Increasing camera low light and range performance

Frank Granovski · June 21st, 2003, 01:40 PM

http://www.dvfreak.com/lux.htm

Charlie Higgins · June 21st, 2003, 10:13 PM

I work as an instructor at a local University. I printed out this thread and took it to several trusted associates: Engineering Professors (and a Dean who is also a cutting-edge optics consultant); folks who, collectively, have over 100 years of experience and knowledge in Optical, Electrical, and Computer Engineering. They all gave me the same answer. Some professed being mystified at why there was a question about this in the first place.

Put simply, you cannot gain LUX without actually replacing the lens, the CCD, the electronics, or all three. You cannot ADD a lens, any lens of any design, and increase LUX sensitivity.

I suggest you go to a good University with this printed thread in hand. Ask someone in Optical Engineering, etc. to take a look at your argument. They will tell you, based on universally respected FACTS of physics, why this is impossible as you argue it.

Vladimir Koifman · June 22nd, 2003, 05:00 AM

I'm still feel unconvinced about the technical side of this discussion. Let me put things upfront - I have absolute zero experience in optics. But even risking to lose my face in this discussion, I still like to learn something.

Let's look on telescope stricture. It has two basic peices - the large diameter lens-objective and small eyepiece. The eyepice is just like camera lens, just optimized for viewing by human eye, instead of CCD. It's also interchangable (usually).
Now, the larger the main lens-objective, the better telescope ability to see weak stars.
Assume someone attaches video camera to the telescope instead of eyepeice. Assume that front lens or mirror of the telescope is very large, say 10 feets. Does this improve the camera low-light abilities?
I'm understand the focus length will be huge and that front lens would make upside-down picture. If somebody values low lux abilities above all, he might use his camcorder upside down to compensate - just joking.

Charlie, you know, after you referenced to professors and 100 years of combined experience, I'm not sure in anything. I'm ready to lose this argument. Please, just explain me where is my mistake.

Tom Hardwick · June 22nd, 2003, 05:41 AM

No mistake Valimir, just a misunderstanding of the way a lens's maximum aperture is formulated.

If you have a 50mm lens and it has a maximum aperture of f2 then the diaphragm blades (the 'aperture' through which all the light must pass) has to be 25mm in diameter. This is optical law and is a simple mathematical equation. If you look at this argument you'll realise that two identical 50mm lenses could both be f2 yet one could be made with the purest glass to be found, the other made with scratched, yellowing plastic.

They'd both be f2 lenses, but one would transmit a lot more light than the other. Welcome the T stop (T for transmission). This is a much more accurate way of measuring the light that exits the lens on the way to the chip, and is the way professionals set up their shots.

Sony have designed the 12x zoom on the VX to have a maximum aperture of f1.6 at the 6mm end. But with production tolerances, 13 elements, three beam splitting prisms and two plane parallel elements in the OIS you'll realise that the T stop is probably some way away from f1.6.

No matter. Increasing the size of the front element (as in your telescope analogy) won't affect the T stop of the lens, only a complete re-design could do that. In the same way adding a wide-angle converter won't help either - in fact it will very slightly decrease the amount of light that finally exits the lens.

tom.

Vladimir Koifman · June 22nd, 2003, 08:17 AM

Tom,

Thank you for your reply. I understand what you are saying about loss of light in the lens elements. So the light loss is unevitable. Theoretically, this can be improved by multilevel coating, if lens cost is no issue. For example, some Hoya UV filters promise 97% of light transparancy. Also, mirrors can be used instead of glasses.

But there is one thing in your explanations I don't understand. In telescopes the eyepiece size is something more-less simalar in all telescopes. I mean, even huge telescopes have relatively small eyepiece. However the bigger the front lens or mirror, the weaker stars the telescope can detect - at given eyepiece size.
I understand there is a flaw in my speculations. In the end, if everybody says I'm wrong, most probably I am. I'm just trying to find a peace of mind and understand why adapters (wide or tele) with lens do not work in a same manner as telescope eyepiece-front end.
Please help me on this.

Tom Hardwick · June 22nd, 2003, 09:27 AM

Yes, multi-layer coating can indeed improve light transmission. Plain uncoated glass reflects back about 8% whereas as you say MC coating can mean that only 0.2% gets reflected (at each air to glass surface) - all the rest goes through. I would suspect that mirrors suffer the same losses as glass, even sufrace silvered ones. Nothing is perfect.

I'm not a telescope man but I'd suspect that larger front elements on telescopes mean more light gathering power, in the same way as wide aperture photographic lenses have bigger front elements. This light is then passed through the eyepiece to the human eye which can vary its own iris diaphragm to expose the retina correctly. If you look through a pinhole all will be dark as you eye is at it's max f1.4 and still there's not enough light to see by. Look through a huge telescope and the eye is fed enough information to sensitise the retina (or film or CCD, whatever).

tom.

Vladimir Koifman · June 22nd, 2003, 09:53 AM

Tom,

Now you put it in even more straight way than I did.
A human eye have f1.4 aperture - I did not know this, thank you for the info.
The ability of eye to see weak stars is extended by telescope, right?
So, if one attaches camcorder instead of eye to the telescope eyepiece, the camcorder will see weaker stars as well, right?
I feel I messed up things somewhere, but I don't see where. Any help?

Tom Hardwick · June 22nd, 2003, 09:57 AM

Absolutely correct.

Vladimir Koifman · June 22nd, 2003, 10:10 AM

Tom,

Does this mean that if one buys telescope as a tele-adapter for his camera (albeit huge, expensive and inconvenient), one can improve the camera low-light capability? I still think there is a catch here, it cant be that professors are wrong here.

Charlie Higgins · June 22nd, 2003, 11:25 AM

Tom's quote, " I'm not a telescope man but I'd suspect that larger front elements on telescopes mean more light gathering power, in the same way as wide aperture photographic lenses have bigger front elements."...This light is then passed through the eyepiece to the human eye which can vary its own iris diaphragm to expose the retina correctly."

You forget that Wayne's argument was that NO INTERNAL CAM MODIFICATION is ncessary, save adding an attachment Lens system. Where does this 'wide aperture photographic lens' obtain its 'wide aperture? FROM AN INTERNAL MODIFICATION, that's where.

Vladimir siad:
"Now you put it in even more straight way than I did.
A human eye have f1.4 aperture - I did not know this, thank you for the info.
The ability of eye to see weak stars is extended by telescope, right?
So, if one attaches camcorder instead of eye to the telescope eyepiece, the camcorder will see weaker stars as well, right?
I feel I messed up things somewhere, but I don't see where. Any help?"

Tom: "Absolutely correct."

Um, no.

The cam cannot change how it 'sees' in the same manner the human eye does so reflexively WITHOUT BEING MODIFIED at its lens, CCD, electronics, and/or software. Without said cam modifications, it is technically impossible to increase the cam's LUX sensitivity. There are losses at every step, from Telescopic front lens to CCD, and these losses are what prevent a sum LUX gain by only adding more glass to the front of any cam.

This, Vladimir, is precisely why it will always remain a tantalizing theory and not be a practical solution- cams are not human eyes, and even if they were, 'attaching' a cam to a huge telescope is tantamount to attaching the cam to an external light source, rather than an 'add-on' lens system. In this case, it's the CAM that's being added to the light source, not the other way around.

Good luck on your search for the answers, fellas...

Charlie Higgins · June 22nd, 2003, 11:46 AM

Oh, yeah...as I recall, basic physics would also dictate that, no matter how much a telescopic device 'gathers light' it's in the service of 'passive transmission, i.e., a small star appears bigger, but the light-gathering serves the magnification of the image, not the modifcation of the device attahced to it- a cam, etc. In other words, at some point, the cam attached to a telescope would be the BOTTLENECK to its light transmission back to a CCD, including whatever technical limitations are built into the CCD itself!

My final 2 cents as a 'lay' observer...

Chris Hurd · June 22nd, 2003, 11:59 AM

Howdy from Texas,

<< a telescopic device... i.e., a small star appears bigger, >>

As an amateur astronomer who has built his own Newtonian reflector (a type of telescope), I can assure you that telescopes do not make small stars appear bigger. All stars (with the exception of our own sun) appear through all telescopes (without exception, including Hubble) as tiny points of light with the single dimension of brightness. Some are brighter than others, which is usually an indicator of age and flavor, not size... and some resolve into binaries (twin stars)... and some resolve into double binaries (two sets of twins). But they're all just points of light, no matter what telescope you're using, no difference in "size," only in brightness.

To steer this back on topic, please note that I've had to edit a few recent responses. It is a rule here that we have NO FLAMING of any kind, so please be nice to each other. Daddy is watching. Many thanks,

Charlie Higgins · June 22nd, 2003, 01:58 PM

See, this is what I mean. I'm not a 'Telescope Guy', either, Tom. Glad to see a reply from one though!

I'll add one more thing: that adding any apparatus to a cam doesn't make it inherently more sensitive to light; it may, at best, increase the amount of light transmitted to the cam itself. This only means that the cam, once again, is the bottleneck of light sensitivity- you ARE stuck with the LUX rating of yer cam until you ebay it and buy a more sensitive one!

Point taken on the 'no flaming' Chris- I get a little carried away sometimes...again, I'm demonstrating a firm grasp of the obvious!

Tom Hardwick · June 22nd, 2003, 11:04 PM

Thank you Chris for calming the flames. I saw the relevant post and thought it harsh. And hey Charlie, I thought I had put up a defence of what I was about to write when I started the paragraph with, "I'm not a telescope man but I'd suspect that ... " I certainly didn't mean to mislead, but bow to your greater knowledge on the subject. And I hope Valadimir now knows that adding anything in front of his front element will soak a bit of light...

tom.

Charlie Higgins · June 23rd, 2003, 12:16 AM

Sorry, Tom. Point well taken; I looked back over the posts and indeed, that's easy to see now...

8-month old son with MAJOR sleep problems got me CCI (Cranky, Confused, and Ill-mannered)...

Thanks for setting me striaght, guys.

*yawn, sniff*

June 22nd, 2003, 11:59 AM	#42
Chris Hurd Obstreperous Rex Join Date: Jan 2001 Location: San Marcos, TX Posts: 27,368 Images: 513	Howdy from Texas, << a telescopic device... i.e., a small star appears bigger, >> As an amateur astronomer who has built his own Newtonian reflector (a type of telescope), I can assure you that telescopes do not make small stars appear bigger. All stars (with the exception of our own sun) appear through all telescopes (without exception, including Hubble) as tiny points of light with the single dimension of brightness. Some are brighter than others, which is usually an indicator of age and flavor, not size... and some resolve into binaries (twin stars)... and some resolve into double binaries (two sets of twins). But they're all just points of light, no matter what telescope you're using, no difference in "size," only in brightness. To steer this back on topic, please note that I've had to edit a few recent responses. It is a rule here that we have NO FLAMING of any kind, so please be nice to each other. Daddy is watching. Many thanks, __________________ CH Search DV Info Net \| 20 years of DVi \| ...Tuesday is Soylent Green Day!

June 23rd, 2003, 12:16 AM	#45
Charlie Higgins New Boot Join Date: Jun 2003 Location: Amherst, MA Posts: 10	light loss, brain loss... Sorry, Tom. Point well taken; I looked back over the posts and indeed, that's easy to see now... 8-month old son with MAJOR sleep problems got me CCI (Cranky, Confused, and Ill-mannered)... Thanks for setting me striaght, guys. yawn, sniff

June 21st, 2003, 01:40 PM	#31
Frank Granovski Outer Circle Join Date: May 2002 Location: Hope, BC Posts: 7,524	http://www.dvfreak.com/lux.htm

June 21st, 2003, 10:13 PM	#32
Charlie Higgins New Boot Join Date: Jun 2003 Location: Amherst, MA Posts: 10	I work as an instructor at a local University. I printed out this thread and took it to several trusted associates: Engineering Professors (and a Dean who is also a cutting-edge optics consultant); folks who, collectively, have over 100 years of experience and knowledge in Optical, Electrical, and Computer Engineering. They all gave me the same answer. Some professed being mystified at why there was a question about this in the first place. Put simply, you cannot gain LUX without actually replacing the lens, the CCD, the electronics, or all three. You cannot ADD a lens, any lens of any design, and increase LUX sensitivity. I suggest you go to a good University with this printed thread in hand. Ask someone in Optical Engineering, etc. to take a look at your argument. They will tell you, based on universally respected FACTS of physics, why this is impossible as you argue it.

June 22nd, 2003, 05:00 AM	#33
Vladimir Koifman Regular Crew Join Date: May 2003 Location: Israel Posts: 115	I'm still feel unconvinced about the technical side of this discussion. Let me put things upfront - I have absolute zero experience in optics. But even risking to lose my face in this discussion, I still like to learn something. Let's look on telescope stricture. It has two basic peices - the large diameter lens-objective and small eyepiece. The eyepice is just like camera lens, just optimized for viewing by human eye, instead of CCD. It's also interchangable (usually). Now, the larger the main lens-objective, the better telescope ability to see weak stars. Assume someone attaches video camera to the telescope instead of eyepeice. Assume that front lens or mirror of the telescope is very large, say 10 feets. Does this improve the camera low-light abilities? I'm understand the focus length will be huge and that front lens would make upside-down picture. If somebody values low lux abilities above all, he might use his camcorder upside down to compensate - just joking. Charlie, you know, after you referenced to professors and 100 years of combined experience, I'm not sure in anything. I'm ready to lose this argument. Please, just explain me where is my mistake.

June 22nd, 2003, 05:41 AM	#34
Tom Hardwick Inner Circle Join Date: Mar 2003 Location: Billericay, England UK Posts: 4,711	No mistake Valimir, just a misunderstanding of the way a lens's maximum aperture is formulated. If you have a 50mm lens and it has a maximum aperture of f2 then the diaphragm blades (the 'aperture' through which all the light must pass) has to be 25mm in diameter. This is optical law and is a simple mathematical equation. If you look at this argument you'll realise that two identical 50mm lenses could both be f2 yet one could be made with the purest glass to be found, the other made with scratched, yellowing plastic. They'd both be f2 lenses, but one would transmit a lot more light than the other. Welcome the T stop (T for transmission). This is a much more accurate way of measuring the light that exits the lens on the way to the chip, and is the way professionals set up their shots. Sony have designed the 12x zoom on the VX to have a maximum aperture of f1.6 at the 6mm end. But with production tolerances, 13 elements, three beam splitting prisms and two plane parallel elements in the OIS you'll realise that the T stop is probably some way away from f1.6. No matter. Increasing the size of the front element (as in your telescope analogy) won't affect the T stop of the lens, only a complete re-design could do that. In the same way adding a wide-angle converter won't help either - in fact it will very slightly decrease the amount of light that finally exits the lens. tom.

June 22nd, 2003, 08:17 AM	#35
Vladimir Koifman Regular Crew Join Date: May 2003 Location: Israel Posts: 115	Tom, Thank you for your reply. I understand what you are saying about loss of light in the lens elements. So the light loss is unevitable. Theoretically, this can be improved by multilevel coating, if lens cost is no issue. For example, some Hoya UV filters promise 97% of light transparancy. Also, mirrors can be used instead of glasses. But there is one thing in your explanations I don't understand. In telescopes the eyepiece size is something more-less simalar in all telescopes. I mean, even huge telescopes have relatively small eyepiece. However the bigger the front lens or mirror, the weaker stars the telescope can detect - at given eyepiece size. I understand there is a flaw in my speculations. In the end, if everybody says I'm wrong, most probably I am. I'm just trying to find a peace of mind and understand why adapters (wide or tele) with lens do not work in a same manner as telescope eyepiece-front end. Please help me on this.

June 22nd, 2003, 09:27 AM	#36
Tom Hardwick Inner Circle Join Date: Mar 2003 Location: Billericay, England UK Posts: 4,711	Yes, multi-layer coating can indeed improve light transmission. Plain uncoated glass reflects back about 8% whereas as you say MC coating can mean that only 0.2% gets reflected (at each air to glass surface) - all the rest goes through. I would suspect that mirrors suffer the same losses as glass, even sufrace silvered ones. Nothing is perfect. I'm not a telescope man but I'd suspect that larger front elements on telescopes mean more light gathering power, in the same way as wide aperture photographic lenses have bigger front elements. This light is then passed through the eyepiece to the human eye which can vary its own iris diaphragm to expose the retina correctly. If you look through a pinhole all will be dark as you eye is at it's max f1.4 and still there's not enough light to see by. Look through a huge telescope and the eye is fed enough information to sensitise the retina (or film or CCD, whatever). tom.

June 22nd, 2003, 09:53 AM	#37
Vladimir Koifman Regular Crew Join Date: May 2003 Location: Israel Posts: 115	Tom, Now you put it in even more straight way than I did. A human eye have f1.4 aperture - I did not know this, thank you for the info. The ability of eye to see weak stars is extended by telescope, right? So, if one attaches camcorder instead of eye to the telescope eyepiece, the camcorder will see weaker stars as well, right? I feel I messed up things somewhere, but I don't see where. Any help?

June 22nd, 2003, 09:57 AM	#38
Tom Hardwick Inner Circle Join Date: Mar 2003 Location: Billericay, England UK Posts: 4,711	Absolutely correct.

June 22nd, 2003, 10:10 AM	#39
Vladimir Koifman Regular Crew Join Date: May 2003 Location: Israel Posts: 115	Tom, Does this mean that if one buys telescope as a tele-adapter for his camera (albeit huge, expensive and inconvenient), one can improve the camera low-light capability? I still think there is a catch here, it cant be that professors are wrong here.

June 22nd, 2003, 11:25 AM	#40
Charlie Higgins New Boot Join Date: Jun 2003 Location: Amherst, MA Posts: 10	Tom's quote, " I'm not a telescope man but I'd suspect that larger front elements on telescopes mean more light gathering power, in the same way as wide aperture photographic lenses have bigger front elements."...This light is then passed through the eyepiece to the human eye which can vary its own iris diaphragm to expose the retina correctly." You forget that Wayne's argument was that NO INTERNAL CAM MODIFICATION is ncessary, save adding an attachment Lens system. Where does this 'wide aperture photographic lens' obtain its 'wide aperture? FROM AN INTERNAL MODIFICATION, that's where. Vladimir siad: "Now you put it in even more straight way than I did. A human eye have f1.4 aperture - I did not know this, thank you for the info. The ability of eye to see weak stars is extended by telescope, right? So, if one attaches camcorder instead of eye to the telescope eyepiece, the camcorder will see weaker stars as well, right? I feel I messed up things somewhere, but I don't see where. Any help?" Tom: "Absolutely correct." Um, no. The cam cannot change how it 'sees' in the same manner the human eye does so reflexively WITHOUT BEING MODIFIED at its lens, CCD, electronics, and/or software. Without said cam modifications, it is technically impossible to increase the cam's LUX sensitivity. There are losses at every step, from Telescopic front lens to CCD, and these losses are what prevent a sum LUX gain by only adding more glass to the front of any cam. This, Vladimir, is precisely why it will always remain a tantalizing theory and not be a practical solution- cams are not human eyes, and even if they were, 'attaching' a cam to a huge telescope is tantamount to attaching the cam to an external light source, rather than an 'add-on' lens system. In this case, it's the CAM that's being added to the light source, not the other way around. Good luck on your search for the answers, fellas...

June 22nd, 2003, 11:46 AM	#41
Charlie Higgins New Boot Join Date: Jun 2003 Location: Amherst, MA Posts: 10	Oh, yeah...as I recall, basic physics would also dictate that, no matter how much a telescopic device 'gathers light' it's in the service of 'passive transmission, i.e., a small star appears bigger, but the light-gathering serves the magnification of the image, not the modifcation of the device attahced to it- a cam, etc. In other words, at some point, the cam attached to a telescope would be the BOTTLENECK to its light transmission back to a CCD, including whatever technical limitations are built into the CCD itself! My final 2 cents as a 'lay' observer...

June 22nd, 2003, 01:58 PM	#43
Charlie Higgins New Boot Join Date: Jun 2003 Location: Amherst, MA Posts: 10	See, this is what I mean. I'm not a 'Telescope Guy', either, Tom. Glad to see a reply from one though! I'll add one more thing: that adding any apparatus to a cam doesn't make it inherently more sensitive to light; it may, at best, increase the amount of light transmitted to the cam itself. This only means that the cam, once again, is the bottleneck of light sensitivity- you ARE stuck with the LUX rating of yer cam until you ebay it and buy a more sensitive one! Point taken on the 'no flaming' Chris- I get a little carried away sometimes...again, I'm demonstrating a firm grasp of the obvious!

June 22nd, 2003, 11:04 PM	#44
Tom Hardwick Inner Circle Join Date: Mar 2003 Location: Billericay, England UK Posts: 4,711	Thank you Chris for calming the flames. I saw the relevant post and thought it harsh. And hey Charlie, I thought I had put up a defence of what I was about to write when I started the paragraph with, "I'm not a telescope man but I'd suspect that ... " I certainly didn't mean to mislead, but bow to your greater knowledge on the subject. And I hope Valadimir now knows that adding anything in front of his front element will soak a bit of light... tom.