Minolta 5400 "Raw-scan" workflow with Photoshop CS

Hello,

 

I've read, or tried to read, all of the mindbending ideas on the web and newsgroups about

adjusting the scanner sw or using Photoshop CS to do the same thing. I have 60 negs to

scan for an art

project. I don't want to re-do these, so I'm trying to nail this out

of the gate.

 

I own a 5400. I tend to subscribe to the 'raw' theory of scanning, i.e.

scan without corrections because scanner sw modifies the scan

post-scan and so just use PS to modify post-scan. I own the CS version, so

working in 16 bit (or whatever, 15+ bit) isn't a problem. I am very

good at Photoshop. I have color neg and b/w neg., a couple of Fuji

velvia chromes.

 

What I am wondering is:

 

1) using the 'raw' scan theory, is it ok to use the Minolta 1.1.5 software, or

would Vuescan or Silverfast make any difference USING THE 'RAW' theory? My

guess is the Minolta sw is fine using the 'raw' theory. I'd rather use PS,

a tool I know well, than to use some other sw that I must become an

expert at.

 

2) If other sw can do a better "raw-scan" by opening some secret

hardware-level nirvana, what is the sw and what are the sw settings to do so?

 

3) AND, whatever the answer to the 2nd question, what are the

setttings in the Minolta software for a raw-type scan?

 

My guesses are:

 

- 16bit linear (vs 16 bit), and can someone explain the difference?

 

- don't know about about autoexposure vs manual,

 

- autofocus unless there's a problem

 

- don't know about multi-sampling

 

- don't know about color matching output space (adobe RGB?) and ICC

profiles

 

- Color neg: best to scan as neg or positive? If positive, invert in

PS?

 

- B/W neg: best to scan as neg or positive? If positive, invert in PS?

 

- Fuj chromes: no idea

 

- ICE on, at least with color neg.

 

- Grain dissolver off with color neg, don't know with b/w neg

 

Thank you for your answers.

 

Best,

David Carson

To start to answer my own question, I saw grant's ideas here:

 

<p>Minolta 5400 vs Nikon 9000 quick comparison: <a href="http://www.photo.net/

bboard/q-and-a-fetch-msg?msg_id=008vfC&unified_p=1">link</a>

 

<p>He says scan at 16 bit linear, color or b/w positive to get the best raw output from the

5400. I still don't know about what's the best way to adjust the 'master gain', as he puts it

(I think he's referring to the master setting in the exposure tab). Or if manual exposure is

best (in the preferences), or if the color matching space is important or not.

yea, 16bit linear, autofocus....using the master gain you can control the exposure with the histogram....works great for me

Thanks grant. You always scan in positive, right? Ever see a need to use multi-sampling?

Ever scan true b/w film (assuming you have scanned color neg and slide)? I'm assuming

you use ICE, and I guess Grain Dissolver comes on too when ICE is used (RTFM, Dave).

 

I guess two black holes are also the effects of the whole profile-schmofile thing, and the

biggie for me is "how to read a histogram", or "what a good histogram looks like." I've

always just went by the idea that you don't want the histo to blow out on either side of the

-+ scale, but blowing off the top is ok. I'm not sure how far the histo should be from

either end. Thoughts?

raw is always positive...as far as i know....i use diffuser when the negs are really grainy, other than that i dont....i never use ice...id rather spot it if anything...

Thanks again, grant!

 

But how about the histogram. I've always just went by the idea that you don't want the

histo to blow out on either side of the -+ scale, but blowing off the top is ok. I'm not sure

how far the dark and light triangles on the histo should be from either end of the curve.

Right on the edge?

i usually put it as far right as i can without it cutting any data...the scans will still be very dense in photoshop, but thats really easy to fix with a few curve adjustments....the tones are far better then the software renders...

Cool grant. I'm getting somewhere here, thanks to you.

 

You don't mention if you manually expose (preferences) and then tweak the master

control, or auto expose and then tweak the master.

 

And have you ever mult-sampled? Worth it?

I have had scanner 3 days but it is my impression that master exposure control is not functional during a 16 bit linear (raw) scan.

i think in raw there is no autoexposure....i havnt seen a diff when i click it off or on...so....and i dont find multisample is really worth it for any of the negs ive scanned so far....just sucks up a ton of time...

As far as I can tell "16-bit linear" is essentially raw data.

 

multi-sampling mostly helps where the film is very dense, i.e. mostly for the shadow of slides, but also for overexposed negs.

 

So far I've not been able to get proper results with color negatives, guess I need to learn more (scanning them as negs results in clipped shadows and highlights, and scanning as positives results in ugly colors).

 

I adjust the histogram to not clip anything, and then as appropriate use curves to compress the highlights/shadows and boost contrast in the mid-tones (or more advanced if needed).

Hell, this thread has been so informative, let's continue down the workflow.

 

- ever tried any of the sharpeners? Photokit, focalblade, nik, QImage's new Equalized

Sharpener, Focus Magic, etc. Gotta work in 16 bit.

 

- Rip: use one or your stock driver? Will the epson 2200 driver handle 16 bit files?

 

- what paper do you like?

 

- how do you center the print on the paper?

 

- have you calibrated your monitor? With what?

 

- and finally, what dpi do you like to print at?

 

whew

heres a few examples for fun...<br>

<br>

1. straight raw scan unadjusted

<p>

<img src="http://www.streetzen.net/bk/54001.jpg">

<br>

<br>

2. straight scan inverted

<p>

<img src="http://www.streetzen.net/bk/54002.jpg">

<br>

<br>

3. scan sdjusted

<p>

<img src="http://www.streetzen.net/bk/54003.jpg">

<br>

<br>

4. 100% crop of 5400dpi original

<p>

<img src="http://www.streetzen.net/bk/5400crop.jpg">

focalblade is excellent, so is a number of imagingfactory effects as well....

Hope this helps and if I can explain it. Auto exposure does work during a raw scan. It is critical that you select just the right section of the film. Keep moving the selection area and doing prescans until you get the brightness you want. Then lock in the auto exposure. Read the help section to understand how to select and lock auto exposure. The help section is the little book with a ? mark on it at the top of the scan utility. I am scanning Kodachrome II (very dark) and this is my work around for exposure compensation on raw scans. Now to really confuse the issue- There are some who feel that making the adjustments color, exposure, ect. with the scan software in 16 bit will give better quality than major corrections in photoshop.

i prefer manual....never really know what that auto crap is doing...

Sorry to interfere folcks but there seem to be a pool of talent on this thread that may be much help for a low techie like me. I scan mainly slides on a Minolta 5400 but have much trouble getting the autofocus to work consistently; sometime it does some other times, not. As it take a hell of a long time to scan each slide if you select the multiple scan option, the uncertainty is a problem. Is there an easy way to make sure that the scan will be accurately focused and sharp before I proceed with the final scan?

Just my 2 bits worth on autofocus - I have found that I get sharper results using manual focus on the 5400 - I know it's a pain to have to set up for every scan, but I have made comparisons with pos, neg & B&W, and in every case, manual focus gave a better result.

 

I find the Minolta driver software irritable - it often hangs up or causes XP to throw up the "......has a problem and needs to close..."dialogue. (Dedicated 40 Gig HDD & 512 mb mem). I believe VueScan is much more stable and will change to it as soon as budget allows.

Should also add that if your final result is for 4x6 prints, then autofocus and batch scan will be fine, but if you are scanning at 5400 and final prints are to be fairly large, manual focus is a must.

Joe

 

if you make corrections to the image using the sliders under the ?Exposure control? tab, then you are producing an image during the scan that is colour corrected and with the correct exposure rather than making post scan corrections which to some extent, will always degrade the image.

 

Lorenzo

 

autofocus works best on high contrast parts of the image, try setting the autofocus point (in the prescan window) to a position about 20% from the center in an important part of the image with fine detail or on a line (eg the edge of an eye or building)

Looks like many of the questions are answered already so I'll contribute my experiences. I don't work much with negatives or black and white. I primarily use slide (velvia, provia and E100VS). I use the Minolta software.

 

multi-sampling - I've experimented a lot with this. For darker slides it can make a big difference. It pulls a lot more detail out of the shadow areas. It gives you more information to play with in PS. I don't think you get much more out of x16 than you can get from x4. For most of my even light (less contrast) scans it isn't worth the extra time to multi-sample.

 

16bit linear vs 16 bit - Personally I don't use the linear option. I have never found an advantage and I don't like having to spend extra time trying to get the exposure right in PS.

 

Autoexposure vs manual - Manual is better as auto can clip. As I'm sure you know you can easily adjust the exposure manually in PS using the Levels (where you can see the histio).

 

Color matching output space (adobe RGB?) and ICC profiles - I turn it off because I use a custom ICC. Some people have good results using AdobeRGB in matching.

 

I use ICE on all scans. I know it can sometimes dull the image but that can be fixed with the unsharpen filter. Oh, ICE can sometimes cause a slight red cast (fix in PS or use custom ICC).

 

Epson 2200 driver handle 16 bit files - yes it should. I have the R800 which is similar in that it uses the same inks and print heads but only does A4.

 

For the best paper to use refer to this link:

http://www.drycreekphoto.com/Learn/printers/epson_media.htm

 

Yes I calibrate my monitor. I use the PhotoCal. Check some other site and news groups to find the best. I don't know much about it.

 

The most important thing I did to get my scans right was to profile my scanner. I know that some other users are happy with the results they get straight from the scanner but I was not. You can get cheap scan targets from Wolf Faust and LittleCMS is a free profiler program that works pretty well. It has saved me lots of time and headache in PS.

 

I agree with Graham about autofocus. The autofocus picks the center of the frame as its focus point by default (can't be changed). If the center doesn't have much contrast there is a good possibility that it won't focus correctly. So in manual mode pick a high contrast area. You should get an idea from the prescan if it is focused correctly.

 

Hope this helps,

Patrick

Lorenzo---

 

After you have done the many prescan checks for exposure, color, focus ect.. (Use manual focus)

Do your high quality scan of just a small portion of the image. You can unlock the crop lines and drag them to a critical part of the photo to selectively scan that section- then check the results (usually around a 12 meg scan) before scanning the entire image. This method is to achieve the very best quality scan of special photos.

Costas & Patrick P.------

It looks like you are in the camp that favors doing as many corrections to the image with the scanner software as possible.

After that doing any fine tuning with photoshop. Do you scan directly to photoshop? If you have saved as a tiff when opening with photoshop you can specify a profile. The profile for the particular minolta scanner and scan type are shown in the pop up box when selecting the scan profile. Such as 5400 pos.

Patrick P.====

 

Could you describe profiling your scanner?

 

Thank you

Patrick P.

 

Also, could you describes the problems that profiling prevents.

 

Thank you

Joe