Camera megapixel ratings - are we being fooled

[shawnstam]

spend money on lighting, buy the best possible it will help you take better pics and if you buy really good stuff you can take it to you next camer setup.

 

[Puffer Fish]

very good program (I use it in windows and linux) but with two issues:
1. doesn't support colour profiles for your devices
2. doesn't support raw

other than that it's the best you can get for 0$ (and much more)

Spent this morning downloading and playing with Gimp 2.2.

The color profile issue is easy to get around - and if you have the money to buy the profiling equipment, getting software should not be a cost concern. If you are working in one format - say sRGB for windows and take pictures in sRGB, calibrate your monitor in sRGB and print in sRGB, then the specific software is unimportant. What messes up the whole industry is attempts at changing from one to another. As you are always limited to the smallest color space, everything else is pretending.

Note: Buying a printer with a bigger color space than sRGB is fairly worthless, buying one that can actually do sRGB is not, but no one wants to advertise the latter.

For free, this is nice stuff and better bang for your buck than many home use softwares. Unfortunately, it sucks from a learning standpoint. A couple of examples:

1. Unsharp mask has a preview, but the preview cannot be zoomed. So you can see the big effect, but not what is happening at the pixel level.

2. There are several very nice color adjustments, but to really understand what is going on, you need to be able to see both the pixel distribution change and how it effects specific pixels. That is missing.

I would still strongly suggest the Corel's Paint Shop Pro X is one really great deal. If the $80 or so cost is too much, then this would at least work and free is always nice.

If you just want to resize, crop and do general stuff, Gimp works really nice, just don't expect it to actually get the best pictures.

 

[Scubatooth]

puffer fish

wouldyou be referring to the guy who is doing the gigapixel project? because he is using a U@ spyplane camera and lenses shooting onto large format film (9.5 x 13 sheets of kodak films)and then scanning and stitching

heres the link http://www.gigapxl.org

 

[Nostromo]

A good example is Adobe CS, if professionals actually knew how pixel information was handled inside of that program, they would never use it. They don't, and it has been better than any other option, so everyone uses it.

As a longtime user of Photoshop, that statement really caught my attention. Can you be more specific?

thanks,
e.a.e.

 

[ScubaLuke]

Like a darkroom, only a 100 time better and you can correct your mistakes for free (unless you printed it). You can do things that no mire mortal should be able to -just takes some learning and practice. Compared to the cost of a darkroom, this stuff is cheap. Good luck.

I've been taking a look at printers today and so far from what I've read I like R1800 best. I saw an interesting of what to do with my old 1280 - convert it to a B&W only printer. There's at least one company that makes a B&W cartridge for the 1280.

Bach to the research!
Luke

 

[Puffer Fish]

puffer fish

wouldyou be referring to the guy who is doing the gigapixel project? because he is using a U@ spyplane camera and lenses shooting onto large format film (9.5 x 13 sheets of kodak films)and then scanning and stitching

heres the link http://www.gigapxl.org

No - this was to show the value of merged images (and make the largest high res image ever made). Used a standard Nikon DSR (if I remember correctly) and a heavy duty, lockable tripod.

He just took hundreds of individual, slightly over lapping images and merged them into one giant image. Was using a 150mm or so lense, so you had the perspective of a view camera, with tremendous detail.

Used to have the link on my machine, but had to rebuild it 6 months ago and that was one of the lost items - will see if I can find it.

The only special thing was the software, because normal ones do not allow that big of an image.

This concept was though to be the answer to a great number of issues with digital camera's, but the technology has never been used very effectively. Studio work was going to be the product of a high resolution closeup stuck on a high resolution background. Has not happened.

 

[Puffer Fish]

As a longtime user of Photoshop, that statement really caught my attention. Can you be more specific?

thanks,
e.a.e.

You get my attention to detail award:balloon

First let me applogize for highjacking the thread and second for all the technical speak.


Adobe CS states that their native color management is in Lab (L=light/dark, a = red/green and b = yellow blue). Camera's, computers and printers use some sort of dot management (RGB, for example).

There are a bunch of issues with even using Lab (this one will bring a cry of foul from a lot of people in the industry - you have to actually look at the math to understand it) - but Lab does not describe a unique set of colors. Each value is the result of averaging a bunch of readings. Using three averages means that that there are several colors that could have the same reading.

The above is not true with RGB - each combination is unique. This issue was discovered by CIE and they released C and h (chroma and hue), which are unique. Problem was, no one could understand what they meant. Color measurement gets stuck between what is easy to understand, but almost worthless and something that works, but not understandable.

This site, recognizes the issue, but only understands part of the problem:

http://www.aim-dtp.net/aim/evaluation/cie_lab/3de6_index.htm

Along comes CMC, and they do a Delta E calculation with LCh, and it works. CIE then redoes their calculation, including LCh in Delta E 2000 and it works. However, no one wants to really tell everyone what the big issue is, because a lot of sales are tied up with this.

Lab is also non-linear, but it is very difficult to tell which of the two issues one is looking at.

Back to Adobe - or any management software. What you need is a single management format. If that is sRGB, then it should be sRGB to sRGB to sRGB. Anything else produces false color or loss of data. One of my favorites is adobe color space, with a monitor that only can show sRGB - wonder what people think they are seeing?

Regarding management - there should only be absolute, absolute, absolute mapping. Anything else is just silly. If you want to increase dynamic range, do it with software.

The next problem with Adobe is the scale issue. Some of their tools use 0 - 100. Some 0 -256 Some have both positive and negative values. There is only one - it is 0 -256 for each channel (unless you have some very special equipment). This one I have talked to them about, and they admit that different sections were done by different groups, and they don't match.

As I am working in the under .3 Delta world and photographic people are talking +/-3 units typically, some of what I see is not something they would. Of interest is that RGB is a scale that if done accurately is under .1 Delta E 2000 or CMC. Nothing wrong with the scale, just the processing.

Note: .3 is the limit an average human can detect, so RGB is not a bad scale.

As you are only as good as your weakest link, pretending that somehow this is not true is really silly. Adobe, long ago, should have adopted a clean work flow system. Not doing so means a lot of errors are introduced that do not need to be there.

A somewhat easy to understand example is just Adobe RGB. The following compares it to a much larger gamut:

http://www.naturephotographers.net/articles1203/mh1203-1.html

But there are several problems with all of this:

1. Even if you could get a monitor to display the range, the steps would all be off (0 -256, just with more space between them)

2. Getting the mapping done on the printer requires some very expensive software, and then you would never know what an image would look like, until it printed. That means doing lots of reprints.

Sadly, I hate sRGB, but it is just about the only scale one can manage from start to finish. There is a place for other gamuts, but not in the normal world.

At least Corel uses the same scale throughout.

Once you see the results of a clean flow, most people are shocked at how much of the issue is the processing and not the actual accuracy.

It should be noted that there are a lot of accuracy issues, but they can be corrected in a clean system easily.

Hope this makes sense.

 

[Puffer Fish]

I've been taking a look at printers today and so far from what I've read I like R1800 best. I saw an interesting of what to do with my old 1280 - convert it to a B&W only printer. There's at least one company that makes a B&W cartridge for the 1280.

Bach to the research!
Luke

You are doing good grasshopper - remarkable technology. I'll bet you will be telling me where to look in no time. Just do a quick math check on the number of dots they are actually putting on a piece of paper.

You did not ask for this, but get one and try the following paper:

http://www.dtgweb.com/shop/product.php?productid=23434&cat=333&page=1

This was the best short term color stable paper I have tested (short as in years), and has almost as good a gamut (color range) as high gloss. You can turn off the gloss optomizer, which saves money. Under glass, expect around 100 years or so of reasonable color accuracy.

I have several family members that we split a box up, to keep the cost down.

You can try something very close in the Ilford "Printasia" paper that is available from the retail market. It appears similar, but not quite identical.

 

[Nostromo]

Hope this makes sense.

I really should be careful what I ask for, I just might get it. Thanks for the explanation, I even understood small parts of it.

The reason I ask was because I did a lot of photography a few years ago and used Photoshop heavily during that time. I always had color management problems, but just assumed it was my fault. Well, a lot of it probably was my fault, but I'll take every available opportunity to blame the equipment. To give you an outline of my workflow, I was scanning transparency film (Velvia) via a nikon LS-2000 into Photoshop at 16 bit per channel in Adobe RGB 1998 color space. This was Photoshop 5.5 so I did my color correction in 16 bit and then had to drop to 8 bit for most everything else. Display was on a set of Apple studio monitors. Output was typically to an Epson 2000p archival ink-jet. I never did get things to where I was happy with them, and yes there were a lot of reprints. Often fifteen or more for a given image. Isn't attention to detail a ***** sometimes? In addition to everything else the Epson 2000p had issues with this weird greenish cast that would show up in certain situations. I figured it was a Colorsync issue, so I contacted Epson for a new file. I don't have the email anymore, but the reply just about made me fall out of my chair. They recommended I open the curves, select the green channel, and reduce the amount of green. I miss doing the photography and digital editing, but I sure as hell don't miss the color management issues...

Anyway, thanks again for the very interesting lesson, and I also offer my apologies to the group for the hijack.

e.a.e.

 

[Puffer Fish]

I really should be careful what I ask for, I just might get it. Thanks for the explanation, I even understood small parts of it.

The reason I ask was because I did a lot of photography a few years ago and used Photoshop heavily during that time. I always had color management problems, but just assumed it was my fault. Well, a lot of it probably was my fault, but I'll take every available opportunity to blame the equipment. To give you an outline of my workflow, I was scanning transparency film (Velvia) via a nikon LS-2000 into Photoshop at 16 bit per channel in Adobe RGB 1998 color space. This was Photoshop 5.5 so I did my color correction in 16 bit and then had to drop to 8 bit for most everything else. Display was on a set of Apple studio monitors. Output was typically to an Epson 2000p archival ink-jet. I never did get things to where I was happy with them, and yes there were a lot of reprints. Often fifteen or more for a given image. Isn't attention to detail a ***** sometimes? In addition to everything else the Epson 2000p had issues with this weird greenish cast that would show up in certain situations. I figured it was a Colorsync issue, so I contacted Epson for a new file. I don't have the email anymore, but the reply just about made me fall out of my chair. They recommended I open the curves, select the green channel, and reduce the amount of green. I miss doing the photography and digital editing, but I sure as hell don't miss the color management issues...

Anyway, thanks again for the very interesting lesson, and I also offer my apologies to the group for the hijack.

e.a.e.

Was not your fault - I started with a 2200 and a very similar setup. Nothing you did would have worked. Green was most likely fluorescence - a scanner issue that exists today. Clean the flow, use the same gamut throughout and don't use a scanner and everything is one time. Add the scanner and sometimes you will have the blue green/green/blue issue show up. It is real, by the way, and no way to control.

Figured you were not an Adobe zealot - they think there world is perfect.

You were going through so many translations, no telling what the final would look like.