zapped

J.Reynolds wrote: "I've read in many places here on PN that down-sizing and sharpening is best done in several stages."

 

I've heard recommendations to *up*size iteratively (110% at a time, typically), but not to downsize in steps. And sharpening is best done at the final print dimensions, not "in several stages".

 

Bicubic or 'bicubic sharper' will down-scale for you in one step. No need to iterate.

 

Also: "How do you decide what dimensions to crop to, before starting this progressive down-scaling?"

 

Whether you exclude the film holder edges by cropping in the preview pan of your scanner software or do it later in Photoshop is irrelevant. Personally I'd scan the whole thing to Tif then use 3:2 rectangular marquee to get the crop you want. After you save that as your "master" scan, you can crop to 8x10, 5x7, or whatever aspect ratio you need for printing.

 

There is no need to be "mindful ... that the pixel count in both dimensions must be divisible by the down-scaling factor".

Glad to help, Stephen. If you want an inexpensive cutter that does a fine job, try the one I use ---

 

http://www.logangraphic.com/products/kits/matcutting.shtml

 

This gives you the Logan 524 Ruler (guide), Logan 302 cutter (head), Logan 500 mat knife, and some other do-dads for a lower price than buying the 524 + 302 + 500 separately. In the US, I used the standard 40%-off single-item coupon from Michael's Crafts to really save a bundle.

 

Have fun,

No Stephen, 3:2 is not a standard matte size. I cut my own with an inexpensive Logan matte cutter (preset to a 45degree bevel), then use a larger frame size that looks decent. I do try to use standard frame sizes, though, because it's substantially cheaper than custom framing. Since you'll always have an asymmetric placement, try to leave the "extra" mat area at the bottom, and prefer standard frame sizes where the excess height is greater than the excess width.

 

For example, let's say you have a 10x15 (approximate) opening and it's portrait-oriented: 10" wide, 15" high. You're considering either a 16x20 or 18x24 frame. With the 16x20, you have 3" excess horizontally but 5" vertically. It would look odd to have 3" borders on the bottom, left, and right, and only 2.5" on top. Same if you tried to leave the 2.5" on the bottom.

 

With an 18x24, though, you have 8" extra width and 9" extra height. So you could have 4" on the top, left, and right, and the larger excess amount (5") on the bottom. Much better balanced.

 

Sketch the combinations & this will be a little more obvious.

"I have some digital images that I'm going to have printed at 10x15 inches on the Noritsu at my local lab."

 

Why 10x15? Is that the largest they make at your local lab, or did you already decide this was the matte size?

 

"As I know very little about matting and framing, the first question is should I leave a border ( like 1/4" ) when I size the image so the matte fits on top of the paper ? Or no border, or a larger one ?"

 

First you have to ask yourself a few questions. Are you cutting a custom matte? Do you want the image set inside the matte (with white paper showing inside the matte) or do you want the matte to completely cover the image?

 

If you want the image set inside the matte, I'd leave between 1/4" & 1/2" on all four sides exposed, then another 1/8" minimum hidden underneath the matte. That means 15" - 2 * (1/4" + 1/8") = 14-1/4" would be your largest image dimension set inside the 15" canvas. Or 15" - 2 * (1/2" + 1/8") = 13-3/4" if you want the 1/2" white border. Either way, you'd want the long dimension of the matte to be 14-3/4".

 

For the large image, you'd go to Image -> Image Size, TURN OFF "Resample Image", then set dimensions to 9-1/4" x 14-1/4". Then you'd do Image -> Canvas size to 10" x 15" (not relative), with white for the selected color. No need to change the pixel counts in your image; the lab should handle that for you. It's a whole 'nuther tutorial to get into printer resolution & color profiles.

 

Now, if you want the image completely covered by the matte, just leave it at 10"x15" but understand you'll lose 1/8" all around if you cut a 9-2/4" x 14-3/4" matte opening.

 

If you're not cutting a matte yourself, you've got to crop & resize your image to the standard matte sizes / opening sizes. Make the same calculations to figure the image size.

 

"Second question is that in the print menu for PS CS I don't see a paper size of 10x15 so is there a way to add that size to the menu ?"

 

You're not printing, the lab is. This is irrelevant.

I'd agree with Scott but add that it's the Fuji Crystal Archive GLOSSY that makes a great combo. Their matte paper makes me cringe.

Camera on tripod, remote release. Meter for ambient, f/8, even if the shutters is 5 seconds or more. Light only the kiddos with the strobes. Time exposure will pic up the rest of the pool.Even better if the ambient light wants 15 seconds or more, you could tell the kids to walk out of frame as soon as the flash pops.

 

Since it's film & you can't check it immediately, also bracket.

 

Lastly, if the color temperature difference between the flash-illumination & ambient is too whacked, take another expoure with no subjects, color-adjust that one, then composite digitally.

Joey's advice is absolutely correct. I personally run Photoshop CS on my tired old office computers (PIII/800, 384MB RAM, 60GB HD) but the difference on my home machine (Athlon XP 2800+, 1GB RAM, 120+250GB HD) is like night & day.

 

In addition to the idea of buying a bare-bones model, you could assemble an excellent machine for no more than $500 if you're willing to steal some of the parts from your old machine. If you can't use the old case & power supply, a new mini-tower & p.s. is only $50-75. Use your old floppy, CD reader, graphics card. Pick up a 2.4-2.8GHZ P4 or a 2400-2800 Athlon XP and an appropriate motherboard, 1GB DDR333-DDR400 RAM, a large HD, and a copy of WinXP Home, and you'll be a very happy camper. Later on get a DVD burner (< $75) for backup.

Thanks, D.B., I did a photo.net search & found this great writeup on DMax & dynamice range, which I'll share for other interested folks ---

http://www.photo.net/learn/drange

Dancing Barefoot --- I'm a D-SLR shooter not familiar with the technical details of scanning slides & film. Could you explain whether you consider Dmax=3.3 to be good, or whether you're implying that even if they claim 3.3 it might not really *be* 3.3?

I stopped considering using fragile hard-drive-based backup storage when 1GB CF cards hit $50. Buy as many as you need at that price!

Lei Liu wrote: "I build computers for my whole life until i bought a dell 2 years ago. ... Get a Dell refurbished P4 2.8 with 512mb for about $350".

 

You're going to have to back THAT statement up with a link. I'm browsing the dell outlet inventory right now & the cheapest refurbished desktop I see is $454 for a Celeron 2.4G.

 

With Dell you'll get a brand-name CPU but the rest of the system will be the lowest-cost, lowest-reliability parts that can be obtained. Underpowered power supplies, sometimes non-standard connectors on the mobo to prevent upgrade, "remaindered" hard drives, marginal system RAM, etc. And the warranty's not that great unless you pay big bucks to extend it.

 

My point about building it yourself is that you can choose sensible mid-range components, and since you built it yourself you won't need to call anybody when you need to swap out a component in the future.

Earlier I wrote: "Here are the times for the individual operations on my tired old PIII/800, FWIW: Increased image size 577% (to 300MB). = 52 seconds. Increase saturation 10%. = 118 s. Added S-Curve (64->56, 194->202). = 142 s. USM (100%/0.8r/2t). = 148 s. Canvas size 10 pixels. = 54 s."

 

I recorded these operations as an action, with history-state cleared after each operation, and ran that action on the same 2048x1536 landscape on my home machine.

 

The machine is a homebuilt Athlon XP 2800+, Abit NF7 motherboard, 1GB DDR333 RAM, WD UltraATA/100 IDE drives. Times were: Increased image size 577% (to 300MB). = 6 seconds. Increase saturation 10%. = 26 s. Added S-Curve (64->56, 194->202) = 3 s. USM (100%/0.8r/2t). = 9 s. Canvas size 10 pixels. = 22 s.

 

Just an FYI,

Don't worry about the typo, but I'm assuming you can't mean what you typed - "I've just resized the same image again from 2048 x 1326 to 11378 x 8534 (277.8MB)". 2048 x 1326 is a different aspect ratio than 11378 x 8534. You probably meant 2048 x 1536?

Numbers like "30 seconds" and "47 seconds" don't mean too much unless we're willing to be precise about the task at hand. I'm not at my home machine (Athlon XP 2800+, 1GB DDR333 RAM), but I thought this was interesting enough to try on my tired old work computer (PIII 800MHz, 320MB RAM).

 

An example of a "9MB image" to start with would be 2048x1536. I chose one that I had lying around, a moderately complex landscape, and did the following operations on it:

1) Image resize, bicubic, 577% (that makes it almost exactly 300MB).

2) Increase saturation +10%.

3) Added a gentle S-Curve (64->56, 194->202).

4) Applied a USM filter (100%/0.8r/2t).

5) Increase canvas size by 10 pixels (the "black line").

 

Here are the times for the individual operations on my tired old PIII/800, FWIW:

Increased image size 577% (to 300MB). = 52 seconds.

Increase saturation 10%. = 118 s.

Added S-Curve (64->56, 194->202). = 142 s.

USM (100%/0.8r/2t). = 148 s.

Canvas size 10 pixels. = 54 s.

 

On my old machine, the saturation, curves, and USM operations seem to reach "done" in about half the time I quoted - the remainder is disk spinning while history is saved. I timed from start until my cursor was working again.

 

At home I'll try this on my faster machine, just for grins.

John, I'm not a technogeek who advocates spending thousands of dollars on the newest/shiniest desktop setup, but I really wouldn't recommend stuffing 1GB of memory into your P2/266 machine, even if the motherboard can handle that much memory.

 

For one thing, you'll be spending money on new memory that can perform much better when installed in a more modern motherboard. In the P2 system it'll be running much more slowly. The same is true for the other subsystems - you could buy a bigger hard drive, but disk speed will be slower due to the old motherboard. At least I'm assuming that your P2 system's IDE ports are not running at ATA100 speeds - 33 or 66 at best. What I'm saying is that each upgrade will be somewhat crippled by the now-outdated motherboard/cpu you're hanging onto.

 

You don't need to buy very high-end, but you should consider a mid-range Athlon XP (2400-2800) or Intel P4 (2.4-2.8 GHz) solution. Build it yourself & it could easily come in under $500-600 for a decent system, including new motherboard, cpu, 1GB DDR ram, 80-100GB H.D., case + powersupply, and a low-end vid card. Add another $100 for WinXP Home and you'll have a solid system for your Digital Darkroom.

You've got nice exposure control and color saturation in the straight color shots. The B&W & duotones are also quite satisfying. Really like the simple gallery presentation - black background, classic white border around photos, no drop shadows. Great casual _and_ posed portraits.

 

My favorite image:

http://members.optusnet.com.au/raymany/bailie/arliah/041A.JPG

 

Thanks kindly for sharing 'em.

Marc - what a great set of shots from your <a

href="http://www.photo.net/photodb/folder?folder_id=432366">September

Weddings</a> album! My favorite was your "Running Woman". Love the

quiet composure of the kiddos contrasted by the frenetic gal in the

background.<br>

<br>

<hr width="100%" size="2"><a

href="http://www.photo.net/photo/2720413"><img

src="http://d6d2h4gfvy8t8.cloudfront.net/2720413-lg.jpg" title=""

alt="M.Williams "Running Woman""

style="border: 0px solid ; width: 511px; height: 512px;" hspace="10"

vspace="10"></a><br>

<b>Marc Williams "Running Woman"<br>

</b>

<hr width="100%" size="2"><b><br>

</b>Thanks for sharing those with us.

Yes, the extra filesize required to work with 16bpp is worth it if you're making significant levels/curves adjustments to the original image, even if the final web-sized version is only, say, 800 pixels wide. If you're only doing minor tweaks to a basically decent exposure, 8bpp is adequate.

If they can see it on the screen, they can always take a screenshot.

Mr. Wegwart wrote: "Woah! that didn't come out too pretty on the screen. Does using AdobeRGB 1998 color do that?"

 

I saved your image, loaded it in PS/CS, saw the aRGB tag, converted to sRGB, and then checked both images out in a non-color-managed app (ACDSee). There's a pronounced deepening of the red- & peach-colored flowers, a deepening of the bride's OOF skin tones, and a little boost to the green leaves in the foreground. Not really *that* huge a difference though, to my amateur eyes...

Does he still owe you the round if the link is broken? :P

I'm just a humble amateur, but given the number of galleries I've

perused over the years I like to think I know good art when I see it.

Your gallery is simply breathtaking, Don. These are three of my

favorites:<a href="http://www.donboyd.com/gallery01/image04.htm"> "Kiva"</a>, 

<a href="http://www.donboyd.com/gallery01/image02.htm">"Snow on Agave"</a>,

and <a href="http://www.donboyd.com/gallery01/image05.htm">"Maple Tree"</a>.<br>

<br>

Thanks for sharing them with us!

I've got a couple of soft pouches with drawstring tops that came with eyeglasses or sunglasses I've purchased. They double as polishing cloths for the glasses so even though I wouldn't polish camera lens with 'em, they're soft enough for carrying my Angle Finder B.

PhotoBob wrote: "If her skin tone is darker than middle gray/Zone V, won't the straight reflective reading make her skin lighter than it really is? So if the skin is more of a Zone VI, then wouldn't I underexpose?"

 

I agree you'd need to underexpose if you metered reflective light off darker skin, and I agree with Brandon's reasoning. But isn't Zone VI *lighter* than Zone V? Zone VI is the general range of Caucasian skin, as I understand it.

 

Perhaps you meant her skin was more in the Zone III-IV range?

 

See also: http://www.luminous-landscape.com/tutorials/zone_system.shtml

I shoot digital, but I did find Las Vegas at night was fairly easy to

capture at ISO400 hand-held. Perhaps you could take along your favorite

400-speed film and give it a whirl.<br>

<br>

This shot was taken inside Paris at 5am, when the casino was completely

empty (<i>click the image to link to a larger

version + EXIF</i>):<br>

<a href="http://www.pbase.com/pricklypear/image/28242473/original"><img

src="http://www.pbase.com/pricklypear/image/28242473/medium.jpg"

title="" alt="" width="400" height="270" hspace="10" vspace="10"

border="0"></a><br>

<i>Exposure: 1/60s f/2.2 at 28.0mm iso400</i><br>

<br>

<br>

An exterior also at iso400 (<i>click for larger version + EXIF</i>):<br>

<a href="http://www.pbase.com/pricklypear/image/28242468/original"><img

src="http://www.pbase.com/pricklypear/image/28242468/medium.jpg"

title="" alt="" width="400" height="270" hspace="10" vspace="10"

border="0"></a><br>

<i>Exposure: 1/79s f/2.8 at 28.0mm iso400<br>

<br>

</i>Generally I worked in Manual exposure

mode with an f/1.8 prime (my Sigma 28mm EX), and metered for an

underexposure of -1 or -2 stops, because I wanted to preserve the

night-time "feel".<i><br>

</i>