T Factor?

[cctv_down_under 0]

the true measurement of a lens should be measured in T factor NOT F stop as F stop is only the diameter divided by the focal point which does not equate to the amount of light let in, But the T factor can truly separate 2 lenses with the same quoted F stop as it can differentiate between glass grounding quality.

 

Is this true?

[Cooperman 0]

yep! ish.

 

The f stop is directly related to focal length and diameter of maximum aperture, which is usually dependent on the maximum diameter of the front optic.

 

If I remember correctly, the old 'T' system specifically relates to transmission of light, which can be affected by loads of factors such as glass density / quality, multicoating, number of elements (individual lenses), element profiles etc.

 

f stops are a nice convenient way of describing a relatively simple mathmatical principle, but in practice they don't tell the whole story.

That said, the theory is that f 5.6 on an 8mm lens should be equivalent to f 5.6 on a 25mm lens, but past experience suggests to me that this need not necessarily be the case.

 

Make things simple I say, life's far too short

[cctv_down_under 0]

coops, discussing this with our staff, my disagreement is that we only sell f1.4 lenses, and asking why we do that with such a good low light camera, the response from the experts is that f stop is not important as much if the t factor is good and therefore we do not need them.???

[Cooperman 0]

DVR,

 

I've got a funny feeling this could end up as one of those circular arguments where we end up disappearing up our own .....

 

Lens manufacturers have historically always quoted maximum apertures (f1.x), but in general terms, it's almost unheard of for them to release detailed engineering information relating to light transmission through their optics (I know I have seen 'T' figures quoted recently, but I just can't remember where I've seen them).

 

In a funny way, this is probably no bad thing, considering that many have problems with even basic optical principles, and if yet more information was made available, we'd probably all end up as unbearable lens snobs.

 

The biggest problem with CCTV equipment in general, is we can get so hung up on what's written on the spec. sheet, it's easy to assume that some thing is going to be good (because it says so on paper), without at least taking the trouble to test whether the claims are true.

 

Quite frankly, it doesn't much matter to me whether a lens is f1.2, f1.4, f1.6 or f1.8 unless it is either required to work at maximum aperture in very low light conditions, or a camera is being deliberately used with a high shutter speed to freeze motion, in which case the lens iris may need to operate at its optically inferior maximum setting, just to get a picture.

 

Ideally, any lens will produce its optimum optical quality at or around it's mid aperture setting (normally around f 5.6 ish), so if there is enough light for a lens to operate at that aperture, in theory, it should be pretty much the same result, whether you are using an f1.4 or f1.8 maximum aperture lens.

 

There is no substitute for personal experience, so if you get the chance to test comparable lenses on identical cameras, you can quickly see whether a particular optic is worth the extra bucks or not.

 

If I'm given the choice (and the budget), I will always go for a larger format lens first knowing full well that the optical quality will immediately appear much better (e.g.on a 1/3" CCD camera, using an 8.5mm f1.5 'C' 2/3" format, rather than an 8mm f1.2 'CS' 1/3").

 

If your guys are able to get hold of 'useful' transmission data for any of the major lens manufacturers products, that could be interesting to say the least.

[McJannet 0]

Gents:

 

Please refer to Vlado's book at

 

http://www.cctvlabs.com/Books/book.html

 

This should be sitting on everyone's shelves, you just need to know this stuff sometimes.

 

To quote on the transmittance factor:

 

"The definition of a T-number takes the F-stop and the lens transmittance into account"

 

So, the f-stop is calculated purely on geometry while the t-number takes the lens material quality into account.

 

If you look at the formula, you'll see that there isn't a big difference between the f-stop and the t-number. Limited as we currently are by PAL / NTSC resolution I would suggest that there is no reason to worry about the t-number.

 

Although with my new Nikon D70 6 mega pixel digital SLR camera it is a concern. I bought this camera primarily because of the well known superiority of Nikkor lenses.

 

http://www.dpreview.com/reviews/nikond70/page3.asp

 

Regards:

 

Dave

[Cooperman 0]

So that's where I've been going wrong all these years, I didn't have the book

 

I'm a bit confused McJannet, what has lens light transmittance got to do with the PAL / NTSC signal system, and indeed what relevance does 'T' have to "PAL / NTSC resolution"?

 

I don't think that lens transmittance should be any more a worry on your new Nikon D70, than it was on my Nikon F2S when I bought it in 1977.

 

Happy snapping!

[rory 0]

Well i know for a fact that a f:0.95 lens on my exview GE cameras, sees "much" more in low low light, opposed to the f:1.4 lens i just put on ....which sees "very little" ...

 

Dont have a clue about the "T" factor though ..

 

Rory

[McJannet 0]

The 0.95 lens lets in about 2.17 x the light that the 1.4 lens lets in. From the definition of F-stop.

 

So the 0.95 is more than twice as good as the 1.4, in terms of light gathering ability.

 

The Ex-View is a good, sensitive CCD providing good low light images. Coupled with a good lens like the 0.95 you'll get a good NTSC picture of about 412k pixels. (how good?: manufacturer will give a lux number at 30 or 50 IRE).

 

Which you then compress the buggery out of.

 

The point being, when everyone asks you to zoom in and enhance; you'll need more resolution and less compression.

 

A super lens is great, but the inherent limitations of the sensor and compression currently limit the system by requiring about 1/3 to 1/2 of vertical field to be filled for identification.

 

Three to five megapixel CCD's must be dirt cheap these days, falling in price as fast as Pentium processors. Why not build them into IP PTZ cameras and start taking some serious images? There must be a huge market for this.

 

Regards:

 

Dave

[rory 0]

just add more fixed cameras.

 

i havent seen one good low light megapial cam yet...

 

but yes the manufacturers need to puick up on thiss

[Thomas 0]

Bandwith.

[Cooperman 0]

originally posted by Mc Jannet

 

The Ex-View is a good, sensitive CCD providing good low light images. Coupled with a good lens like the 0.95 you'll get a good NTSC picture of about 412k pixels...

...A super lens is great, but the inherent limitations of the sensor and compression currently limit the system by requiring about 1/3 to 1/2 of vertical field to be filled for identification

 

Forgive me for being picky, but the 'NTSC' bit refers to the colour system, and doesn't relate to the specific resolution of a camera. So you can have 330 line, 460 line, 520 line (or greater) 'NTSC' cameras; likewise the pixel count for the imager is quite obviously not going to be affected by the maximum aperture of a lens.

 

A 412k imager will still be a 412k imager whether you use an f0.95 or f1.4 lens.

 

If a manufacturer (or expert) states that the f0.95 lens has 2.17x more light gathering ability than an f1.4, far be it from me to argue. That said, coming back to DVR Expert's original point, whilst the larger aperture will theoretically offer this huge jump in light gathering ability, if the lens elements, coatings and design are total c**p, most of that apparent advantage will count for little when looked at in relation to light transmission through the lens.

 

It's also worth remembering that at larger apertures, overall optical performance is somewhat reduced, and the 'depth of field' will be minimal.

 

As regards the percentage of target required to occupy the picture for effective identification, this problem is generally nothing to do with simple engineering principles, but in most cases, the camera / lens combination are either incorrect, located in the wrong place, or set up in a less than ideal way; and that is down to lack of knowledge or experience on the part of the installer.

[cctv_down_under 0]

5 mega IP cams have been around for while, but Thomas hit the nail on the head... Bandwidth!!!!

[Thomas 0]

Realisticly you could do it now by doing a seperat network with the DVR grabbing it through a gateway. The bottleneck is framerate, but that's nothing new.

[Cooperman 0]

Light transmission through optics .... and network bandwidth restrictions ... all in one thread

[Thomas 0]

And soon when someone finds an easier way to splice fiber, it will all be the same.

[C7 in CA 0]

And soon when someone finds an easier way to splice fiber, it will all be the same.

 

There are no-polish fiber solutions that are easy. It basically just butt splices the pre-polished connectors onto your fiber strands. The mechanical splice kit is around 1000 bucks. The fusion splicers start around 15000 bucks.

 

So optical fiber can be easy if you are using the tight buffered fiber. Stay away from the loose tube stuff! it requires threading each individual fiber into a jacketed fanout kit. Big pita.

[Thomas 0]

Doesn't that still take a while?

[C7 in CA 0]

Well I cant vouch for the camlock type systems because I haven't used them. But I have used Levitons Threadlock system which is a mechanical connector (vs. an epoxy connector) and I can reliably connectorize an optical strand in about 5 minutes. But it takes another 10 or 15 minutes for me to polish the connector. Probably closer to 10 minutes though. I'm sure if you worked fiber everyday you could be much faster. I notice I'm much faster on my last connector of a job then I was on the first.

 

So, with no polish It probably takes about 5 minutes per strand. Not bad. That's only about twice as long as a Cat5 jack. Of course there is some cable prep time before you can terminate the fiber. Plus you have to put it all in a nice fiber distribution can. Can't just let it all hang out like you could with coax.

[Cooperman 0]

Originally posted by C7 in CA:

I notice I'm much faster on my last connector of a job then I was on the first.

 

Ain't that the truth! I once spent two weeks solid terminating loose tube fibres - and that was when the epoxy had to cure for a minimum of 24 hours before you could polish it off.

 

At the end of the job I realised there's a much easier way to do it .... sub contract!!

[cctv_down_under 0]

Picky Picky