]

My Nikon 9000 will generate misplaced pixel bands if it has other Firewire devices running or plugged in at the same time.

[seth_prince]

Jack, I'm 99.5% certain it was caused by my walkman.

 

I wonder, is there anything that I could insulate the firewire cable with, in order to get

an even cleaner signal.

 

Any engineers out there?

[fast_primes]

A fascinating post! It's known in the audiophile world that CD (and DVD) players are susceptible to external vibration. In fact if you were to suspend your stand alone CD or DVD player on 3 rubber leg tips from the hardware store so that they bounce slightly like the turntables of old (decoupling them from room and floor vibrations), you'll find they sound substantially better and track better also! And I've noticed that when I turn off a non-used VCR and tuner, that the sound of my CD or DVD player improves still more-an audio analog of electrical interference!

 

So it seems that electrically and physically isolating one's slide scanner can reap similar benefits and that possibly, a lot of complaints about scanners might have been more the result of environmental factors than the scanners themselves!

[walterh]

"I wonder, is there anything that I could insulate the firewire cable with, in order to get an even cleaner signal. "

 

Yes you can:

 

1. put your walkman or any other electronic device as far away from your scanner and cable.

 

2. buy a decent firewire cable

 

3. complain to minolta for poor shielding of the cable input

 

4. this should not be necessary - just for overdoing it: buy a shield tubing (whatever its called in english :-P its like the shielding net you can see if you take a coaxial cable apart). feed your cable through this and connect one (!) end of it to ground, try first the computer case.

 

i am still a bit in doubt if this is radio interference. if it is your walkman is a terrible radio source and should be returned and complained about to whatever the prober authority is where you live. also the scanner is really on the cheap side as far as protection against radio interference is concerned.

 

hopr this helps.

 

cheers walter

 

DL3KW (ham radio operator)

[stepan_pylyp]

You didn�t say what kind of banding you found which makes an analysis a little harder.

The usual banding produced by the 5400 (my 5400 included) are usually one pixel wide and run in the direction of the scan. These are not (in my opinion) caused by interference during the scan because they always affect the output of the same pixel. It may however affect the calibration of the individual pixels and therefore produce consistent errors across the length of the scan. I have a Photoshop plug-in to filter these bands out and I can provide anyone who is interested with a copy. You also didn�t say whether you re-calibrated the scanner between the walkman scan and the shower scan. If so, then the scanner may just have been in need of a re-calibration.

If you found banding running in a direction perpendicular to the scan direction then it is quite likely that they were caused by external interference during the scan. It is possible that the 5400 is (too) sensitive to interference but it is also quite possible that your walkman generates too much noise.

In order to determine whether the walkman does indeed affect the scan quality simply let the system warm up for say 30 mins. and then scan once with the walkman, then once without, then again with and perhaps a fourth time without. BUT DON�T CALIBRATE OR CHANGE ANYTHING BETWEEN SCANS. It should then become evident as to whether the operation of the walkman is in any way related to the banding.

And most important - post the results!

[seth_prince]

Stephen. Recalibrate the scanner? How do you do that? I didn't make any adjustments

to anything. The lines run in the same direction as the scanner opperates. I will post

examples of both images when I can.

[lwg]

Also try turning the walkman on and off during the scan see if the banding is intermittant in the scan.

 

If it the cause of the banding is it more likely to be the firewire cable or the scanner picking up the walkman's noise?

[stepan_pylyp]

To re-calibrate the scanner from within the DiMAGE Scan software press Ctrl + Shift + i.

 

The firewire interface is digital and therefore has an intrinsically high noise immunity. However, if the walkman causes a high level of interference and it is very close to the cable then it quite possible that this could cause problems. The sensors on the other hand process VERY small signals and are therefore more susceptible to interference. Minolta SHOULD have designed their case and earthing system to prevent normal levels of electromagnetic radiation from affecting the result but ...

[steven_clark]

On the other hand if the ground on the Firewire connection isn't sufficiently isolated from the rest of the system it might be possible to get some signal induction onto the ground line and have that carry over to the ground for the sensor where it can affect the analog part of the process. This however would be something of a long shot, but it might be interesting as the interference pattern of the inducted signal and the sampling frequency could generate some very interesting patterns.

[stepan_pylyp]

I quite agree Steven. I would hope that they have separated the analog and digital grounds well (optically) and that the sensor circuitry is well shielded. It is however quite possible that some interference could still get through. Another likely path is over the power supply. Although the input is already DC I don�t expect everything to run directly from 24V so there is most likely an additional DC-DC converter to step down the voltage for the control and sensor electronics. I�m guessing, but I would expect the analog circuitry to be powered from an isolated output of such a converter. Although these converters isolate well, their ability to block high frequency disturbances depends on the effectiveness of the filters, the layout of the printed circuit boards and the design of the case. As I say, its my guess, any yours is as good as mine (or better)!

 

You also mention that the induced signals may generate some �very interesting patterns�. That�s roughly what I would expect to see, or simply apparently random disturbances, if the interference was induced during the scanning process. However, I see very clearly defined, one pixel wide stripes running in the direction of the scan. Seth also reported banding in the direction of the scan (although we haven�t seen the results yet) which leads me to the provisional conclusion that the affect is constant throughout the scan and is therefore the result of either defective pixels or sub-optimal calibration (which could be due to interference during the calibration phase). Comparing some of my images it appears that the stripes are always in the same place but vary with intensity probably depending on the degree of success of the calibration. This leads me to believe that the banding that I'm personally observing is not a result of external influences (I'm prepared to be convinced otherwise). I find them very annoying but I can filter them out effectively. Others may however be observing different effects.

[imagination_celebration]

Yes, the monolta Dimage 5400 will cause banding when you multisample. That is a flaw with it. But if you don't multisample, it won't appear.

Plus, I don't even think its multisampling improves the image. I didn't see any percievable improvement when I multisampled a slide 4x I had taken from Santorini. It took an incredible amount of time. And upon close examination, there was a redish line that cut through the picture. I could have probably cloned it out in photoshop. It's not perfect, but my images look a heck of a lot better than those grainy drum scans I had made. The 5400 pixel polish, grain disolver, and digital ice smoothe the grain out and produce a superior image than the drum scans I had done in Fort Worth.

[ruslangaripov]

<p class=MsoNormal><span lang=EN-US style='mso-ansi-language:EN-US'>Dear Imagination

Celebration, did you try the latest software?<o:p></o:p></span></p>

<p class=MsoNormal><span class=SpellE><span lang=EN-US style='mso-ansi-language:

EN-US'>DimageScan 1.1.3 English<o:p></o:p></span></p>

<p class=MsoNormal><span lang=EN-US style='mso-ansi-language:EN-US'><a

href="http://download2.minoltaeurope.com/OpenMind/Technic/SWFW/mswdipho.nsf/logdwl?openagent&F92BF77B7B9925E249256E450045DE47/$File/Ds113e.exe">http://download2.minoltaeurope.com/OpenMind/Technic/SWFW/mswdipho.nsf/logdwl?openagent&F92BF77B7B9925E249256E450045DE47/$File/Ds113e.exe</a><o:p></o:p></span></p>

It removes

banding problems produced by 5400 itself, but not the walkman ones, <span

class=GramE>unfortunately:</span>)<o:p></o:p></span></p>

<p class=MsoNormal><span lang=EN-US style='mso-ansi-language:EN-US'>It helped

many users on the net to enjoy multisampling after new version was produced

month ago.<o:p></o:p></span></p>

 

<p class=MsoNormal><span class=SpellE>Best</span> <span class=SpellE>regards</span>,

<span class=SpellE>Ruslan</span>.</p>

[lukep]

I haven't tried the new software yet, but I'll tell yo umy experience from yesterday. While scanning a colour slide at 4000 dpi, 16 bit and 4x multisampling, I got horrific pinstriping vertically down the image. I put it down to vibration and tried again more carefully - the stripes were worse, so I called Minolta. They suggested it was a resource problem so I doubled the memory allocation for the scanner software (the one out of the box, which by the way, I don't think is well thought out) on my Powerbook G3. Since then I haven't had this problem! It may be coincidence, but I think it's worth a try if you have this problem. Incidentally, the first time round I found the dust and scratches filter in PS gets rid of the lines, but whether it affects the image detrimentally I don't know.

[simon_batley2]

I was intrigued by this thread when I saw it, I too got very slight discoloured pixel banding in the highlights even after the upgrade to the 1.1.3 software which had nonetheless significantly improved things.

So after reading in this thread about Firewire noise and possible RFI issues from other devices, I isolated my 5400 from possible culprits such as a standalone USB disk drive that had been about 6" away from the scanner, and changed to USBII from Firewire.

And yes it was true, - all the slight banding issues I had had went. I can now zoom into individual pixels in the highlights and its absolutely perfect.

My only problem now - being perfectionist such as I am - I feel I want to redo all my previous scans using this new mode of operation!

 

BTW Ed Hanrick advises me he is using the newer Minolta firmware code in Vuescan 7.6.80 which improves the earlier significant 5400 banding issue � and with the 5400, for me using Vuescan calibrated with a Wolf Faust IT8 target is the only way to go � no colour casts, vivid colours, absolute minimum retouching � if at all, in Photoshop. I also find that with using a calibrated scanner I can see the benefits of using different colour balanced films such as the differences between Kodak E100G and E100GX.

 

 

Simon Batley

 

www.suffolkimages.com

[benj_c]

I recently read a discussion at photo-i.co.uk about the problem with green lines being improved by wrapping the power adaptor to scanner cord in aluminium foil. Worked for most, but not for some.

[rhett_toler]

I sent my Minolta Dimage II to Minolta for repair. I had a horizontal scratch across most of my scans. Very noticeable on backgrounds with lots of blue. Minolta sent me a 120+$ estimate for repair. While the scanner was gone I replaced my LCD with a Viewsonic 19" monitor. After working with the monitor a bit (Optical, ect) I noticed a faint rolling horizontal line in the monitor. Not as fast as 60hrz hum, but a rolling line.

 

What was it? My Siemens 2.8 GHZ cordless phone base. I had also declined my repair on my Minolta scanner. When the scanner returned (un-repaired) and the Siemens out of the room, perfect scans.

 

Lesson, I should have known better. I have gone through great pains isolating devices and power in my listing room, why should the PC be any different?