photo scanning vs negative scanning

[VegasBud]

gadees,

From the spec sheets you provided links to, the V500 doesn't appear to offer any advantage over the V330 in a "prints only" scanning environment. Either scanner far exceeds the needed capability for scanning prints. Keep in mind that there is no point in scanning prints at more than 600 dpi (arguably at more than 300-400 dpi), and that Digital ICE doesn't work when scanning prints. Technically, the V600 (and above) scanners are supposed to scan prints using Digital ICE, but the few references I've seen about the matter indicate it introduces more problems than it solves.

Strangely enough, however, the best answer to your question appears to be "It depends on the relative prices you actually have to pay".

Over here (in the US), Amazon has the V330 for $120, and the V500 for $140. For $20 more, it might be worth it to cough up the extra money to have the Digital ICE option available, just in case you later find yourself with some transparencies (slides or negatives) to scan, and might want to give Digital ICE a try.

At the Epson store (in the US) the V330 is $120, and the V500 is $180. For $60 (50%) more, it's not worth it (at least not to me), just to keep the "Digital ICE for transparencies" option open.

At the Epson site (in Australia) which you linked to, the V330 is $249, and the V500 is $449. If that's the prices you actually have to pay, you have my condolences, and my advice is to absolutely, positively, completely forget about the V500.

[gadees]

Oh my gosh, thankyou so much, you have no idea how much better this makes feel! I've been itching to purchase, but holding off the last few days to research some more - had trouble finding information that I felt sure about.

I can get the V330 for $200 and the the V500 for $350- possibly $300 after $50 cashback. $20 difference would have been a no brainer, wish it were that little difference here. But even the $100 difference is what started making the decision harder again, because I do have some negatives - but very few it's probably not worth worrying about them, and I'm not even sure if they are anything good!

In any case, I feel comfortable about buying the V330 now and I think I will ... just ... go... for ... it!

Will be ordering today, hope to get it by the end of the week. Actually looking forward to scanning! Will be researching best methods and tips to get the best possible results during the week and hopefully understand all I need to know by the time it gets here.

Thanks again for your reply vegasbud, your help is truly appreciated.

[lordsmurf]

Benefit of V500 is ICE. That is all, nothing more. ICE = remove dust, etc
The V500 is better for archival scanning, for that reason.

V300 has some limited software dust removals that work in certain scan modes.
I use it for quick print scans for Facebook.

[gadees]

Hi lordsmurf, thanks for the clarification.

Ordered the V330 now.

I know I wanted to get it the other day, but I had a look and found about 20 roles of negatives - thought I only had a few, though only 3 I probably want the best results from. And I couldn't decide if I should maybe get it just in case the digital ice would really help them. Plus just in case I discovered another pile later on.

Would you have any recommendations of tutorials to follow for best scanning results?

I've downloaded Vuescan and Lightroom trials. I've been practicing lightroom and following some youtube tutorials to get the hang of adjusting photos. Getting better at it.

I think I will go with scanning them in RAW using Vuescan, and then come back and process them later. Wish I could process in lightroom, but looks like I have to do it in Vuescan.

[VegasBud]

gadees,

As long as your negatives are 35mm, the V330 should scan them fine. Digital ICE is optional...not mandatory. Whether any given person uses it depends on the individual, their workflow, and their preferences. Personally, I disable Digital ICE even when it is available.

It sounds like you really want to do the best job you can, which is always good, but (much like the V300 vs V500 decision) buying Vuescan and Lightroom seems like a bit of overkill to me. There isn't much point in paying for features you're unlikely to need, or even use.

Lightroom, for example, is a great program...I use it almost every day...but it seems to me that it's main virtue is as a streamlined workflow platform geared to the professional digital photographer, which isn't what your use really is.

It's been a while (years) since I used Vuescan, but my recollection is that it's strong point is the ability to integrate a considerable amount of post processing in the scan process. If that's the way you want to go, by all means go ahead...it's not the way I would go, however.

Since I don't know you, the materials you'll be working with, or what your ultimate goal for this project is, my "best practice" recommendations might miss the mark by an unknown amount. In that case, just let me know, and I'll adjust. At any rate, the following is what I would recommend for general scanning of prints and transparencies on a flatbed scanner (in this case, the V330):

Use a two step process:

1) Save the best (unprocessed as possible) scan you can get out of the scanner as your "original"

2) Use a separate program to post process to "optimal"

By doing so, maximum flexibility is retained. The more your "original" contains scanner-applied processing, the fewer your options. The farthest you can "undo" is back to the "original". Any processing applied prior to saving the "original" is locked in, and can't be "undone" later (except by rescanning).

For the scanning part:

At least start with the scanning software that comes with the scanner. The V330 has a "Professional Mode", so select that. Turn off any auto-sharpen, auto color correct, etc. The only part you may need to adjust is the settings to get the histogram so there's no clipping of the highlights and shadows (the histogram curve is fully contained in the window).

For prints, scan at 600 dpi to be absolutely sure all information is captured. Scan at 300-400 dpi to be 99% sure you captured all that's available.

For transparencies, you can scan at the full optical resolution (4800 dpi on the V330), but that's probably over the top (in most circumstances). A 35mm transparency, scanned at 4800 dpi, printed at 300 dpi, is basically a 16 inch x 24 inch print. Put another way, the scan would be approximately 6800x4500 pixels (about 30.6 megapixels). To be blunt, neither consumer scanners nor consumer 35mm film merit that level. As best as I can recall, about 2400 dpi is the sweet spot...beyond that, you're just wasting disk space and overburdening the editing software unnecessarily.

Another consideration is whether to scan at 24 bit or 48 bit. Hard drive space is pretty cheap these days, so I would recommend 48 bit, mostly so you won't have to redo the scanning somewhere down the road. It seems like most editing software can at least open 48 bit, even if they only use 24 bit internally, so that shouldn't be a problem.

Save the "original" (straight out of the scanner) as a TIFF.

Don't underestimate the subject of this thread...if you have the negative, scan that...if you don't have the negative, scan the print.

For the post processing part:

You should probably start with whatever editor came with the scanner. It may be all you need for your current purposes, and is probably the easiest to learn and use. Once you are familiar with what it can and can't do, you can decide whether you need "more".

The next step up would be to a freeware program. Off the top of my head, "Raw Therapee" is probably a good choice to start your exploration with. It has the advantage of being able to fully make use of 48 bit images, and has the tools you need for most general post processing tasks. If you want to stick to 24 bit (for now), "GimpShop" seems like a reasonable choice in the freeware category.

Choice of which editor to use is very much tied to personal preference, so you shouldn't hesitate to try other freeware editors to find what what works well for you...and what doesn't. There is, however, a learning curve to consider each time you change software. If time is of the essence, that may be something to consider.

If freeware doesn't cut it (for you), at least you'll have a better idea of what features you want, and what features are unnecessary, when you go shopping for commercial software.

A couple observations concerning post processing:

Learn (or create) techniques/workflows which please you and your audience. There are a ton of tutorials out there which are pure crap. Even well reasoned tutorials are only representations of how a given person approaches a given problem. What's right for one person may be wrong for another person. Use your eyes and your mind (and a little time) to find what's best for you.

"Above all else, do no harm" is just as relevant to post processing as to medical ethics. If something doesn't indisputably make the image better, undo it, and try something else.

[gadees]

Hey VegasBud, thanks so for much for the detailed information. Wish I noticed this reply sooner. I was stressing over what dpi to scan at for a long time. I finally started the first 100 with 450dpi but have been worried it was too high. It was an inbetween the common 300 and 600 people suggested.

I think now I'll up to 500 or 600 as you suggest. For 4x6 color prints. I want to get as much quality as I can. I was worried, that going much higher then 300dpi wouldn't be any good, read around that as you can't get more then the original resolution anyway, you might end up degrading the image quality by adding more pixels then the source.


I got Vuescan, because I thought it was the way to go and wanted to use the RAW feature so I can then process in lightroom. But having second thoughts now, it seems this might not be the best way to do it, and that I should probably do some starter correction in the scanning software..


After reading your reply I had a play with Epson Scan software, and actually think it might be easier. So what I've done is I've gone into epson Professional Mode > turned off all the correction options. However it has an auto exposure that can't be turned off - is this okay?

I don't really understand the clipping. Do you know any tutorials I could look at to learn how to best adjust the Epson histogram, or should I just look up histograms in general? I don't know what sliders I should move and where to get the best results. Is it like in Photoshop Levels, where you edit each color mode and drag the sliders to the end points of the histogram?

By doing so, maximum flexibility is retained. The more your "original" contains scanner-applied processing, the fewer your options. The farthest you can "undo" is back to the "original". Any processing applied prior to saving the "original" is locked in, and can't be "undone" later (except by rescanning).

I found that in Epson Scan, if you go into the configuration settings > Color tab > you can select "No color correction". This turns off all possible (I think) corrections, including the histogram. Would it be any good scanning this way - and then correcting the Levels in Photoshop (just because I know how to do that ? Or is this too little correction during the scan?


I didn't realise how hard scanning is to learn. It's so baffling. I start to think I'm on the right track and then find out all these other things I should be doing or trying. Plus I haven't really found many tutorials / walkthroughs. Especially for the Epson Scan, seems to be nothing there. (User one anyway)


Any help is appreciated, I'm basically just stuck at how to scan with color correction / no clipping in Epson Scan.



Oh, I really like Lightroom. I followed a ton of tutorials and really find it so easy to correct images now. But I'll test out some other programs just in case.

[VegasBud]

gadees,

I apologize for not answering you sooner. I didn't hear you knocking on the door. I hope the quantity of information in this post, and hopefully the quality of the information, will offset the lack of timeliness.

Yes, it is true that the best you can do is capture the detail which is present in the original you're scanning. You can't capture something that isn't there.

No, it doesn't "degrade the image quality by adding more pixels than the source (requires)".

Scanning at too-high dpi still retains all the detail (information) available in the original, no matter how crazy you get with it. Scanning at too-low dpi, on the other hand, discards detail which can't be recovered later (except by re-scanning). When scanning, you're better off erring on the side of too-high dpi than erring on the side of too-low dpi.

Scanning at too-high dpi does have a price, however, which is paid in time and storage space. Scanning takes longer, saving/loading the much larger files takes longer, and editing those files takes longer (all the non-contributing pixels have to be processed, too).

As to what dpi is preferable to scan at, the individual doing the scanning (along with their goals, and the intended use of the scans) is one factor. The choice isn't from a very large assortment of values, though, for someone who "want(s) to get as much quality as I (they) can." As to why that is so, there are two ways to go about it...I can give you the list to select from, or give you the reason why some values are acceptable, and some values are "less acceptable"...or I can give you both.

The list of preferable dpi values for scanning prints (on your V330) would be: 300, 320, 400, 480 or 600.

The reason those values are preferable has to do with the way flatbed scanners work. The long, thin, bright stick-thing (which moves from one end of the scanning bed to the other as it scans) has the optical sensor on it, and processes one (and only one) row at a time. To cover more than one row, the sensor assembly (carriage) is then moved down to the next row. In the specifications for your V330, it gives "4800x9600" as the resolution. The "4800" part (optical resolution) means the optical sensor works on 4800 points per inch. The "9600" part means the "stepping motor" (which moves the sensor assembly up and down the scanner bed) is using 9600 rows per inch. In a nutshell, the sensor reads in one row, and then the stepping motor moves the sensor down 1/9600th of an inch to let the sensor read in the next row.

Now consider what happens when you specify a dpi to scan at. As an example, let's say you chose 300 dpi. A little math shows 9600/300=32...which means the stepping motor should move 32 steps down for the sensor to be in the right position to read in the next row. What happens when you pick 450 dpi? Well, 9600/450=21.3333333. The stepping motor can move the sensor down 21 steps, or 22 steps...it can't move the sensor down 21 and 1/3 steps. All it can do is move the sensor down 21 steps twice, and then 22 steps once (in every group of 3 rows). Let's do one more example at 500 dpi (since you mentioned it)...the math says 9600/500=19.2, which works out to 4 rows at 19 steps, and one row at 20 steps. As you can see, distortion is being introduced whenever the stepping motor has to deal with fractional steps. If you "want to get as much quality as you can", it's a good idea to not introduce distortion needlessly.

Next up would be Vuescan and RAW files. This post is already weighing this thread down with material likely to give some people headaches. Rather than add to the bulk and the readers' pain, I'd prefer to leave this particular topic to the simple observation that a Vuescan RAW file from a V330 is functionally identical to a standard TIFF file from the Epson Scan software with all the "enhancements" disabled (as I mentioned in a previous post). If you really want more information, I will provide it.

We can probably combine "auto exposure", "histograms", "clipping", and "configuration settings" into one recommendation...looking it up at the Epson site, and playing around with the Epson 4490 I got to use for this thread (which is the only scanner I currently have), using "No Color Correction" in the configuration settings is definitely the way I would go.

Just to make sure you have the answers to questions you asked:

Yep, histograms all display similar representations of the pixel values. Photoshop histograms are much like other histograms...as are the histograms in Lightroom, which you've probably noticed.

"Clipping" just refers to situations where the pixel values are truncated. If there's a wall of data right up against the right side of the histogram, the highlights have been clipped (some quantity of the tones in the highlights have been lumped together into pure white. If there's a wall of data on the left side of the histogram, the shadows have been clipped (some quantity of the tones in the shadows have been lumped together into pure black). Clipping is a bad thing, something to be avoided (if at all possible).

...is this too little correction during the scan?

I had to think about that one for a minute...the concept of "too little correction" is a foreign concept to me. Hmmm...now at the three minute mark, I still can't conceive of a correction during scanning which is a good idea. If you want the maximum amount of image data from a scan, don't do anything to that data until it's safely stored away in the "original" TIFF file.

On other things you mentioned:

Things often seem hard when you first start doing them, but get easier the more you do. Scanning may seem difficult and complex right now, but there are really only a few things that can bite you on the backside, and you learn those pretty quick.

I really like Lightroom.

So do I. If Lightroom is fast, effective, efficient, and comfortable to you, that's about the best you can hope for. I only mentioned other alternatives to make sure you understood that you don't have to spend $300 for an editing program to work with scans. It's your decision, your workload, and your money...only you know what the right way for you is. For my uses, Lightroom is worth $300. If Lightroom is also worth $300 to you, I would be the last person to try and convince you not to use it.