Why a pentaprism?

[ckhaos]

Dear SLR experts:

The placement of a pentaprism together with a reflex mirror in an SLR allows through-the-lens viewing.

However, why not have two mirrors? i.e., replace the pentaprism with another mirror that is fixed at such an angle that it reflects light from the lower reflex mirror into the viewfinder. (I guess this arrangement would be maciam the submarine's periscope that we all learn in secondary school.) In this way, TTL viewing is still preserved, and the lower reflex mirror can still be tilted up at the shutter release. So, please let me know why a pentaprism is needed??

 

[ckiang]

The cheaper SLR DO use another mirror in place of the prism. I presume the prism is much more accurate than the mirror, and brighter also. And doesn't have a silver surface which can wear out.

Good binoculars and optical instruments like some microscopes also employ prisms as opposed to mirrors.

Regards
CK

 

[denizenx]

air scatters light mah.. and ppl dun tell me use vacuum, the thing wun flap...
but a prism is like fibre optics, it keeps most of the stuff intact...

 

[roygoh]

I tried tracing the light path and came to the conclusion that if you use a single mirror in place of the pentaprism to reflect the image on the focusing screen towards the user's eye, the image will be inverted (both up-down and left-right).

Of course more complex optics can be added to invert the image back to normal, but that makes it no better than using a pentaprism in the first place.

On TLR cameras where the photographer looks down towards the focusing screen, the image appears upright but flipped horizontally.

- Roy

 

[Ian]

Originally posted by denizenx
air scatters light mah.. and ppl dun tell me use vacuum, the thing wun flap...
but a prism is like fibre optics, it keeps most of the stuff intact...

The fact is that a prism actually has higher light loss levels than first face coated mirrors.

The actual reason and not one of conjecture is that the distance from the reflex mirror to film plane has to match exactly the distance from the reflex mirror to the focusing screen which is mounted at right angles to the film plane.

In order to allow the whole of the focusing screen to be used for focusing you require a light path that is equidistant at all points and a mirror doesn't meet this criteria, however a pentaprism does as it has a flat base and is designed so that the internal light reflection paths are always equal from the "input" to "output" faces.

As for light scatter .. it's complete bollocks in scuh a short path when using front surface mirrors.

 

[chriszzz]

Originally posted by Ian


The fact is that a prism actually has higher light loss levels than first face coated mirrors.

The actual reason and not one of conjecture is that the distance from the reflex mirror to film plane has to match exactly the distance from the reflex mirror to the focusing screen which is mounted at right angles to the film plane.

In order to allow the whole of the focusing screen to be used for focusing you require a light path that is equidistant at all points and a mirror doesn't meet this criteria, however a pentaprism does as it has a flat base and is designed so that the internal light reflection paths are always equal from the "input" to "output" faces.

As for light scatter .. it's complete bollocks in scuh a short path when using front surface mirrors.

Interesting explaination, but not quite right. The focusing screen is used for focusing on manual cameras, NOT in AF cameras ( I presume by focusing screen, you mean the little matte glass ABOVE the reflex mirror ). In fact, for many manual cameras ( such as the Canon F1 and numerous Medium formats ), you can remove the pentaprism and still focus manually with no problems. Therefore, the pentaprism has nothing to do with manual focusing.

In AF cameras, the AF Sensor is actually BELOW the mirror, at the base of the camera. You can flip up the mirror on your AF camera ( use BULB mode ) and see the AF sensor. Thus, the prism has nothing to do with autofocusing either ! In fact, as already pointed out, some cheap cameras ( such as the Canon EOS Rebel / 300/500/etc series) actually use a mirror in place of the pentaprism. If what you said is true, then these cameras would fail to focus !

Here's my take on why a prism is used. Remember your O-level physics. Light reflecting off the internal of a prism does so via "total internal reflection". What this means is that the light is reflected completely, with almost no loss. Light relfects off the inner surface from a single pane ( ie, the internal surface of the prism ), thereby avoiding faint double images. This is the same reason why good periscopes use prisms instead of mirrors as well !

Use your favourite search engine and look up total internal reflection. It will explain alot.

 

[denizenx]

 

[ckhaos]

Hmm... intriguing...

Ian: Would you care to elaborate on your equidistant theory? I don't think I understand it completely. Thanks,

Roy: I tried tracing the light path using two mirrors and the image comes out upright, so I don't think the use of the prism is to invert the image.

Chriszzz: If I remember my O level physics, I don't think TIR reflects 100% of the light, in fact, the light loss at a glass surface is HIGHER than that at a silvered surface! So, I don't think your reasoning is correct.

 

[Peculiar]

Plane mirrors are not used because they produce multiple images, thus high quality optical equipment does not employ them. Instead prisms are used as reflectors. Using total internal reflection, only one image is produced.

Multiple images produced by mirrors are quite faint to be noticed. They are caused by the silvered surface and the glass surface of the mirror.

In total internal reflection, NOT 100% of the light striking the prism is refracted, insteaded, a small fraction of the light is reflected off the surface, but it is negligible.

(Adapted from "Principles of Physics")
Hope my explanation helps..

 

[chriszzz]

Originally posted by Peculiar
In total internal reflection, NOT 100% of the light striking the prism is refracted, insteaded, a small fraction of the light is reflected off the surface, but it is negligible.

Exactly. Light enters the prism at 90 degrees, so there is a little reflection, and ZERO refraction. All light (100%) that enters the prism is reflected off the "back" of the prism, off a single pane, again with zero refraction and zero loss due to reflection. Finally, light leaves the prism at 90 degree to the exit surface, with a little reflection but again, ZERO refraction. This is as good as it gets.

Even more importantly, all surfaces are "single pane", which avoids double images, no matter how faint. Some light loss is inevitable though, but it's still better then a mirror, since silvered surfaces will also absorb and reflect light.

 

[chriszzz]

Originally posted by ckhaos
Hmm... intriguing...

Ian: Would you care to elaborate on your equidistant theory? I don't think I understand it completely. Thanks,

What he meant was that the distance between the mirror and the film MUST BE EQUAL to the distance between the same mirror and the focusing screen ( for Manual focus ) or the AF sensor ( for AF cameras ). Unfortunately, the prism is BEHIND the focusing screen, and only serves to transmit the image from the focusing screen to the photographer's eye. Therefore, it plays no part in the actual focusing for both MF and AF.

Chriszzz: If I remember my O level physics, I don't think TIR reflects 100% of the light, in fact, the light loss at a glass surface is HIGHER than that at a silvered surface! So, I don't think your reasoning is correct.

As explained, some light loss is inevitable. What's important is that the transmitted image is as "pure" as possible, free from aberrations caused by refraction and double images. In fact, people who have used a low end Canon SLR that uses mirrors instead of prism complain that the viewfinder is not as bright as that of a higher end SLR which uses the prism, when compared using the same lens. I don't have such an SLR, so I can't comment if this is true or not.

 

[roygoh]

Originally posted by ckhaos
Hmm... intriguing...

Roy: I tried tracing the light path using two mirrors and the image comes out upright, so I don't think the use of the prism is to invert the image.

Well, I said if a single mirror is used on top of the reflex mirror (as propsed by yourself in your original post) it will will present an inverted image.

If 2 mirrors are used, one directly above the focusing screen but tilted towards the front of the camera and the other from the front of the camera reflecting the image from thefirst mirror towards the back of the camera, the image will be upright bu still left-right inverted.

There has been low-cost SLRs that uses a mirror array to replace the prism. Here's the reference I found in Popular Photography's website:

http://www.popularphotography.com/Camera/PrintArticle.asp?ArticleID=144

I guess what is done is to replace each reflecting surface of the prism with a front surface morror. Did not have time to dig too deep into this, but seems like the reviews on such mirror systems were that they produce darker images than the prism.

If you look at the focusing screen image from the top with the camera pointing towards our front, the image will be upright but left-right inverted. To redirect this image 90 degrees towards the user for eye-level viewing, my guess is that 3 reflections are required to make the final image upright as well as left-right correct.

I really don't understand how the image can be left-right correct with only 2 reflections in the pentaprism. Can anyone enlighten me or point me to a site that will give a more detail demonstration of the light path in a SLR pentaprism?

- Roy

 

[Peculiar]

To understand how an image is left right corrected using 2 mirrors, one must stand in front of 2 mirrors, where by the 2 mirrors are at perpendicular to each other.
By doing so, you will observe that your image is actually left-right corrected, and not the conventional mirror image.

It appears that the light beams hitting the mirror actually sort off like interchanged so as to give a corrected image..

This also explains why a pentaprism has a roof top

 

[roygoh]

Originally posted by Peculiar
To understand how an image is left right corrected using 2 mirrors, one must stand in front of 2 mirrors, where by the 2 mirrors are at perpendicular to each other.
By doing so, you will observe that your image is actually left-right corrected, and not the conventional mirror image.

It appears that the light beams hitting the mirror actually sort off like interchanged so as to give a corrected image..

This also explains why a pentaprism has a roof top

I understand how my own image can be left-right correct after 2 reflections. That is a different situation from the pentaprism because I started off with an image (myself) that is not inverted!.

The image on the focusing screen as seen from the top of the camera is upright but left-right inverted. The pentaprism has to re-direct this image towards the rear of the camera while making it upright as well as left-right correct.

I am still not convinved that this can be done with only 2 reflections.

Could you elaborate on how the light beam got interchanged to give a corrected image?

- Roy

 

[roygoh]

So my summary (kind of long) of the answer to the original question is:

1. The focusing screen on an SLR is located at the same distance from the reflecting surface of the reflex mirror as the film plane. When the reflex mirror is at 45 degrees to the light path, it forms the image on the focusing screen. When the reflex mirror is lifted and the shutter is opened, the image is formed on the film plane. This enables TTL manual focusing - if the image is focused correctly on the focusing screen, it will be focused correctly on the film plane.

2. To enable eye-level viewing, the image on the focusing screen has to be re-directed towards the rear of the camera. More than 1 reflection is required to maintain an upright and left-right correct image. Traditionally this is achieved by using a pentaprism.

3. The pentaprism cannot be replaced by a single mirror (without additional optics), as more than 1 reflection is required to redirect the image on the focusing screen to the rear of the camera while keeping it upright and left-right correct.

4. The pentaprism can be replaced by some sort of mirror array. It has been done on some budget cameras, as quoted in the link in my earlier post. The performance is not as good because the final image is darker than a pentaprism implementation. This seems to imply that mirrors are not as efficient as prism TIL.

5. To prevent double or even higher number of overlapping images, the mirrors used to replace the pentaprism has to be front surface mirrors. In other words, the reflective coating is on the top surface instead of under a layer of glass like common household mirrors.

6. I believe front surface mirrors will not last as long as pentaprisms.

- Roy

 

[denizenx]

got qn, potential duh:
are lenses still projecting an upside down image nowadays?
is that why there's a discrep here between roy's cannot-correct-inversion and the rest's "periscope design" understanding?

 

[denizenx]

 

[roygoh]

Originally posted by denizenx
got qn, potential duh:
are lenses still projecting an upside down image nowadays?
is that why there's a discrep here between roy's cannot-correct-inversion and the rest's "periscope design" understanding?

That's my belief - SLR lenses should produce an inverted image.

Using the diagram you've posted, which is the most common one found on the internet, the image will be left-right inverted though upright. At least that is what I think, and I am waitiing for someone to wither confirm that or prove me wrong.

 

[ckiang]

AFAIK, lenses ALWAYS produce upside down, laterally inverted images. Rule of optics.

In the most common SLR diagram above.

1. Mirror upside down, l/r inverted on mirror.
2. Image rightside up but l/r inverted on the TOP of the screen.
3. After 1st reflection on the pentaprism, rightside up, l/r inverted.
4. After 2nd reflection, the inversion is corrected and passed to eyepiece.


Regards
CK

 

[denizenx]

but like roy says.. the LR->RL cannot be corrected unless there's a non-flat surface somewhere.. and dun believe they do that cos it will give cowabunga CAs... unless the elements already produce either LRinv or full corrected image lor....

as for aiding the light preservation they silver the pentaprism too, and since the prism is obviously non double-plane (never leave the glass at all right?) there's no double reflection and less light loss through practical TIR..