Go Back   Steve's Digicams Forums > Digicam Help > General Discussion

Reply
 
Thread Tools Search this Thread
Old Dec 30, 2002, 11:23 AM   #11
Senior Member
 
Join Date: Aug 2002
Posts: 2,162
Default

jawz.. I think you are way ahead in your understanding, I posted the previous reply to explain that even though you've read the spec, and played with all the little check boxes (like dither/weave etc). These become academic if you can't get the ink properly and consistently on compatible media.

Sometimes, people wonder why they can't get the same result as the hard copy demo prints. You never see the source file, so I bet that's a very large electronic, noise free, (possibly even downsampled HD image) uncompressed bit map or scan. But we're actually in the real user world.

With such clean image sources and tip-top paper I'm sure the effects of res. enhancement may be seen on large printouts. Whether they are repeatable day to day might be an interesting point. With our relatively small compressed image files we're really looking through 'dirty windows' so I wouldn't guarantee you would see every level of enhancement detail.

So, if I was looking for what works best I'd start asking about any test printouts, type and size of files used, capture resolution, type of picture, detail, at what settings they were printed, etc. If they understand their features, they will exploit them to show the best demo prints. Bit restricting though, to only photograph these types of scenes!

However, this doesn't guarantee good results for our JPEGS. In fact, sometimes the opposite. More money spent on higher print res. just opens the window to the artefacts previously hidden by blur, so we then go out and buy better cams! It's already happening with MPEG in TV/dVd.

With some experience of the latter, I can say that deriving a picture source in native resolution, processing native, and displaying native often works better than using conflicting re-sampling and conversion algorithms, although there are fewer problems with stills, than movies. So number crunching, if you don't understand it, isn't always consistent in its results. So why do we do it? Simply because ccd performance, raw memory and cpu speed cost in the product and number crunching can exploit what most of us can't see.

Many printers work well, but after all, they are only based on mechanical components pushing and pulling a clever hyperdermic around and tugging at paper. I don't mean to offend!
voxmagna is offline   Reply With Quote
Old Dec 30, 2002, 4:27 PM   #12
Senior Member
 
Join Date: Dec 2002
Posts: 430
Default

How right you are. <alas> Thanks for sharing your insight.

When I stumbled into the world of digital photography, I found that making the transition from film to digital on the image capture side of things was not a particular problem for me, but I completely underestimated the myriad complexities involved with the image output part of the process. I failed to appreciate just how many variables get involved when going from a captured image file to visually pleasing image on a piece of paper. Most all of these numerous variables directly effect the quality of printed output either by producing artifacts that can be measured or by producing less quantifiable human visual perceptive effects.

Because there are so many variables to get under process control, it would take an overwhelming number of trials to truly optimize the whole process by brute force. Seems like this is a good candidate for a Taguchi style DOE (design of experiment). . . . Sorry, the engineer in me keeps leaking though the cracks.

In the mean time, I shall have to be content tweaking and fiddling and keeping the ink and paper segment of our economy flourishing.

Thanks again.

jeff
jawz is offline   Reply With Quote
Old Dec 31, 2002, 4:30 AM   #13
Senior Member
 
Join Date: Aug 2002
Posts: 2,162
Default

I was thinking we could start a new thread, unfortunately it would contain long debates boring to some.

We could contrast and compare the crude typewriter and printing processes really designed for high volume, current jet printer technology using media from dead trees, with raster based image processes such as scene to film via lens, then remind ourselves how our eyes work and ask the question 'has image printing for photography moved very far forward'?

It's the new Year soon, so I'll just get ready for some digi shooting and accept current technology constraints - whilst hoping for better in the future!!
voxmagna is offline   Reply With Quote
Old Dec 31, 2002, 11:44 AM   #14
Junior Member
 
Join Date: Dec 2002
Posts: 9
Default

Jawz,
Get "Qimage", learn it, love it!

http://www.ddisoftware.com/qimage/
hzxlvf is offline   Reply With Quote
Old Dec 31, 2002, 8:55 PM   #15
Junior Member
 
Join Date: Dec 2002
Posts: 12
Default

Jawz,

When resizing, never "resample". You're just throwing away pixels when you do. In Photoshop, uncheck this option. Change the resolution to about 195. I believe this is as close to 8x10 as you cant get. This just changes the print size, so you will see no difference on the monitor. In effect, you are cramming the same number of pixels into a smaller area when printing, so you get higher resolution output. I don't remember all the technical details as to why this works, but this method made a VERY significant difference printing photos on my Canon F60.

Dan
Dan29944 is offline   Reply With Quote
Old Jan 1, 2003, 6:43 AM   #16
Senior Member
 
Join Date: Aug 2002
Posts: 2,162
Default

Jawz... Try QimagePro. Not re-sampling for printing is OK up to a point. But when sending large image files as email attachments, or web posting, it can be pretty essential.

Just because you don't re-sample in your image file, doesn't mean there's no re-sampling happening, because it just moves to the printer driver software! That's why the printer spooler generates such big files, particularly with a 'dumb' printer (most Epsons) that don't have a big internal memory, fast dedicated raster processing, and hardware resolution mapping.

On a finer point, some might argue, a sophistocated editor, might do a better or worse job than the printer via its dedicated pc driver. If you understood the 'native' printer resolution, all the resolution enhancement and weave stuff could be done by third party software. If Canons do better than PS, then they must be out in front!- to be expected since they designed their print system.
voxmagna is offline   Reply With Quote
Old Jan 1, 2003, 9:37 PM   #17
Member
 
Join Date: Dec 2002
Posts: 39
Default

You've gotten some good advice here but I didn't see anyone take a hard stand on answering the original question.

You wrote:
>will the viewed image quality by better by using lower print >resolution (200 pix/inch in this example) or by forcing 300 pix/in >and resizing and cropping to the 8" x 10" output size

If you resize before printing, the software will interpolate. If you reduce the printer resolution it will also interpolate (it has to if the size is constant).

In the past, high quality software interpolation almost always looked better than what a printer driver could do. Hence one of the reasons for QImage or our software which uses a similar technique.

However printer drivers have gained a lot of image processing intelligence in the last couple of years.

For example, some printers now use advanced edge based interpolation that can be superior to most software.

So the final answer is, it depends on what printer and what software. But the race is closer than it was 1-2 years ago.
lwhitney is offline   Reply With Quote
Old Jan 2, 2003, 3:29 AM   #18
Senior Member
 
Join Date: Aug 2002
Posts: 2,162
Default

But don't most people buy the printer with the biggest claim to dpi, Even though they don't always understand what this means? When you see that window option for 2400dpi don't you select it, whatever raw image res. you send? In other words they are accepting the printer and software will do the work well and are putting their money on it!

Seems a shame that they could have paid less for the printer, and 3rd party software vendors could do just as good a job. But is this only true if those vendors know how to drive the printer to get the best from it, and this info might be with the printer vendor?
voxmagna is offline   Reply With Quote
Old Jan 2, 2003, 10:35 AM   #19
Member
 
Join Date: Dec 2002
Posts: 39
Default

I think you're right voxmagna that most people just "Set it and forget it".

On your second point I think 3rd party software vendors do have most of the info they need to do an excellent job optimize printing, but only a few packages do it.

For example why do PhotoFinale and QImage do it for $35 each, yet PhotoShop does not do it for $600?

I think the problem is that high price software assumes their users want to target a certain resolution, like web design, linotype, etc. so they don't bother.

I assume low price software from big companies don't do it because they assume "those kind" of users don't care.

Ideally it would be nice if printer vendors would partner with software vendors to integrate the best of what can be done at the driver level, with the best of what can be done in software.
lwhitney is offline   Reply With Quote
Old Jan 2, 2003, 12:42 PM   #20
Senior Member
 
Join Date: Aug 2002
Posts: 2,162
Default

You're very right, some of the best packages I've bought with little or no pretence in their marketing, excel at what they do. I have a great deal of respect for software writers who test their packages as Freeware, then move to improved saleable quality product.

That's why Forums like this are worthwhile to users who can extol value (and criticise), based on their experience, rather than merely comment on features and the colour of packaging.
voxmagna is offline   Reply With Quote
 
Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off



All times are GMT -5. The time now is 7:57 AM.