Visual Sensor Size Comparison

<p>So here's what I'm trying to find and I'm struggling. I'd like to see a visual comparison of different sensor sizes. So, for example a scene shot with a d700 at 300mm, then the same again with d300 at 300mm equivalent and the same again with p100, all at the same resolution. Does anyone know if there is anything like this on the web? I know that everywhere says that larger sensors increase picture quality, but I've never actually seen it, I've only read the technical reasons why.</p>

<p>Any help or links would be awesome :)</p>

<p>Unless you're also able to specify and control how the images from these different devices are all processed before you can see them, such a test doesn't help, really. Post production (or even just in-camera JPG rendering) will have as much more to do with the perceived differences than anything else.<br /><br />Of course, if you're doing your test on low-light subjects, and thus are using higher ISOs, that's going to produce very different results as well. And, you've got to be very careful when describing "300mm equivalent" on different bodies. What do you mean by that? Are you talking about keeping the <em>perspective</em> the same, including the background and foreground elements? And, are you talking about ignoring, or trying match up the inevitable (and significant) apparent differences in depth of field that each format will render, under some circumstances?<br /><br />This is a lot more complex than just "take the same shot three times with different camera bodies."</p>

<p>Hmmm... I don't think I appreciate the complexity of it all. Firstly let me try to explain my definition of the "equivalent". In this test I would think that (changing the length) the FX is simply at 28mm, the DX about 18mm and the P100 at 4mm, and if all the camera's were set to the same settings (iso, aperture, etc), and all shot at say 10MP, would the image quality (i know it's a subjective word) suffer, and more importantly how.<br>

I'm not talking about perspective or other photographic elements, though they do interest me, but more about the technical quality of the images. I also know that it's not purely down to sensor size, but to do with the engine of the camera too, but again, from what I've read the size of the sensor makes a huge difference.<br>

If there are no visual examples online then I'm interested to debate the differences too.</p>

<p>How would you evaluate the "technical" quality of the image? Sharpness, for example? That's why I brought up depth of field. In practical terms, the larger the sensor, the less of it you have without stopping down more to make up for it. So if your evaluation of the image includes noting that there would be very (very!) noticeable differences in what appears to be in focus - in the scenario you describe - you could easily draw the wrong conclusions.<br /><br />Some people show up here complaining that their expensive new DSLR doesn't get as much in focus as their simple point-and-shoot, and thus the image is of lower quality, despite the higher price. You get the idea. Each rig is optimized for difference priorities. Though you'd notice less of a difference between the DX and FX formats that you would between the DX and the P&S camera, that way.<br /><br />There are also going to be lens issues. Ironically, some lenses that perform fairly well on FX might actually produce "better" images on the DX body because the DX sensor is only paying attention the flatter, less distorted, sharper sweetspot closer to the center of the projected image circle. So your comparison has to also include very different lenses - in the interests of making the most out of each sensor format. And that makes for an uneven playing field, too.<br /><br />The important differences (getting past the not trivial optical issues out in front of the sensor) involve signal to noise ratios. If you really ARE proposing three different size sensors that have the same number of actual sensor elements on them (say, 10mp), then the larger buckets on the larger sensor will result in lower noise, and greater sensitivity. These differences usually show up when you're shooting in more extreme conditions. <br /><br />Which brings us back to having you describe your test better. Not just in terms of what's the same about each exposure, but in terms of what you're shooting. Are we talking about available light portraits by candlelight, or kids' soccer in bright daylight?</p>

<p>[[i know that everywhere says that larger sensors increase picture quality, but I've never actually seen it, I've only read the technical reasons why.]]</p>

<p>While I completely agree with Matt here, you can compare the image quality of test/target shots on dpreview.com using their so-called interactive wizard.<br>

I don't think they have this as a stand-alone application, but within a specific review, you can choose from a couple different cameras.</p>

<p>For example, in the recent Panasonic G10 review, for JPG images (processing done in-camera):<br>

http://www.dpreview.com/reviews/panasonicdmcg10/page3.asp</p>

<p>You can pick up to 3 other cameras. Move the selector around the scene to see the differences between the reviewed camera and ones you chose. </p>

<p>But even this is of limited value because of all the factors Matt describes above. (For example, these tests are all taken at different focal lengths.)</p>

<p>[[would the image quality (i know it's a subjective word) suffer, and more importantly how.]]</p>

<p>Suffer for what application? 100% views in Lightroom/Photoshop? Web images? 4x6" prints? 30x40" prints? Your intended output is as much a factor as anything here.</p>

Other factors are at play besides sensor size. For example, CMOS is worse quality than CCD (proof: Canon SX1 vs SX10, Panasonic FZ100 vs FZ45). Also, newer technology usually beats older technology.

 

http://www.dpreview.com/previews/panasonicdmcfz100/

 

I recommend going to the above URL and replacing the Coolpix sample with Canon EOS-1D Mark IV. You can see that the FZ35 is giving generally competitive results at low ISO, despite a much smaller sensor, and despite pitting a 27-486 zoom lens against a 50/1.4 prime lens. You can move the rectangle around to compare different bits.

 

P.S. Oops, Rob said it better than I did.

<p>Paul, as others have pointed out, there's no easy comparison. Under the right cirmstances, my 10-ish MP G11 and my 10-ish MP 40D will produce images very similar to my 10-ish MP (or actually 12-ish) 5D. That said, the smaller you go in format, the greater your limitations are. If you want the same shallow depth of field as on a larger format camera, you're out of luck. If you want to shoot in the same dim light, you're out of luck. If you're struggling with excessive dynamic range in the scene, you're out of luck.</p>

<p>Most formats are created equal on the deeper end of their depth of field capabilities. Smaller formats will yield the same depth of field at larger apertures, but they can't be stopped down as far as with larger formats. Then maxing out DoF at diffraction limits, one has a much slower exposure with the larger format, but then the ISO can be bumped to compensate. In the end, both formats are going to limit at the same DoF with the same noise. </p>

<p>Thus the differences are on the shallow DoF and low light end. There are many photos my G11 can take. My 40D is capable of taking all of those, in addition to many, many more. My 5D is capable of taking all of those that the 40D can, plus even more still.</p>

<p>Comparing formats is all quite involved. I wrote something about how formats compare here:</p>

<p><a href="http://www.graphic-fusion.com/fullframe.htm">http://www.graphic-fusion.com/fullframe.htm</a></p>

<p>I tried to make it as comprehensive as possible. I'm always tweaking and adding, so if you have any comments or suggestions, please speak up!</p>

Mind that a given sensitivity (ISO) may not actually be what is reported. Right there you would have differences in cameras with same sensor size but differing sensor hardware & firmware. Check DxO sensor comparison site; same had been reported elsewhere too but don't have the URL handy.

<p>There is the other aspect illustrated on dpreview of pixels/cm2. I wondered why I was able to make good A3 prints from my original little Canon s20 until I discovered though it only had 3.3Mp sensor there are 8pixels/cm2 which is not too much different from the DSLR at around 5Mp/cm2 and vastely better than the G11 at 23Mp/cm2 :-) Or my lovely FZ50 with its 26Mp/cm2. So much for the pixel counters :-)</p>

<p>Hi all,<br>

Thank you all for your responses, and very thought provoking they are aswell. The widget on dpreview was especially useful and exactly the sort of thing I was looking for. From looking at that it's clear to see the effect a larger sensor has with noise reduction at higher isos, but really it's not a huge difference at lower isos, which really does make me question how large I need it to be for the work I'm interesting in shooting, which at the moment is primarily landscapes, and what I'm looking to shoot in the future when I can afford a longer lens or two, wildlife.<br>

I'm thinking to be honest for the landscape work, as I mostly just shoot on daytrips with friends, that a smaller sensor would be fine as it's mostly low iso work, and even when shooting wildlife with long slow lenses, it should still be ok as long as the light is good. I think an upgrade would only be necessary if I intend to shoot when the light goes down. Anyway, to cut to the chase, the reason I got onto all of this was for the purchase of a new camera system in the future, and I'm torn between a superzoom with a tiny sensor and a DSLR with a big sensor but less zoom and a lot more cost.<br>

I suppose my next question would be about density, as someone mentioned previously the density of the MP on the sensor affects pq also, so would I be better off with an older camera with a less densely packed sensor, or do all the added benefits of a more modern camera (such as the image stablisation) make up for the density?</p>

Paul, I see that you have FD lenses. The Panasonic G1 with excellent 14-45 kit lens (28-90 crop equivalent) is on closeout for $499 now. With a $32 FD lens adapter you can continue using your lenses, and if you can deal with electronic viewfinder, this is a smokin' bargain.

<p>I see the 4/3 as the logical step up for me if only they came with other features I prefer, they don't so I'm sticking with what I have Pany FZ50. I dislike intensely the pre-occupation with compactness which leads designers to make zoom lens which close down a couple of stops as they zoom out. Panasonic probably spoilt me with their FZ20 which has a constant f/2.8 through the x12 zoom range and they seem to have been going backwards ever since despite improvements in noise and other good features on the cameras. The FZ100 is pathetic in this respect despite rave reviews :-)<br>

The other aspect of 4/3 is the price of the lens, if you use the FD adaptor and are happy to work manually, I was considering adaptor to use my Pentax range of lens. The only option which comes near to matching what the FZ50 is giving me is a 4/3 body and the x10 zoom but the price is horrible for these days. Again I feel it is a step backwards to use manual lens in this day ofthe automatic , including OIS or other systems. Olympus has it in the body which is where it should be. Having it in the lens is just a money making rort I think for some abstract IQ reason.</p>

<p>Paul, go to DXO Mark, you will see that larger sensors fare considerably better in all catergories - noise, ISO sensitivity, color, dynamic range, etc.<br>

<br /> What will be painfully clear is that that the "superzooms with tiny sensors" that you reference are frankly horrible when compared to DSLRs, especially those with full frame sensors. The divide between the two is enormous.<br>

<br /> The very best sensor tested by DXO, the Phase One D65 (65mp and very expensive), actually doesn't do dramatically better than something like the Canon 5d2 with it's full frame sensor, but even the Canon digital rebels are vastly better than tiny point and shoot sensors.<br>

<br /> If you're on a budget something like the Canon t2i would seem an obvious choice.</p>

<p>Interesting looking website, I've not seen that one before - I'll have to investigate - thanks for the link :)</p>