5D full frame question

Started by kkirksea, 12/15/2005 05:18PM
Posted 12/15/2005 05:18PM Opening Post
Bear with me if this has been asked already ( and I couldn't find it in the search ) or if this is obvious to all the pros.

I don't have a good grasp on how DSLR's work with astrophotography yet, so I'm trying to get my head around it. I'm trying to figure out which way to go, 20da/5d/Sbig.

With the 5D's FULL frame sensor, does that mean that at any given magnification the image recorded by the sensor will be wider field ( 1:1 vs 1.5 or 1.6:1 on non full frame DSLRs ).. or will it be the same field with just more resolution?

Will vignetting aspects of the 5D/Scope or 5D/eyepiece/Scope be a bigger issue?

Thanks
Karl
Posted 12/15/2005 07:10PM #1
Hi Karl,
The 35mmm chip size will afford the Canon 5D a apparent "wider field"- the same way using a medium format 6cmx4.5cm gains over a 35nn camera, but someone please correct me if I'm wrong- I read somewhere that the pixel density is lower than the 20D? That means you cannot magnify an image as well as a 20D, all things being equal. ot sure if it is enough to make that difference though.

Although vignetting is an issue to think about if your focurser is small,what is more important is the focal plane curvature of APO scopes. Even with a reducer/flattner, some premium scopes will show star image elongation at the outer areas of the frame, thus rendering the extra gain of the additional chip area useless, unless you are not fussy. Scopes like the FSQ-106N will work great however. not sure about a Sky 90+reducer config.

Ivan
Posted 12/15/2005 07:37PM | Edited 12/15/2005 07:39PM #2
Karl,

Here's my answer, and I am willing to be corrected by others:

A "Full Frame" digital chip means it is as big as a 35 mm piece of film (36x24 mm or so). The 20D is not that big (22.5 x 15). SO, the 20 D only senses the middle area of what the lens is putting out.

However, DSLR sellers have decided to turn this into some kind of an advantage--Your "100 mm lens now acts as if it were a 160 mm lens." In fact what is happening is that the camera is throwing away all the light that comes in from the lens but is beyond the central portion where the chip resides.

With the 5D's FULL frame sensor, does that mean that at any given magnification the image recorded by the sensor will be wider field ( 1:1 vs 1.5 or 1.6:1 on non full frame DSLRs ).. or will it be the same field with just more resolution?

The shortest answer is--take a 35mm negative or slide, and cut out the central 22.5x15 mm. That is the picture you now have. Throw the rest away. If you print out that central portion you have left--at the same magnification as you printed the original, and it will be smaller. Blow it up to the same size as you printed your whole frame, and it will be an enlargement of the central portion--with the rest thrown away. (Yes, then your full frame is getting a wider field.)

The problem is with your question--there is no such thing as "for any given magnification." Let us instead say--for any given focal length. We need to make this distinction because the final magnification is determined by how you view the image. If you take an image that is 43m across the diagonal (the diagonal of a 35 mm negative)on an 8x12 piece of paper you will get one magnification. But if you take only the central 27 mm cross section and use that to fill the whole 8x12 piece of paper, then you will get a bigger magnification (by that famous factor of 1.6).

But--and this is the important factor--you have not changed the native "magnification" of the focal length. You have just thrown out some light that was on either side of the chip, and spread what you got in the center over the whole piece of paper.

So--has the resolution gone up? That depends on the size of the pixels. If the two chips have the same size pixels, and you are using the same length lens, then the number of arc seconds you spread over each pixel will remain the same. And that is what determines resolution. (But note that if you plan to make a "print" of the image, you will not have as many pixels to work with if you have kept the pixel size the same.)

But in order to make the chip larger, you must either increase the number of pixels or the size of each pixel (or both!). If you have increased the number of pixels (and kept the size the same), the resolution remains the same. If you have kept the number of pixels the same, and merely increased the size of them, then your native resolution has actually decreased. You are getting more arc seconds onto each pixel.

How's this for an answer.

And as for vignetting, etc. Since that generally shows up around the edges of the frame, it is largely ignored by the smaller 20D chip--it is in the area that the chip cannot see. So it is not as big a problem. But that does not mean it is non-existent.

Alex
Posted 12/16/2005 11:40AM #3
http://www.bobatkins.com/photography/digital/canon_eos_5d_or_20d.html

Paul Gustafson
Posted 12/16/2005 02:04PM | Edited 12/16/2005 02:11PM #4
http://www.dpreview.com/reviews/canoneos5d/

Includes 20D comparison.

Dan
Posted 12/17/2005 04:26AM | Edited 12/17/2005 05:11AM #5
"OK, firstly, when you put a 300mm lens on a D60 you do NOT get a 480mm lens – it is still a 300mm lens. What has changed is the format of the camera – from 36x24mm film to a 22.7x15mm chip."

Also, "The smaller format uses less of the normal projected image circle of a particular lens and thus edge distortions are ‘cropped’ out. See this site's review of the Canon 16-35mm for proof of this."

The Canon L series lenses seem to be required for that 5D especially for astro work. A 20D gets by without the edge aberrations mostly and even using the same L lens the 20D will exhibit less distortion at the edges.

Quotes from: http://luminous-landscape.com/tutorials/understanding-series/dslr-mag.shtml

It's more complicated than we think. Thankfully

By the way, this is one first class photography site. I'm sure you'll be amazed at the info and images as I was.

Dan
Posted 12/20/2005 05:27AM #6
Alex did a really good job of answering your questions and gave you great information. I have used a 1Ds Mark II for wide-field astrophotography for almost a year, and have been using a my 5D a little since I got it in early October. The one big thing you will find is that astrophotorgraphy is the most unforgiving of all photography. Corner aberrations that are not visible in regular images will make the corners of astrophotos unusable.

For example, my fabulous Canon 24mm f/1.4L lens looks great wide open for all kinds of photography--but the stars are terribly astigmatic and suffer from coma in the corners. To mostly eliminate the effect, you have to stop down to f/4; it takes f/8 to totally eliminate it. On the other hand, I can also just crop the corners; I still crop off way less than a 20D does.

That said, I still have a modified 20D and a 20Da for astrophotography--because they are cheaper. The 1Ds Mark II and 5D are for landscapes and birds, etc. where FF has real advantages.

Jeff

Jeff 8O