| Author |
 Replies: 23 / Views: 3,594 |
|
Pillar of the Community
United States
4038 Posts |
I have been thinking about "Diffraction Limited Aperture" (DLA)lately and effect on my imaging. DLA is dependent on the size of the pixels on the sensor. The smaller the pixels, the larger the aperture required to not have the image sharpness degraded by diffraction. For a few select Canon cameras, the DLAs are: XS / 40D; 10.1MP; pixel size 5.7um; DLA = f9.1 XSi / T3; 12.2MP; pixel size 5.2um; DLA = f8.3 T1i / 50D; 15.1MP; pixel size 4.7um; DLA = f7.5 T2i / 60D; 18.0MP; pixel size 4.3um; DLA = f6.8 Any smaller (effective) aperture than the above and your shots will be limited by diffraction. I got to thinking that because of the above, perhaps there would be little benefit for coin photography to having more and more pixels in the same sensor size. Larger sensors with more of the same size pixels gives a bigger field of view, but has other issues (such as lower effective magnification and of course $$) that make full-frame cameras less attractive for macro work. So what about going backward to earlier generations of cameras with same size sensors but fewer pixels, for effectively better DLA? A problem is that if you have fewer pixels to work with, for same down-sizing ratio you will have smaller final image size. But is that such a big deal? To test it all out, I got onto ebay and looked for a suitable guinea pig. I found a Rebel XS for sale for $109 (cheap!), body only (perfect, I don't need any more lenses), and with a minor problem with the flash card port (I don't use flash cards on bellows anyway). A few days later, my "new" XS arrived and I was doing comparison shots. I only looked as far back as the XS and the 40D not because of sensor size, but because those are the first cameras in the Canon lineup that offer Live View, which is of course ultra-critical to have. So I had not done the calculations before, but once I got the camera and started downsizing I discovered a nice mathematical property to the Canon sensors...going from XS at 10.1MP (3888 x 2592) to the T2i at 18MP (5184 x 3456) the ratio is a convenient 33% increase. This means that I can downsize the T2i images by 4x, and the ES images by 3x, and end up with the same 1292x864 final image size for web publishing. I usually crop whole coin images to 800x800 and either camera allows this while still maintaining an integer downsizing ratio, which I consider important for maintaining sharpness. I leave variety shots un-cropped. So how's the result? Here are the two images, 5x variety shots of a Lincoln Cent RPM. I did this first since I wanted to prove the EFSC was functional on the ES. At 5x, with no EFSC the loss of sharpness would be very apparent. Canon Rebel ES:  Canon Rebel T2i  My conclusion is that these are essentially equivalent in sharpness. The difference in DLA is offset by the sharpness improvement of 4x vs 3x downsizing. What this means to me is that I can recommend ANY of the recent Canon lineup shown above for coin photography. More pixels doesn't give a better result. I believe all that matters is the camera is new enough to have Live View and EFSC. Would love to hear your thoughts on this...Ray |
|
|
|
Pillar of the Community
United States
4132 Posts |
I'm gonna have to disagree. The second image is noticeably sharper. If I pop them up in separate windows, line them up and alt-tab between them, the difference is pretty clear.
On the other hand, it might just an aesthetic difference. Down-sized, I can't say that the T2i image actually reveals more detail, or if it just has a sharper "look". And given how cheap a used, older-model Rebel body is, it could be a really good route for somebody who doesn't want to spend a lot of money for a marginal difference. Combined with an el-cheapo lens and bellows, cost is getting down into point-and-shoot territory.
Edited by CaptainFwiffo
09/23/2012 12:20 pm
|
|
Pillar of the Community
 United States
4038 Posts |
There are subtleties you need to consider...
The two images aren't exactly apples to apples. Each one is a focus stack (required at 5x) and it's hard to get the focus at exactly the same focal planes so there are slight differences in focus. To punctuate this...the T2i is a stack of 3 images, while the ES is a stack of 2 images.
I also did not process these from RAW, instead relying on the jpg outputs from the camera (I know, shame on me...). I expect the T2i jpg algorithms to be a bit more robust. In fact, when I look at the T2i result, it has more "sparklies" than the ES. This is possibly due to more internal sharpening happening before the image exits the camera. It may also be due to a different level of sharpening happening within the stacking program, I'm not sure. The extra apparent sharpness of the T2i may just be the result of extra sharpening.
I plan to shoot a full-coin shot today with both cameras, single image, and will process from RAW so the workflow is the same. Then we can see if there really is a sharpness difference.
Contact me for photographic equipment or visit my home page at: http://macrocoins.com |
|
Pillar of the Community
United States
4132 Posts |
I will withhold judgement then, aside from the point that the old-model rebels are plenty, even for very subtle variety photography and VAMming.
This is part of the reason why I don't care for the "megapixel" nomenclature. Because it's a measure of pixel-area instead of linear resolution, it greatly exaggerates the difference between sensors. 18 megapixels sounds like it ought to be almost twice the sharpness of 10 megapixels. But it's really only 33% more resolution. That's not a lot, considering how many other factors contribute to sharpness.
|
|
New Member
United States
16 Posts |
Good thread, like the idea on what can be done with some adjustments and going oveboard not necessary to meet the intended goals. My question is on the stacking. Did you play around with the focus or stay the same with all shots and stack.
|
|
Pillar of the Community
United States
4038 Posts |
You can stack in two ways: adjust focus (distance from lens to sensor) or camera position (distance from lens to object). Both methods will give you a similar effect. For focus stacking of non-macro scenes (landscapes, portraits, etc) it's better to adjust with focus. But for macro shots, it's better to adjust the camera position. I use a microscope stand for my coin photos, so it's well-suited for position focus stacking.
|
|
Pillar of the Community
United States
4038 Posts |
Here are the overall coin pics comparing the two Canon cameras. Which do you like better and why? There is more difference than I expected. Either the jpg engine is fairly different between the cameras, or I may have some settings differences.   Contact me for photographic equipment or visit my home page at: http://macrocoins.com |
|
Pillar of the Community
United States
4132 Posts |
The first one gives me the impression of being sharper, but it's not a big difference. The white balance seems closer to correct on the second one, but since I always shoot raw, that doesn't matter at all to me. The first also seems more contrasty. The greater impression of sharpness may be due mostly to the difference in color/contrast.
I still think the original point stands. If somebody wants to have a dedicated coin photography setup on the cheap, they can combine a used, older-model Rebel body with a bellows and el-cheapo duplication lens and get 95+% of the performance of the state-of-the-art coin setup for whole-coin images.
|
|
Pillar of the Community
United States
4038 Posts |
The first one is the XS, the second is the T2i. I looked over the settings and found that I had the Auto Lighting Optimizer or whatever it's called on with the T2i. The XS doesn't have that feature. I turned it off and re-took the pic and now the contrast looks more similar. Interestingly, the ALO affects the RAW image...  Contact me for photographic equipment or visit my home page at: http://macrocoins.com |
|
Pillar of the Community
United States
4132 Posts |
The auto-lighting optimizer probably only affects RAW files if they're processed with DPP.
|
|
Pillar of the Community
United States
3076 Posts |
Ray and SSD, As you may know,I joined this Forum 2 years ago for coins and was drawn to photography..How to get a good shot of the coin...With the camera I and we, happened to have. IN SHORT..This photography Spot, has truely grown from just the average person trying to get a very nice shot INTO THE TECHKNO details of what makes a great shot!!   In the past we have compared and found many an old glass is truly what we desire in this coin shooting arena... OLD GLASS and its applications... NOW OLD CAMERA'S..there Mega pixles and there differences...and to the public.. the more means better.... A post like this is well thought of and revealed...AND THANKS RAY!.....Due to the fact that this sight is extreme in it's use of the camera and its lenses....for there intended usage... I have often wondered what difference the MP component of the camera VS the sensor implies for our usage,, THE SAME may apply....AND I ASK...The Rebel body or Ti bodies are the low end affordable bodies with the lessor of the sensor capabillites if the high end Canon cameras.... WHICH is why I think this is a great example and thread... that the most..high end.. in our purposes is not the best...... Its a thread like this which compares the high end to the low end........All though its the not the high end canon ..it is what most of us all own or can afford...Ti2 to XT...and such wonder full pics we get from these cameras...My point is.. most of these cameras and the other camera manufactuers are all within the same ball park figures...But ray has broken it down from XS to Ti2 for us to see the total difference...as shown for our view and ideas of the out come....And it relates to other manufactures who also follow the same climb to better and better...pics...I must say,,,coins are there own monster, Thats why we are here |
|
Pillar of the Community
United States
4132 Posts |
Comparing the XS image with the new T2i image at work on my calibrated monitor at work, I can't draw a conclusion about which is sharper. The XS image seems sharper on the side of Abe's face, but the T2i seems sharper in the field on the left side of the coin. That makes me think that any differences in sharpness are attributable to slight differences in alignment/focus. Probably actually differences are only visible at higher resolutions by pixel-peeping.
The color is more accurate in the T2i image (assuming this is your average red cent and not purple toned). I actually think the ALO version is closer to a normal impression of "copper", but I think a minor white balance adjustment would get them into alignment.
|
|
Pillar of the Community
United States
4038 Posts |
I wanted to shoot these with an effective aperture at or larger than the DLA of the cameras, so I took all these shots at f4, for an effective aperture of f7. This gave good sharpness but insufficient DOF for Cents, so I also focus-stacked the shots to eliminate any focus issues. These are all 3-shot images, with critical focus on high points of cheek, center of throat, and field in front of throat, stacked with CZP. I realize now I should have shot with the Printing-Nikkor to eliminate any aberration issues that could obscure camera differences. Maybe I will do that, since I do want to know the real differences between the cameras.
Contact me for photographic equipment or visit my home page at: http://macrocoins.com |
|
Valued Member
Canada
331 Posts |
In terms of experimental setup, would it also be favourable to shoot at a relatively modest magnification, say, a whole Morgan dollar, so as to eliminate focus stacking from the list of variables? Just a thought. |
|
Pillar of the Community
United States
4038 Posts |
Perhaps that would be better. I think stacking still may be required at the DLA limit for the T2i even for Dollars, but maybe only 2 images would be required (features and field). I'll give it a try and see what happens. The XS will likely not need stacking at its DLA limit for Dollars. I suppose that says something right there...
Contact me for photographic equipment or visit my home page at: http://macrocoins.com |
|
Moderator
United States
23522 Posts |
Random comment: I'm not achieving sufficient DOF to get even a Morgan focused both at devices and fields, even with a mechanical aperture of f/8 on the ARD. It's easy to see when attempting to focus using Zoom Browser's Live View full-size detail. Done with rigor, it'd take me a stack of three for a Dollar, much less something at greater magnification.
|
| |
Replies: 23 / Views: 3,594 |
Posted 09/23/2012 12:44 pm
