Mike Chaney's Tech Corner

Technical Discussions => Articles => Topic started by: admin on August 28, 2009, 04:45:27 PM

Title: September 2009: Digital Photography Reality Check
Post by: admin on August 28, 2009, 04:45:27 PM

September 2009: Digital Photography Reality Check


Visit just about any online forum related to digital photography and you'll find what amount to "digital gangs" where people hang out and talk about the latest features found on some model camera or printer and how their equipment is better than the equipment used by that "other gang" using a different brand.  Occasionally, one gang member will break etiquette and go visit another gangs territory, sometimes just to "troll" and start trouble.  For some of us who started our passion for digital photography early, we remember the old days of the Nikon Coolpix 990 or the Olympus D600L or maybe even the first "real" consumer camera that I can remember: the Kodak DC40.  Some of us took great pictures back then.  With all the talk online about "14 megapixels just isn't enough, you need 21 megapixels", it makes you wonder how we ever survived back in those archaic times around a decade ago.  I was in a local restaurant lately where one of our lunch buddies sells large (13x19 and larger) prints that hang on the wall, mostly of vintage cars.  They all look like they came from a recent pro camera but when you ask the photographer "What did you use for that one, a Canon 5D Mark II?" the answer you get is "Nope.  I shot that one with my old CP990".  That's a 3 MP camera and some people would swear you could barely get a decent 8x10 at that resolution yet back then, when we had to, we found a way.  We produced 13x19 and even larger prints from those photos with the aid of sophisticated resampling software like Qimage.  Now we turn our nose up at anything under about 10 megapixels.  Seems like now might be a good time for a reality check!



Seems almost surreal that just a decade ago we were working with digital cameras that were pushing the limits trying to reach the 2 megapixel boundary!  Since that time, the "more is better" mentality has driven the digital photography industry and in 2009 our high end consumer dSLR cameras sit comfortably at around 12 to 14 megapixels.  Looking back to what we did with our photos a decade ago and the prints we were able to produce just a few years ago with only 5 to 6 megapixels has to make you wonder.  How many pixels do you really need?  People spend lots of money selling old cameras and buying new ones just to get a few more megapixels and a little less image noise.  What is that really doing for you?  Aside from some pretty significant improvements that are starting to appear in the mirrorless dSLR designs, upgrading your camera every year or two just doesn't make sense.

Sure, you may be able to squeak by with a little lower light or produce a photo that, when examined under a 10x loupe, has a little less noise but what did that do for the overall photograph and was it worth $1000 to upgrade?  Let's do the reality check.  Most of us print 4x6 photos for family snapshots and print some 5x7 or 8x10's for the mantle.  Some of us who are a little better at photography might print some 11x17 prints or 13x19 prints to frame and display on a wall.  Very few of us need enough pixels to fill a billboard nor are we such poor photographers that we have to do extreme cropping to get a decent print.  For those of us who have printers big enough to even print a 13x19, how many pixels do we need?  Well, you'll find plenty of references to the old myth that you need 300 PPI for "photo quality".  To get a razor sharp 4x6 that passes all scrutiny without any resampling when holding the print up to your nose and examining it, that might be true.  Fortunately we have programs that can resample photos automatically and we don't normally walk up to a 13x19 print and examine it with a magnifying glass so we can get away with resolution a lot lower than 300 PPI when we print large posters.  That's how the old CP990 pulls off poster size prints.

In reality, an old 6 megapixel camera running 3000x2000 resolution would produce that 13x19 print at just under 160 PPI.  With resampling, that 160 PPI looks great on a 13x19 print.  A 12 megapixel camera coming in at around 4200x2800 pixels could reproduce the same 13x19 photo at about 220 PPI.  Would the print from the 12 megapixel camera look better/sharper than the print from the 6 megapixel camera?  That depends a bit on the subject matter but in general, yes, the 12 megapixel camera would produce a sharper print at that size.  But here's the question you should be asking: how does that change the overall photo?  Does it look sharper if you walk up to it and examine it closely?  Probably; a little.  Does it look sharper from a distance where people normally view a 13x19 print?  I doubt it!  Does that really deter from the photo, other than those who are looking for pixels or trying to show off a particular camera?

The bottom line here is that more than the number of pixels, noise characteristics of the camera, or camera features, the most important thing is the photo itself.  We're at 12 megapixels and above now on most affordable consumer dSLR cameras.  That's probably enough for anyone printing just about any size photo short of the few people who have specialized needs such as printing large panoramas or billboards for commercial use.  Stop obsessing about your pixels and go take some great photos!  Let the software engineers (like me) obsess over what to do with those pixels so you don't have to.  :-)


Printers, ink, and 16 bit printing

Another trend that is gaining some (albeit slow) steam is 16 bit (per channel) printers.  I've said for years that in 99.9% of cases, the benefit of 16 bit printing is something that will never be seen in practice.  I'm in the business of designing software for photographic printing and I have yet to see an example where someone can demonstrate the difference between 8 bit per channel (24 bit) printing and 16 bit per channel (48 bit) printing other than maybe a slight color shift due to the fact that they are two different drivers with different settings.  Some have tried, but I've turned right around and produced the same print through the normal 8 bit-per-channel driver that was indistinguishable from the 16 bit print.  That said, the manufacturers will continue to tout it and with the release of Windows 7 and what Microsoft calls "HD imaging", you're going to be hearing a lot more about 16 bit printing.  Why?  Because "more is better" and it sells.  Reality check time!

Your monitor is 8 bits per channel (24 bits).  Even when you select 32 bits for your display type, that's still 24 bit color with a mask or alpha channel which doesn't add to the color depth.  Does the fact that your monitor works at 8 bits per channel bother you?  Have you ever noticed any banding or color problems as a result?  If so, they likely came from a source other than the fact that the monitor is running in 8 bits per channel!  Printers are now running 8, 9, 10, or more inks just to try to keep up with the color gamut available on your monitor and that works fine at 8 bits per channel so why do you need 16 bits per channel on the printer?  Well, in reality, there are some areas of color where your printer's color gamut exceeds that of your monitor like yellow and magenta.  To actually benefit from that extra gamut using a 16 bit printer, however, ALL of the following must happen:

  • First, you have to take a photo of something that has super saturated yellow or magenta.  Okay.  Let's say you've found a bright yellow or magenta flower that fits the bill.

  • Next, your camera has to be able to record the colors in that wide gamut.  Camera systems (the sensor plus the processing software) have gamuts more narrow than most printers in the yellow/magenta region but let's just say the yellow/magenta flower you are shooting just happens to have the right shade of yellow to be within the gamut of the photo you captured.

  • Now you have to have taken the picture in raw photo mode.  If you didn't, your JPEG will be tagged with either sRGB or Adobe RGB, both of which are smaller (in some areas) than your printer's gamut.

  • Next, you have to process your raw photo to a large enough color space to hold the color gamut and you must produce a 48 bit image.  Most people use ProPhoto as their large color space (see the next section).

  • Finally, and here's the difficult part to understand sometimes, you must remember that a 24 bit image can cover the exact same amount of gamut as a 48 bit image!  There is no difference in the color gamut that a 24 bit versus 48 bit image can use.  The difference is in the smoothness of that coverage.  So to notice any difference in the prints, the printer's handling of the 24 bit image must be such that banding shows up in areas of slowly changing color like blue skies.  That normally never happens unless you use a color space that is larger than it needs to be in the first place (again, covered in the next section).

The reality check here is that all of the above almost never happen at once.  If you think you need 16 bit per channel printing and you already have a 16 bit printer, just print your photo through the special 16 bit driver that came with your printer and then reprint it through the normal 8 bit per channel Windows driver and see if you can see a difference.  If you find a difference other than some slight color cast or other shift that is likely due to the two drivers having different settings, it is likely due to a problem somewhere else in the workflow rather than the bit depth of the printer driver.

The bottom line for printer is, if you think you need a 16 bit per channel printer or a printer that is advertised as having a 16 bit driver, hang on to your wallet for a while longer.  Most of us have been printing astounding photographs from printers with typical 8 bit printer drivers for a long time.  Don't waste your money on a printer just to get 16 bit printing when you'll likely see no visible difference in your prints as a result.  Of course, there are other reasons for wanting/needing a new printer but you get the idea.


Large color spaces

I'm not going to spend a lot of time on this subject as I've covered it in detail before in my prior article on 16 bit printers.  Some people have started processing all their raw photos into very large gamut color spaces like ProPhoto RGB because they are afraid if they don't, some of their colors will be clipped.  In reality, we've been processing and printing photos from Adobe RGB for many years without worrying that our photos might be "missing" some colors.  With newer wider gamut printers running lots of inks, isn't a wide gamut color space needed?  Simply put, not in most cases.  Again, a lot has to happen to get a "wide gamut color" to go from image capture to print.  You have to have a subject that has a super saturated, bright color.  Your camera and developing software has to be able to process that color: simply stated, the camera itself has to have the gamut to pick it up.  And lastly, the color in question has to reside in the two or three (usually small) areas of your printer's color gamut that exceed a typical color space like Adobe RGB.  I won't even go into the pitfalls of ICM color management and how there are many compromises being made there that affect this subject: that subject alone is big enough for another article. 

Where's the reality check?  Process your raw photos to Adobe RGB developed photos and print them.  Then process them to ProPhoto RGB and print those.  Do you see a difference?  You'd probably win the lottery before you came upon a combination where you could see a difference.  And using a very large color space can actually cause color banding where none would exist when using a smaller color space like Adobe RGB.  Simply put, if you are like most people and you use sRGB or Adobe color space for you photos, 16 bit printing is of no real benefit.  If, however, you are one of the "purists" who insist on using ProPhoto color space for your images, 16 bit printing may actually improve your photos by providing smoother color gradations.   My advice?  Don't complicate your workflow by using specialized (over-sized) color spaces unless you have a specific need to do so.  Most of us don't have such a need!


Summary: Reality check on the reality check!

Okay, I admit, I'm oversimplifying a bit in this article but that was my intention: to bring people back to reality and encourage photographers to get back to what is important and not get caught up in the "mine is bigger than yours" digital gang wars.  The reality check on my reality check is that, while many people may get caught up in the technology wars while forgetting that we are trying to make great photos, I'm not advocating halting forward progress!  Do I think manufacturers should never try for more than 12 megapixels?  Of course not.  Do I think 16 bit printing is useless and manufacturers should quit touting the benefits?  Definitely not!  The reality is that we have equipment now that can produce incredible photos and I want to get it across that these new technologies don't make you a better photographer and often times don't make visibly superior photos or prints!  These technologies are simply stepping stones to better things down the road.  Don't let them stall you from doing what is most important: getting great photographs!


Mike Chaney

Title: Re: September 2009: Digital Photography Reality Check
Post by: Fred A on August 28, 2009, 07:10:44 PM
Excellent article Mike.

I read it and came away with loads of info, but the main point/issue I got out of it was that to get better prints, instead of buying the latest and greatest, I would be better off honing my photographic technique (assuming I had any).

Thanks again!

Title: Re: September 2009: Digital Photography Reality Check
Post by: Seth on August 28, 2009, 08:27:20 PM
author=Mike Chaney link=topic=308.msg1893#msg1893 date=1251477927

For some of us who started our passion for digital photography early, we remember the old days of the Nikon Coolpix 990 or the Olympus D600L or maybe even the first "real" consumer camera that I can remember: the Kodak DC40.

Oooooooooohmygod, you're trying to antiquate me!!  No drugs and you gave me flash backs.  IF you don't mind, I would like to relate how both those cameras(CP990 and DC40) helped "change" photographic history.

Title: Re: September 2009: Digital Photography Reality Check
Post by: Eljae on August 29, 2009, 01:08:55 AM
Thanks for the great article Mike!

Let the software engineers (like me) obsess over what to do with those pixels so you don't have to.  :-)

...and this is why I ordered Qimage this afternoon, as well as the great group on your forum.


Title: Re: September 2009: Digital Photography Reality Check
Post by: Ya Me on August 29, 2009, 12:39:53 PM
Mike, thanks for the great article.

It's an article I'm sure I will read several time!

Ya Me

Title: Re: September 2009: Digital Photography Reality Check
Post by: Keith on August 31, 2009, 01:11:41 AM
Hi Mike,
Your software is just great, and I have enjoyed all the articles read to date. In this case I would like to offer some of my own input that somewhat seems to run against the grain of this piece. For me the constant improvements with cameras, computer software and printers has really helped  noticeably. My reality check has been that even with my limited expertise the photos are now of a much higher quality, the software precesses things much better and easier and the printers do a great job on making up for my short comings. I look forward to improving my knowledge and skills, especially with Qimage, but my argument is that the upgrades across the board are helping very noticeably and I find myself befitting by purchasing the newest technology I can afford. That is my reality check.

Title: Re: September 2009: Digital Photography Reality Check
Post by: photoalbummaker on September 01, 2009, 12:41:35 PM
My first DSLR, the Nikon D70, still serves me well. Since my target print sizes are typically 12x12 or 11x14 or less to fit photo albums I'm under no pressure to upgrade to higher resolution cameras. Eversince I bought my first Epson Stylus IIs, a primitive first generation color printer, I began the quest for an ideal printing vehicle. This proves to be a long challenging journey. Image capture has been made easy with the digital camera, perhaps a little more time consuming with film and scanning, but printing is quite another animal.

After going through several Epsons up to my last two R1800's, I'm now settled with the HP Designjet 30, a dye based printer, for its reliable operation, ie. no clogs whatsoever. And to get the best print from this printer I use Wasatch SoftRIP with a Gretag Macbeth profiling package to print on Monadnoch 100 LB cover paper coated with Inkaid or Golden Digital Grounds (I don't have to worry about a dwindling supply of swellable polymer anymore - there are now only two sources: HP's Premium Plus and Ilford Classic).

So I doubt it if I'll upgrade my camera any time soon, and it looks like I can take a breather in the enduring quest for a printing approach that dries quickly, accepts archival coatings well, and tough enough to endure the clamping and hammering as it goes through my album binding process, and the print engine that does not consume 90% of my time to get it to work properly.

Thanks for sharing your thoughts and for listening to mine,

Title: Re: September 2009: Digital Photography Reality Check
Post by: Terry-M on September 01, 2009, 02:19:25 PM
And to get the best print from this printer I use Wasatch SoftRIP
Have you ever tried using Qimage, you can use the trial version for 30days? A fraction of the cost too  ;)

Title: Re: September 2009: Digital Photography Reality Check
Post by: photoalbummaker on September 01, 2009, 04:28:07 PM
Actually I do use QImage, as well as Photoshop depending on the current digital imaging requirement. I may use the QImage upsampling engine when I need to. Then I import the edited image into the RIP and print in CMYK mode.

If I stay with the HP Premium Plus paper which is designed specifically for the Designjet 30/90/130 printers, I don't need SoftRIP. I like my homemade paper much better for its surface characteristic, and I find the RIP gives me much better output due to its ability to linearize and control ink limit for each CMYK channel.

So I need the RIP to create CMYK profiles and then only for the printing phase. And yes, I agree that QImage is very affordable for what it can do.


Title: Re: September 2009: Digital Photography Reality Check
Post by: bgrigor on September 06, 2009, 12:31:25 AM
Thanks Mike, very well said and so true. I am a digital artist and I also print images and do art reproduction for other people for a living. I come up against the "300 ppi" myth all the time. Yet, in practice, I have produced amazing prints on canvas from point-and-shoot cameras at effective resolutions as low as 79 ppi. By "amazing", I mean I didn't expect it to look as good as it did and the client was delighted. And those prints were done with Qimage.



Title: Re: September 2009: Digital Photography Reality Check
Post by: rpcohen on September 16, 2009, 02:30:43 AM
Great column Mike and good reality check.  My first digi-cam captured 1.5M images which produced very nice 8x10 prints until I learned that such a thing wasn't possible without many, many more megapixels.  :)


Title: Re: September 2009: Digital Photography Reality Check
Post by: jbhaber on September 25, 2009, 08:47:50 PM
I've been printing up to 20x30 images from my 6mb Canon Rebel XT for years. Been winning camera club awards and have been selling some images, too. Use Qimage for up to 13x19, and send out for the larger ones. Lots of complements, and not one complaint!

Thanks for the great utility, Mike!

Title: Re: September 2009: Digital Photography Reality Check
Post by: Fred Too on September 26, 2009, 02:09:24 AM
Well said, Mike! I've been doing this for a long time and agree strongly with your position. I like to stay "current" with my camera but know that this really isn't necessary but it's still fun!
It's a different issue, but I love your new Flashpipe software. The convenience of sending photos to my networked computer in the back room is great and I feel this convenient "backup" my save my neck one of these days! Many thanks for your work on this.

Title: Re: September 2009: Digital Photography Reality Check
Post by: hedwards on October 03, 2009, 01:30:11 AM
Nice article, I think that it points out a few things which have been evident for a while, if overlooked. 6MP is indeed enough for many purposes, and except for those that really need interchangeable lenses, a typical P&S is probably just fine.

The other thing though is that it's not just that more pixels aren't necessarily necessary, it's that we're going to the point now where the sensors are over resolving what the lens is capable of providing, and not just on cheaper lenses, mind you, but at some point we're probably going to hit the point where anything less than L quality is going to not really make the cut.

But then again, I'm sticking with my 10D because I really need better glass and by the time I upgrade my body, the much more important issue of noise on the sensor is going to be so much better than it is today. Even now I've seen some ISO 1600 images which were to my eye usable, not necessarily perfect, but somewhat better than my ISO 400 images off my current gear.