Megapixels...

[PJL]

@gsgary - don't be a douchebox.

I realized right after I posted the reply last night that the eos 1 line was 1.3x cropped. My bad.

And technically, I would need to "get [my] FACTS right."

Right...the 1DS line is full frame. You could also get a 5D mk 1 for around $700.

 

[JTPhotography]

The answer to the question is no. And it is all relative to the size of the image in question. If you're viewing thumbnails, not much difference between 8 and 36 megapixels. But if printing 30x40s for a gallery, 36 megapixels will look like crap against a 200 megapixel (if one existed).

 

[lennon33x]

Then let me ask this...(this is purely in relation to the replies to my original question)... Which format maintains the closest resolution to the RAW file (i.e. JPG, TIFF, etc)? And can you ever improve resolution in post (I'm assuming the answer is no, much like compressing a wav file in music to an mp3, where there is a loss of quality).

 

[robbins.photo]

Easy killer. I meant that exposure settings - aperture, ISO and shutter speed (and therefore the info contained in exif info) are used in IQ. If an image is underexposed because one of the components of the exif info not correct, then it has everything to do with the quality of the image.

Regardless, I'm looking for a general rule of thumb that says "any number of megapixels above [x] megapixels will not drastically increase your IQ"

Well a lot more goes into image quality than just the MP rating of the sensor, so I'm afraid you can't really boil it down to such a simplistic equation.

 

[robbins.photo]

There is no such rule of thumb.
Exif is just information about the camera, camera settings, and other information about the photo.

MP count relates to image resolution. The more pixels there are the more detail a digital image can record.
Lenses also have a resolution limit, and with fewer MP the image sensor is the limiting factor, not the lens.
At some point an image sensor has enough pixels that the lens becomes the resolution limiting factor.

A digital camera processes the information the image sensor records in several ways before we get to see an actual photograph.
Most of the gains in image sensor low light performance are a result of better in the camera software rather than increasing MP counts.

Camera makers use CMOS image sensors instaed of CCD image sensors for several reasons:
CMOS image sensors cost less to make
CMOS image sensors use less power so Li-ion batteries last longer between recharges.
Because CMOS uses less power they generate less heat, which reduces thermal noise in digital photos.

Note that astronomers make images of very faint objects and use CCD image sensors instead of CMOS image sensors.
But astronomers are making long exposures of several hours each, and to combat thermal noise cool their CCD image sensors with liquefied gas.
Nitrogen is a liquid between -196°F and -210°F.

So, if someone Is shooting with the Nikon d800 (I think it's the one with the monstrous 36 mp image), it then limits itself to the resolution limit of the lens it's shooting through?

the main reason I ask is that I want to switch to a low budget full frame (like the EOS 1D II) from a Rebel T3. The Rebel is right around 11.5 MP, but the 1D II is 8mp but full frame. I want to make sure that I'm getting the same, if not better quality, even though the 1D is much older.

Ok, well the EOS 1D actually has better image quality than the T3 - roughly about 10% better on average thanks to it's superior color depth and better dynamic range.

 

[TCampbell]

You can "interpolate" images in post to increase resolution, but you can not actually improve the detail resolving capability.

You can also apply sharpening in post to create the illusion that you've improved resolving power -- in reality you have not because there are limits to how sharpening works.

If I have a barcode pattern and I'm close enough to the barcode that you definitely can see some what space between each black line, then even if the edges of the line are fuzzy, I can apply sharpening and make the barcode look a bit better. But this only works because there is enough spacing between the stripes that we can clearly resolving one stripe from a neighboring stripe. If the image quality were not good enough to clearly resolve that gap between the stripes (say we position the camera farther away), then the stripes and the space between them no longer can be fully resolved... and at this point the "sharpening" process breaks down and can no longer save you.

Here's a good article: LensRentals.com - Have You Seen My Acutance?

Here's another: Diffraction Limited Photography: Pixel Size, Aperture and Airy Disks

BTW... there is also a limit based on the physical diameter of the lens (and this is another limit based on the assumption of theoretically "perfect" glass - so it's not a question of lens quality) and it's called Dawes' Limit.

Both diffraction limits (based on apertures) and Dawes' limit (based on physical diameter) are dealing with the wave nature of light -- properties at the quantum level -- and so these are laws of physics that lens makers and camera makers can't really side-step.

But the bottom line to your original question about some rule of thumb -- the closest to a rule is that it really depends on how you plan to present the image and the distance at which you expect your viewers to stand when looking at it. The image will appear to be high resolution when the human eye can no longer resolve the individual pixels. How many "dots per inch" will really depend on how far away you are standing while looking at the image. Billboards can be SHOCKINGLY low dpi (I once read they can be as low as just 11 dpi) -- but it turns out that works when you expect your viewer will not be closer than a few hundred yards (so clearly the viewing distance is a huge factor here.) We tend to not use most of our data when using images.

 

[KmH]

Note that astronomers make images of very faint objects and use CCD image sensors instead of CMOS image sensors.
But astronomers are making long exposures of several hours each, and to combat thermal noise cool their CCD image sensors with liquefied gas.
Nitrogen is a liquid between -196°F and -210°F.

Oh man I wish. I'm making do with a 60watt peltier cooler on the back of my CCD. I think I need an upgrade. :lmao:
Actually I want my own space telescope.

I should have mentioned amateur astrophotographers use cooling systems that don't get nearly as cold as LN.

As far as a space telescope, make it small and light and contact this guy :thumbup: -

 

[Light Guru]

The answer to the question is no. And it is all relative to the size of the image in question. If you're viewing thumbnails, not much difference between 8 and 36 megapixels. But if printing 30x40s for a gallery, 36 megapixels will look like crap against a 200 megapixel (if one existed).

Very good point. The OP has not said what he plans on doing with his images. If it's just for web viewing then sure a older full frame would be fine but if he wants to make large prints I would not go with a older lower megapixel camera.

There actually are cameras that you can get a 200 megapixel image from and they have been around longer then anyone on is site has been doing photography. A quality scan of a large format negative specifically 8x0 negatives will easily top that 200 megapixels.

Then let me ask this...(this is purely in relation to the replies to my original question)... Which format maintains the closest resolution to the RAW file (i.e. JPG, TIFF, etc)? And can you ever improve resolution in post (I'm assuming the answer is no, much like compressing a wav file in music to an mp3, where there is a loss of quality).

Well there are several types of JPEGs and TIFFs and the amount of compression can vary. That said a uncompressed TIFF will the the best option.

 

[480sparky]

As far as a space telescope, make it small and light and contact this guy :thumbup: -

Pffft.

 

[lennon33x]

Theoretically, the largest I would print would be about maybe a 8x11 or the occasional canvas print (it seems as though the canvas gives a decreased resolution regardless of the image quality). Most images would be internet or the standard 5x7 or 3x5

 

[gsgary]

Theoretically, the largest I would print would be about maybe a 8x11 or the occasional canvas print (it seems as though the canvas gives a decreased resolution regardless of the image quality). Most images would be internet or the standard 5x7 or 3x5

The camera you have is ok for what you want, i own all the cameras mentioned 1D, 1Dmk2, 5D for me lens quality is more important

Sent from my GT-I9100P using Tapatalk 2

 

[Derrel]

I'd say the answer is probably 42. Forty-two megapixels. Anything more would be a waste.

42 (number) - Wikipedia, the free encyclopedia

 

[raventepes]

This might answer some of your questions and make you think about it...

The Full Frame Debate | byThom | Thom Hogan

 

[lennon33x]

This might answer some of your questions and make you think about it...

The Full Frame Debate | byThom | Thom Hogan

Interesting read. It made me really think about what I'm "needing" when it comes to photography. Faster shutter speed, bigger focal area (i.e. smaller crop) and faster sync. Sounds like the 1D might be up my alley. The only caveat is relatively low ISO and decreased battery life. I think I can handle that though. Plus it's in my price range (as most of the time I have to get rid of gear to get new gear).