V600 and Holders

I believe I must be missing something in the transition of a scan. I started out with some negatives today, scanned to .tif and although some went from 155mb to several hundred I thin converted to jpeg to do some post processing although the still were large, resized but still get too large a file to send out for printing with many being 20-30mb. I scanned at 48 bit and 6400dpi best I recall on most of them

Mpix can't print from such a large jpeg I was told and really, that size isn't needed for something like a 11x14. I don't know if I'll lose quality continuing to resize a jpeg after post processing when it still appears too big. What am I missing here?

I'm not n expert on this. But if you're only re=opening one jpeg, there really shouldn;t be too many problems. I think the problems come when you do the saves and re-open a number of times..

Also, what's the problem with post processing tiffs? is it that your program will only see jpegs? If that's the case, save the jpegs with no compression. The file will be as large as a tiff but will be a jpeg to be opened in your post processing program. Also, you're really not going to get more resolution than 2400 from the V600, probably less. So there's no point setting at 6400. If you need more resolution for printing, then up-rez after editing in post procesing when you're ready to print. Then save as a jpeg with minimum compression to keep the file size no bigger than can be used by MPIX.

There's no problem with editing the .tifs but I thought I'd just leave an unadulterated copy of the scanned negative and save the .tif as a jpeg. When I did that is when it was still way too large(100+mb). I'd then do my tweaking with that file, brightness/contrast/color/sharpness, etc., and save...still too big so I'd resize. It just seems like too many steps to get you where you want to go.

Some of those I went with 4800 instead of 6400 as I was doing a bit if experimenting. I wanted to extract all the info I could from the negative. However, as you mentioned, if the V600 can't go that far I don't know why it's there as an option.

Having the larger scan sizes are used to advertise and compare to other scanners to gain a sales advantage. Also, if you need a larger "up-rezed" photo to print large, you already can get that from the scan program rather than doing it in post.

Regarding PP and saves, since I use Lightroom, the original scanned photo is never changed. I don't have to save a photo until I want to print it or use it on the internet. Only then do I have to create a file to print or transmit to the internet. When I use to use Elements, I had to do what you did and create extra files. Extra work and more space used in the drive.

What I was attempting to do was have the .tif file untouched and then save it as a jpeg once I PP it.The .tif is so large even saving as a jpeg for further work prior to sending it out for a print and/or using it for the Internet, that the jpeg size was still very large. I was thinking it would be somewhere in the neighborhood of less than 10mb but I'd have to keep resizing after PP to do that. I suppose I just don't have a better handle on this as yet.

When I use the digital camera(Nikon d7000) the size it yields is right at 21mb for a raw(NEF) file as Nikon calls it. Then I'd work on that NEF file to get any corrections I wanted to make and satisfied with the results I'd save as a jpeg file. That usually ended up 3-8mb.

I just scanned some images at 7.4 Gigs. It was the right size for what I was doing for my client. Scanning software is not supposed to yield a perfect scan, only one that's color balanced, and not to contrasty, so that the right amount of contrast can be added later. it's impossible to remove contrast from a scan, the detail is already gone. Scanning tools don't have masking capabilities, not even the real expensive ones, so they are naturally hampered in their abilities. By definition it can't be perfect.

I wouldn't use the levels (at least for the primary corrections) or the white and black points (never). Teach yourself to use curves. If you are going to go to jpeg, there is no need to calibrate the scanner. I would suggest you don't go to jpeg until you are ready to send a copy of your master file (downsized and converted to sRGB off to Mpix or wherever.

Scan in as RGB, use PhotoShop to do your corrections, with curves and hue/saturation adjustment layers. Learn how to mask. Then you can get exactly what you want. It's better if you get your own printer so you can see the results immediately.

Everyone always expects it to be automatic. However, any pro will tell you that color management is a crock, and that these adjustments are quite easy once you learn the basics...

I think this is an improvement.

Francis 007aa by David Fincher, on Flickr

OldWoman011b by David Fincher, on Flickr

Kathy009 by David Fincher, on Flickr

Lenny Why woulodn't you use Levels? I found that after the scan (Epson flat bed V600) the image usually is flat, little contrast and the range is no where near the 0-255. When I use levels, either on auto or manually too bring in the histogram to the edges, the colors bounce to almost normal. The I use contrast, curves etc. What do you suggest?

Alan,

You can take some small consolation in that Levels before a scan is different from using levels after. Digital PhotoLab (the software I use) has a Levels dialog they call Histogram, and I do use it. However, I use it for a very gross adjustment. I don't get anywhere near the edges. I certainly do not use the droppers, which I think are the dumbest invention ever. Brightness/Contrast is also terrible. I ignore these totally.

The reason that Curves work is that its a curve, after all. When you move something on a Curve, everything else stays in an appropriate relationship. This is not the case with Levels, where you can be sure its clipping as you go. A great scan is made when you maintain al the tonal relationships, and expand the midtones, to contradict the natural compression of the device.

I am sure you know that you can add contrast to anything but taking it away is another matter. If there are 5,000 distinctions (just to choose a number) between zone 5 and 6 and you make that area very contrasty, you can cut the number of tone distinctions by 50-75%. You can't get those back by lowering a curve, or removing the contrast with a slider. Once you toss those extra tone distinctions away, they are gone. All great scans are a little flat. That way, the person making the print can add just the amount they want and they have the style of print they want.

Using Levels after the scan, say in PhotoShop, is just silly. Curves are powerful, and will keep everything in place...

Hope this helps.

Thanks for the info I'll have to play with your suggestions.

Alan Klein said:

Thanks for the info I'll have to play with your suggestions.

Click to expand...

You are very welcome.

I should add that I think the Hue/Saturation tool is fine to use for color...

I think Photoshop is pretty hard for the first while. It wasn't designed by photographers. When I first looked at curves I was totally confused. I resigned myself to moving only up and down until I could get the idea into my brain. I finally saw the grid as the zones (not quite exactly) and it started to make sense. Now its totally intuitive... and it gets easier all the time.

Finally, the last thing I'll say on this matter this evening is that there are all these books... and some of them are -- well, no, they are all written for different purposes. However, they often have the phrase "just do this" followed by who knows what technique. I have trained myself that when I see those words my b.s. meter is automatically activated... and I figure most of the time, its probably not as easy as all that, or if it is, its going to destroy things in ways I don't want. All good for a 5x7 print, but maybe not a 16x20...

Hope this helps...