PDA

View Full Version : Megapixels mean nothing!


SolarTempest
03-24-2006, 10:42 PM
I just got back from my first exhibit at an art gallery! Had several of my photo sets put up on display along with work from other amazing photographers from the area. Very very exciting :toothy: . Also sold my first print!

After seeing some of the shots in the exhibit, it is clear to me that the "megapixel race" is complete garbage. I always knew that a few megapixels difference in cameras doesn't have a large effect on image quality, but never really saw the full comparison.

One of the photographers there had some absolutely stunning 8"x12" shots there that just blew me away. The colours, quality, and subjects were out of this world. And what did he use to shoot? A 4mp Nikon D2H. As an aside, he's now shooting a Nikon D200.

Four megapixels and the quality was tack sharp.
There was visually no difference between his 4mp shots and my 6mp shots.
In fact, I'd even say that his shots had far superior quality than any the 8mp Canon photos there (20D and Rebel XT - same sensor).

Just really shows that lenses are the most important factor in getting the best shots possible!

Ollie
03-25-2006, 05:45 PM
I just got back from my first exhibit at an art gallery! Had several of my photo sets put up on display along with work from other amazing photographers from the area. Very very exciting :toothy: . Also sold my first print!

After seeing some of the shots in the exhibit, it is clear to me that the "megapixel race" is complete garbage. I always knew that a few megapixels difference in cameras doesn't have a large effect on image quality, but never really saw the full comparison.

One of the photographers there had some absolutely stunning 8"x12" shots there that just blew me away. The colours, quality, and subjects were out of this world. And what did he use to shoot? A 4mp Nikon D2H. As an aside, he's now shooting a Nikon D200.

Four megapixels and the quality was tack sharp.
There was visually no difference between his 4mp shots and my 6mp shots.
In fact, I'd even say that his shots had far superior quality than any the 8mp Canon photos there (20D and Rebel XT - same sensor).

Just really shows that lenses are the most important factor in getting the best shots possible!


I'd say the photographer is the most important factor in getting the best shots possible.

That being said, I do like having a high MP camera. It's pretty useful for when the lens zoom just doesn't cut it. You can still get a decent shot even if you can't fill the frame. But you do make a good point. Once you get past 4MP (about what one needs for a 8*10 print, IIRC), the extra resolution isn't always that big a deal.

Nicepants
03-26-2006, 04:03 AM
Don't forget to adress the main reason for megapixels being an issue, though. digital cameras, as oppossed to film cameras are most convenient for people who are doing photo editing or manipulation on a computer. In such events, sometimes having an outrageously high megapixel is a HUGE advantage. It doesn't make it impossible to do the work if the mpx are too low, but it sure helps if they are high.

Then, of course, there's printing. Getting a sharp image at 8" x 10" is all well and good, but what of twice that, or what about poster size? If you are taking photos that you KNOW will go on a poster than you'll definately want something with much higher megapixels.

See, honestly, the megapixel race is more for the benefit of the computer users or the printers, and not the photographers. As far as cosplay pictures are concerned, I doubt there would ever be a reason to go above 6 megapixels, and most could get the job done with 4. But for anything else, those megapixels can make a difference.

SolarTempest
03-27-2006, 09:11 AM
Once you get past 4MP (about what one needs for a 8*10 print, IIRC), the extra resolution isn't always that big a deal.
For sure. 4mp does a great job at 8"x10", but higher might be tough. (Never gone higher on 4mp myself.)

I do love working with the higher mexapixels and it does help a lot. There's definitely a lot more room for error in a picture for composition when you've got pixels to spare. Like when all you have is a 50mm prime and you wish you had a 105 =P

What are the details on printing poster size? I've never done this myself and only know one person who has. What px dimensions and DPI are posters printed at?

I really think the megapixel race in more a marketing gimmick for consumers who don't know better. Much like the megahertz race of computing. Only now I think people are beginning to realize that megapixels are the determinant of photo quality and that sheer clock frequency isn't the the determinant of computing power.

Of course, there are some people who actually use their computers and a lot of photographers who see the benefit of higher resolution (especially professionals). I'm just really excited and interested to see where the technology is going to turn over the next few years as people really aren't needing much higher resolution cameras.

staereo
03-27-2006, 10:18 AM
Solar-
I am feeling the need to go ahead and correct much of the misinformation that you are discussing. While many of your observations are based in subjective opinion, my problem is coming from the fact that your facts are being either mistaken of fabricated.
I have avoided going into this because I didnt want to ruffle your feathers, but at the same time, I don't want people to be led to improper conclusions based on facts that are false.

I just got back from my first exhibit at an art gallery! Had several of my photo sets put up on display along with work from other amazing photographers from the area. Very very exciting :toothy: . Also sold my first print!
After seeing some of the shots in the exhibit, it is clear to me that the "megapixel race" is complete garbage. I always knew that a few megapixels difference in cameras doesn't have a large effect on image quality, but never really saw the full comparison.
One of the photographers there had some absolutely stunning 8"x12" shots there that just blew me away. The colours, quality, and subjects were out of this world. And what did he use to shoot? A 4mp Nikon D2H. As an aside, he's now shooting a Nikon D200.
Four megapixels and the quality was tack sharp.
There was visually no difference between his 4mp shots and my 6mp shots.
In fact, I'd even say that his shots had far superior quality than any the 8mp Canon photos there (20D and Rebel XT - same sensor).
Just really shows that lenses are the most important factor in getting the best shots possible!

First of all, being in focus is just that. Any camera can do so, and to make up for a lack of focus with a bolstering of megapixels never works. I could show you pictures that are easily as sharp from an old 3.1mp DIGIcam I used to use. ( http://www.dpreview.com/gallery/photo.aspx?gallery=kodakdc4800_samples&photo=20 )
Being in focus is part of the skill of the photographer, and that is the first step. Another point to be made is that many of the covers to image sensors of newer cameras today deliberately soften an image before the light is picked up by the sensor.

The thing is, an equally focused camera from a higher megapixel camera will produce a more detailed image. Its just a fact. For an exxagerated proof of this fact, please feel free to visit the gigapixel project, at the following link: ( http://www.gigapxl.org/gallery.htm ) The quantity of pixels DOES mean something, as does the sensor size.

Furthermore, you quote that the 20d and the 350d (xt) share the same sensor. Thats plain wrong. The 20d sensor has a max resolution of 3504 x 2336; and is a 8.2 million effective pixel/ 8.5 million photo detector sensor. The 350d has a max resoultion of 3456 x 2304; and is a 8.0 million effective pixel and 8.2 million photo detector sensor. They are inherently DIFFERENT sensors. And this is just brining up the photo detector size. There are MANY MANY differences between these two sensors, which Iwould be happy to list here, but my point is just to ensure that we know they are in fact DIFFERENT sensors.

It has long been said that lenses are important, to which I highly agree. But the problem is, youre negating teh fact that image sensors of today are the film of yesterday. Photographers have long been choosy about what film they use for their work. Today, the digital photographer makes a single 'film' choice to commit to when they purchase an image sensor, which in our case, comes in teh form of a camera body. Unlike the days of chemical exposure, we can't grab any old body and fill it with the film of our choice. When we purchase our camera body, we our now choosing the exposure media for the life of that body. So not the body plays a much larger role in the type of image we shoot, as it has the exposure media (ie: sensor) built in. So to write of a quality image to a lens alone, would be just as foolish as to quote the opposite.

Another point: If you took a 4 mp camera, and interpolated the image such that it matched the same resolution of an 8mp image, you WOULD notice a difference between the 2 images. One would be correctly recorded, the other would be software's best guess as to what half of the pixels should look like. Now, if youre looking at a 4mp image, next to a shrunken down 8mp one, of course the images are going to be far more similar in quality. This is simple logic, as you are not 'making up' information to increase the size, you are removing information from the larger image, that the 4 mp sensor may not have ever recorded to begin with.

Megapixels DO mean something.

To add to this post, the D2H is an old pro camera. When it was released in mid 2003, it was considered a professional level camera. Much like the canon 1ds. Youre talking a LOT of power pulled out of that little image sensor. Every piece to that camera was maximized, and that 4 mp camera is many fold the quality of your average 4mp point and shoot of today. Theres a reason those pixels are as beautiful as 4mp can be. The processor in that camera is TOP NOTCH, not to mention its not a CCD NOR a CMOS sensor. Its an entirely different technology called JFET.

Thats like saying... a porsche 911 turbo is just a v6 and is faster than a ford mustang v8. Cylinders mean nothing. Thats blatently WRONG. The porsche just taps every cylinder for all that its worth, if they were given more cylinders they would be even faster. (Isnt porsches new supercar a 10 cylinder?) Its a matter of exploiting your resources for what they are worth. That little 4mp camera, the d2h, was 3500. Thats far above and beyond what most 4mp cameras cost when it was released. Its a trade off.
Either way, megapixels DO mean a lot. A LOT. It just happens some cameras pull more out of an equivilant number of pixel sensor than others. And usually have a price point to back it up.

This leads me to my next point. In another thread you posted about cameras, http://forums.cosplay.com/showpost.php?p=1339172&postcount=25 , you insinuate that when megapixels are within 2mp of each other, they might as well be the same. Well, as I stated throughout this whole post. Megapixels DO matter, but its EVEN MORE about how the camera extrapolates the recording of light as to the quality of the body youre using.

Lets compare the two you were referring to in that post. The Canon 20D and the Nikon D70.
First, you say "The biggest ups that the 20D has over the D70s is that it's 8mp (vs 6mp) and 5fps (vs 3fps)." I didn't find that either of these were the main elevators for the 20D's higher pricepoint.

You go on to say ".....almost anyone really can't tell the difference between the three (film/20D/D70s)." I beg to differ on this. Perhaps general population, however anyone choosing the higher priced 20d probably is able to note the difference between film, a 20d, and a d70's image. Besides the resolution, the entire image is recorded differently and subtle differences in how each records can be noted by people with a photographic eye.

You also note: "I think a key advantage is that the D70s have built in wireless capabilities (Nikon Creative Lighting System). That allows me to set up remote flashes which can be wirelessly trigged by my on-camera flash, which isn't possible with the Canon unless you spend extra cash to get slave triggers or a commander."

If it's being wirelessly triggered by your on camera flash, doesnt that mean you're using a slave lighting system, which uses photocells to capture incident lighting thereby triggering their slave strobes? If this is the case, you betcha canon can do that.

Now, lets go ahead and review the actual differences between these to cameras.

1) Magnesium Alloy case on the 20D. I believe the d70 uses plastic. ?
2) 8.5 million total pixels CMOS on the 20d. 6.24 megapixel (total) CCD on the d70.
3) 8.2 million effective pixels on the 20d. 6.1 million effective pixels on the d70.
4) 3504 x 2336 resolution on the 20d. 3008 x 2000 on the d70. Thats an entire 500x300 picture missing between the two's resolution.
5) 20d uses, I believe, a 24bit RAW capture. d70 uses a 12bit.
6) 20d has a 9 point manually selectable AF, I believe the d70 is a 6 point non selectable system. (as in it uses auto modes only for deciding af point)
7) 20d has an asa range from 100-3200. d70 is 200-1600.
8) the 20d uses second generation e-ttl metering. The d70 uses first still (i-ttl).
9) The 20d is flash sync cord ready. the d70 is not (requires hot shoe adapter)
10) Not only does the 20d offer 5fps to the d70s 3; but the 20d can continue for 23 frames, the d70 tops off at 12 frames.
Edit => 11) http://www.dpreview.com/reviews/CanonEOS20D/Samples/Compared/Studio/20d_p2_iso0100_crops.JPG http://www.dpreview.com/reviews/CanonEOS20D/Samples/Compared/Studio/d70_iso0200_crops.JPG Which do you think is better?

Bah, I could honestly go on and on about the advantages of the 20d, but there would be no point. I really do think the d70 is a fantastic camera. It is GREAT! But the 20d was built to a different price bracket. And to say that the two cameras are nearly identical leads people to believe that the canon is priced higher without justification.

Basically, I just want everyone to get a fair idea of the facts involved in what you discuss, and I think this post helps clear some mistakes up.

I hope this helps,
Bruce

SolarTempest
03-27-2006, 04:05 PM
No ruffled feathers at all, Staereo.

I mentioned sharpness, you mentioned interpolation - right on! Resizing an 4mp image up to the 8"x12" size, images still look great and sharp. The average photographer will never print larger than this, so ~4mp image resolution is just fine.


My newspaper photo editor last year hated one of the photographers that shot for him SO much. Every shot he took was terribly out of focus. Photographer's explanation? "My photography is a representation of the world; It's fuzzy and never quite clear, so that's how my pictures are."

Of course, he never did end up shooting for the paper. :toothy:


My bad on thinking both sensors are the same. Both sensors are extremely similar in size (within a few percent) and should yield almost the same image quality. Is the image quality between the two sensors very different? anyone?


Of course, it's foolhardy to say only the lens or the sensor or the camera body alone is THE determining factor of a camera, my point exactly Staereo! Thank you for clarifying this!

Relevant to this thread, I think it's wrong that people are advertised cameras on the premise higher resolution=better camera. Anybody out there with a DSLR, what are the most frequent questions people ask you about your camera?

For me:
1. What is your camera?
2. How many megapixels?
3. Is it yours?

At least it's the second question and not the first..... :untrust:


I like how Stereo has discussed the D2H in more detail. He's really hit on all the points on how the quality of an image sensor does not only lie in megapixels, but really circuit design.

JFET indeed. Any type of transistor, MOSFET or JFET, really isn't what make a sensor what it is. It's such a gross simplification... Believe me - I've studied microelectronics, transistors, and electrical circuits in almost all my time over the past 4 years in Computer and Electrical Engineering.

Mmm... maybe the discussion of 20D vs D70s should be moved into another thread.

staereo
03-27-2006, 05:23 PM
In this case, I absolutely agree with you. The megapixel rating of a camera has become a trendy catch-spec for digicams all over. When people discuss a purchase, or a camera which they already own, the first thing I hear is 'megapixels', even in cases where they cannot recall the manufacturer.

So without a doubt, many manufacturers have turned to inflating one spec in an attempt to market an inferior product as being better. Its ingenious in the business sense, but unfortunately leaves a poorly informed public at risk of buying blind.

As far as the image sensors between the canons go, I think what youre recalling is the last dual-tier lineup. The canon 10D (the forerunner to the 20d) and the 300d had the identical image sensors. They shared the same tier difference as the 20d and 350d do today; but canon realized a huge mistake in doing this. People were reflashing the 300Ds to have nearly all of the features and attributes of the 10d! So people stopped buying the 10D, because they could reflash and have the same image sensor as the 10D by buying the 300d for hundreds less. So I think thats where you may have remembered this from. They made sure to make the tiers a little more divided in this generation, to keep their platforms valued. The sensors are noticably better with the 20d, specifically in low light, as I recall. But ya, canon made that mistake that youre remembering with the 10d/300d generation of cameras. Lesson learned on their part, I guess. LoL

So I agree with ya here. Good points!

Bruce