Newly used LED lighting

[Mark von Kanel]

Sorry WM , we got off topic ! but as i said earlier as long as they put out a constant spectrum (that the camera sees) and you correct using the passport it makes no difference which you use does it?

its just a case of what you have to hand or which you prefer to work with. or cost or anything other than the colour of the light really.

[ajohnw]

It does ,make some difference that may matter. That's what CRI is all about. Colour Rendering Index. It relates to the fact that these light sources don't have continuous spectrums. The higher the CRI the better, hence photo compact fluorescence bulbs. They use more phosphors so the spectrum is more consistent rather than spiky. The ones I linked to on ebay are >90 but I vaguely remember that they are available up to around 95. The problem is that different colour reflect differently so it needn't be just a case of having the usual colour balance problems. Single shades can be out and the rest ok. The reason for RGGB led's as against ordinary RGB led's is the appearance of certain colours when things are illuminated with them.. They use 2 shades of green rather than just one.

One of the problems with high CRI led's is that they have to emit rather a lot of blue - of interest to people who have to pay government agencies for street lighting - some fit high CRI that cost more to run. Rather a lot more actually.

John
-

[Cliff Roamer]

You guys are doing nothing but confusing me.

Me to

Just why are these LED's so expensive, £300, apart from being made by Manfrotto and the mounting is there anything special about them?



Hybrid+ LED, the 84 LEDs emit continuous light at a temperature of 5600°K (daylight), 294 lumens with a luminance of 850 LUX, at a distance of 1 m.

[Inkanyezi]

Terms that most people don't know the meaning of are tossed around. "CRI" is in fact two different entities altogether, and it is not a physical property, but a subjectively evalutated property. The CRI of a light source is evaluated by a board of people with supposedly good colour vision. It is done with a standardised colour map, not very unlike the colour passport. The board of viewers set a number to the likeness of the colour of each patch as perceived when illuminated with the light source under trial, and a standard light source. The average of those numbers for all patches and all viewers, expressed in hundredths, is the CRI for the light source tested.

The standard for colour temperatures of 5000 K and above is "daylight".
Daylight is an elusive beast. Natural daylight is ever changing, and the standard has to be artificial, with a constant output. Natural daylight, our real daylight, does not have a smooth spectrum (often referred to as "continuous"), which does not necessarily include all visible wavelengths. The standard light source for daylight is usually fluorescent, but also filtered incandescent can mimick daylight. For this a neodymium oxide layer is used to make the spectrum less smooth and more like daylight, and a blue filter to raise colour temperature by cutting red and yellow.

So in fact we do not want a "continuous" spectrum if we wish to render colours as daylight does.

The CRI standard for colour temperatures below 5000 K is a black body.
Any incandescent filament is a close replacement for the black body. That is the reason why incandescent has a high CRI, close to 100. The usual figure mentioned is 99, but you could regard any incandescent light source as CRI 100.

An incandescent light source, CRI 100, cannot, and should not, render colours "truly" as daylight does. The primary reason is that incandescent has a smooth spectrum comprising all wavelengths, with its peak in the red and yellow region, but it is very poor in blue and violet as well. There are also deficiencies in the green part of the spectrum, compared to the standard for daylight. The differences in colour rendition compared to daylight are what we call "metameric mismatch". By passing incandescent through a neodymium filter, we even out that metameric mismatch and can indeed make the incandescent light render "true" colour, resembling daylight. A simple blue filter, the Wratten 82 series does not do that, but can complement the neodymium filter to balance colour temperature to a higher value.

But it does not end there. Our very eyes are not equally sensitive to all wavelengths. Our colour sensitive receptors will not respond equally to the energy of all wavelengths. This is purportedly corrected in our brain, but in essence, we have three kinds of receptors, most sensitive to a rather broad band of wavelengths around red, green and blue. Those are the same colours that are filtered by the Bayer filter in our cameras, and the same as emitted from our computer screens; the three additive colours. There are differences in perception between different people, not only that some may be colour blind, but we don't see the same colours as our neighbours do.

The CRI chart of colour patches are therefore compared, subjectively, by a board of observers, setting a number to just how much likeness between the one lit by the source to be evaluated and the one lit by the standard source. Smoothness of spectrum or continuity is disregarded here, it has no meaning in a discussion about quality of light and colour rendering indexes.

And our images, as long as looked at on a screen, will always be discontinuous, comprised of basically three colours, but light in nature as well will be received by our brain as those same three colours, because those are the colours to which our receptors in the eye are tuned.

If you wish "true" colour, such as perceived when your subject is illuminated by "daylight", you must use a light source of 5000 K or above with a high CRI, but if you want "true" colour, as perceived when illuminated by incandescent, you must use an incandescent source or one that mimicks incandescent sufficiently well. In the latter case it would not always need to be continuous, but have the same mix of the three basic colours as incandescent. However, fluorescence from your subject may cause metameric mismatch, when the spectrum is not exactly the same, as the colour from the illuminated surface does not necessarily correspond to the colour of the light source. Particularly when there is a substantial portion of light in the blue to violet part of the spectrum, fluoresced colour of lower wavelength from illuminated surfaces may be present.

Whether a lamp is LED or fluorescent tube is irrelevant, but its CRI is the crucial factor – together with colour temperature. A lamp with colour temperature below 5000 K cannot, and should not, render colours similar to daylight, unless its CRI is much lower than 100. A neodymium filter usually gives the incandescent lamp a CRI of between 70 and 80. Its crooked spectral curve then will be similar to that of "daylight", the standardised daylight which is compared to the light emitted from our closest star, the Sun, a hydrogen fusion nuclear reactor, filtered through the atmosphere. The Sun's radiation is not smooth even in outer space, and our atmosphere scatters it and cuts off a variety of wavelengths as well as emphasising other wavelengths.

[Colin Southern]

Trying out LED lights...color temp seems to be about 2700 kelvin using Color Checker Passbook.
What say you about the colors of the artificial flowers?

William,

Love you like family buddy, but I've just gotta come right out and say it - that image is awful; the hard shadows and poor tone control make it hard to evaluate the colour, and from what I can see, the orange flower to center right at the bottom probably has too much blue in it, but it's hard to tell.

Sorry mate

[AB26]

[*] Colour: Warm White

A very bad choice for Photography!

You should try the Cool Daylight.

[AB26]

You should be looking for something like this

http://www.ebay.co.uk/itm/105W-eq-52...item54068f0f7c

But ideally with a better cri rating.

John
-

I will definitely get some of those!

Thanks John.

[AB26]

William,

Some LED shots:

The tool I used: only 55 lumens, 3XAAA batteries = 130 hours. It is part of my camera kit, always have it in my bag.
WB was set to 2700K to compensate for down lighters. (Don't ask me what tipe of light it is, I don't care as long as I get the WB right.)


In this next shot I left the K setting at 2700 to illustrate the difference in colour temp between the down lighters and the LED.
Notice how blue the light is from the LED.



Next shot: Horrible, tripod mounted but I had to touch the tripod to stabalise the light.
Shot at 2700 K



In the next shot I set K to 5600. The colour rendition is pretty accurate considering the in camera settings:



The value from this post of yours: There is a lot that can be done with portable LED lights.

A Led Lenser X21 (1000 Lumens) will light up a mountain.
How will you light up a mountain with 18 W energy saving bulbs?

Oh yes - all the shots are SOOC Jpegs.

[ajohnw]

I would disagree with Kathy about LED or fluorescent tube being irrelevant, They are entirely different animals. One uses visible light to excite phosphors and the other doesn't. A LED lets this scarcely visible blue out as part of lt's light. Also in LED terms so called warm white is generally around 5k, daylight 6.5k. That requires and awful lot of this blue.

This all relates to something that has been around for a long time really - viewing conditions, monitor or prints. The difference is that it's the camera that is doing the viewing so in some respects CRI is irrelevant other than higher values MAY indicate a more even spectrum in the light output. A better guide really is things like the words tri/multi phosphor on fluorescent systems or that they are intended for photographic illumination applications.

May be of interest. Some info on viewing conditions here. Some of the links are dated so search for what ever the link states on the sites when not found pops up. There are details of the ISO standards on one site.

http://www.northlight-images.co.uk/a..._lighting.html

I did wonder about sorting my room lighting out.

LED's come in for portable applications but on the other hand does it really matter if colours shift. As far as I am aware most of these come with warm white led's not 6.5k daylight types.

John
-

[ajohnw]

Me to

Just why are these LED's so expensive, £300, apart from being made by Manfrotto and the mounting is there anything special about them?



Hybrid+ LED, the 84 LEDs emit continuous light at a temperature of 5600°K (daylight), 294 lumens with a luminance of 850 LUX, at a distance of 1 m.

They aren't power led's by the look of it so certain aspects such as the blue colour that comes out of power led's go out of the window. Some can even have the colour outputs individually adjusted.

John
-

[Cliff Roamer]

Terms that most people don't know the meaning of are tossed around. "CRI" is in fact two different entities altogether, and it is not a physical property, but a subjectively evalutated property. The CRI of a light source is evaluated by a board of people with supposedly good colour vision. It is done with a standardised colour map, not very unlike the colour passport. The board of viewers set a number to the likeness of the colour of each patch as perceived when illuminated with the light source under trial, and a standard light source. The average of those numbers for all patches and all viewers, expressed in hundredths, is the CRI for the light source tested.

The standard for colour temperatures of 5000 K and above is "daylight".
Daylight is an elusive beast. Natural daylight is ever changing, and the standard has to be artificial, with a constant output. Natural daylight, our real daylight, does not have a smooth spectrum (often referred to as "continuous"), which does not necessarily include all visible wavelengths. The standard light source for daylight is usually fluorescent, but also filtered incandescent can mimick daylight. For this a neodymium oxide layer is used to make the spectrum less smooth and more like daylight, and a blue filter to raise colour temperature by cutting red and yellow.

So in fact we do not want a "continuous" spectrum if we wish to render colours as daylight does.

The CRI standard for colour temperatures below 5000 K is a black body.
Any incandescent filament is a close replacement for the black body. That is the reason why incandescent has a high CRI, close to 100. The usual figure mentioned is 99, but you could regard any incandescent light source as CRI 100.

An incandescent light source, CRI 100, cannot, and should not, render colours "truly" as daylight does. The primary reason is that incandescent has a smooth spectrum comprising all wavelengths, with its peak in the red and yellow region, but it is very poor in blue and violet as well. There are also deficiencies in the green part of the spectrum, compared to the standard for daylight. The differences in colour rendition compared to daylight are what we call "metameric mismatch". By passing incandescent through a neodymium filter, we even out that metameric mismatch and can indeed make the incandescent light render "true" colour, resembling daylight. A simple blue filter, the Wratten 82 series does not do that, but can complement the neodymium filter to balance colour temperature to a higher value.

But it does not end there. Our very eyes are not equally sensitive to all wavelengths. Our colour sensitive receptors will not respond equally to the energy of all wavelengths. This is purportedly corrected in our brain, but in essence, we have three kinds of receptors, most sensitive to a rather broad band of wavelengths around red, green and blue. Those are the same colours that are filtered by the Bayer filter in our cameras, and the same as emitted from our computer screens; the three additive colours. There are differences in perception between different people, not only that some may be colour blind, but we don't see the same colours as our neighbours do.

The CRI chart of colour patches are therefore compared, subjectively, by a board of observers, setting a number to just how much likeness between the one lit by the source to be evaluated and the one lit by the standard source. Smoothness of spectrum or continuity is disregarded here, it has no meaning in a discussion about quality of light and colour rendering indexes.

And our images, as long as looked at on a screen, will always be discontinuous, comprised of basically three colours, but light in nature as well will be received by our brain as those same three colours, because those are the colours to which our receptors in the eye are tuned.

If you wish "true" colour, such as perceived when your subject is illuminated by "daylight", you must use a light source of 5000 K or above with a high CRI, but if you want "true" colour, as perceived when illuminated by incandescent, you must use an incandescent source or one that mimicks incandescent sufficiently well. In the latter case it would not always need to be continuous, but have the same mix of the three basic colours as incandescent. However, fluorescence from your subject may cause metameric mismatch, when the spectrum is not exactly the same, as the colour from the illuminated surface does not necessarily correspond to the colour of the light source. Particularly when there is a substantial portion of light in the blue to violet part of the spectrum, fluoresced colour of lower wavelength from illuminated surfaces may be present.

Whether a lamp is LED or fluorescent tube is irrelevant, but its CRI is the crucial factor – together with colour temperature. A lamp with colour temperature below 5000 K cannot, and should not, render colours similar to daylight, unless its CRI is much lower than 100. A neodymium filter usually gives the incandescent lamp a CRI of between 70 and 80. Its crooked spectral curve then will be similar to that of "daylight", the standardised daylight which is compared to the light emitted from our closest star, the Sun, a hydrogen fusion nuclear reactor, filtered through the atmosphere. The Sun's radiation is not smooth even in outer space, and our atmosphere scatters it and cuts off a variety of wavelengths as well as emphasising other wavelengths.

I rely on people who understand all that to simply tell me, 'Yep, that'll work'

[Cliff Roamer]

They aren't power led's by the look of it so certain aspects such as the blue colour that comes out of power led's go out of the window. Some can even have the colour outputs individually adjusted.

John
-

Ah, thanks, maybe some of the advertising hype is for real.

[ajohnw]

I wouldn't trust it. One thing that is very noticeable in many things that contain these is that the light is very very blue. Done as it's a way of getting more power out so the numbers look better.

John
-

[Inkanyezi]

Of course the whole thing boils down to the quality of the light source used. Still we have no LED sources that have as good rendition as fluorescents, but anyway it is misleading to talk about an even, smooth or continuous spectrum. The best light sources for daylight all have discontinuous spectra, just as natural daylight has as well, although they may have peaks that are not present in the natural daylight.

It is a matter of tricolour response, as the important blend of wavelengths can be built up from the three colours, to which our colour sensors in the eye are tuned.

There are indeed LED bulbs that closely match incandescent in colour temperature, although not spectrum caracteristics, as no photoelectric emitter or phosphor will display a smooth spectrum.

But there are great differences between LED light sources. Those that are 2700 K, same as incandescent bulbs, mostly are good enough for photography, taking into account that their metameric response will be similar to incandescent light. As far as I know, we still don't have any daylight LED sources of similar quality, but they may be produced in a near future.

When using LED:s for photography, I think the best choice would be those that are intended to replace incandescent lamps. Setting white balance will correct colours sufficiently for most purposes, however not for critical colour comparison.

For daylight, there are very good fluorescent lamps around, with colour temperatures of around 5000 k to 6500 K, and CRI around 97, usually declared as above 90, for tube colours 950, 954, 955, 960 and 965, The first digit indicating 90+ and the two last digits colour temperature with the two zeros stripped. Those fluorescent tubes are usually not very expensive, about twice the price of ordinary tri-phosphor tubes with CRI of 80+. In my home, most lamps have compact FL tubes of colour 954. I have found that colour to blend well with daylight.

There is another caveat with alternating current and LED or compact fluorescents, that due to negligible smoothing of the rectified DC, bulbs emit modulated light, with twice the grid frequency, i.e. 100 Hz modulation for Europe and many other countries, and 120 Hz for the USA and most American countries. In photography, this means that the light is not suited for short exposure times. To avoid banding due to this modulation, exposure should preferably be several cycles of the modulation. So exposures of less than 1/30 second should be avoided with those compact light sources.

Availability of different light sources is not the same in the USA as in Europe or Asia. It may also be difficult to find sources of good quality in retail.

(FYI, before retirement, I was working as a technician at Waldmann Lighting for about ten years. Lighting has been my profession.)

[Mark von Kanel]

Hi Urban,

you obviously know what your talking about so how about a link for a suitable lamp for photography? i need some for some still life i want to try, What about dimming? are there any available or do we just buy different wattage?

thanks

[chauncey]

i need some for some still life i want to try, What about dimming? are there any available or do we just buy different wattage?

Because I'm groping around with these lights, I just buy the 100 watts that are dimmable...

that image is awful; the hard shadows and poor tone control make it hard to evaluate the colour

Yes it is but...I wanted something with lots of colors to evaluate a new LED bulb that was used with the Passbook for color correction.

BTW, what type of inexpensive bulbs do you guys use for studio set-ups?

[ajohnw]

Urban a little off topic but are 2700K power leds available?

On mains led lighting that just replace normal bulbs to be honest I have never come across any that low and as you point out daylight types are far from that. I do know that people like the UK's National Trust and English Heritage who maintain stately homes etc that are open to public visits seem to have settles on a certain type of compact fluorescent lamps as they want the public to see colours correctly but they might also be concerned about fading.

John
-

[chauncey]

Urban a little off topic but are 2700K power leds available

The color checker indicates these are 2700K as shown in the passbook image.
http://www.amazon.com/gp/product/B00...pf_rd_i=507846

[Inkanyezi]

I have noticed that the merchandise is different in different parts of the world, and here in Sweden I cannot get any other high quality LED bulbs than the 2700 K ones. They come as complete spotlights or as screw-in or E14 or E27 spots or replacement bulbs, as well as G10 socket bulbs for 230 V.

Cree makes several different power diodes, and they are retail marketed in Philips and Osram spots.

They are not dimmable, and most are Ra 80+.

There are also cheap spots of the bluish kind mentioned, which are not tri-phosphor, but blue emitters with yellow phosphor. That is the most common type in flashlights and in bicycle headlights. I abhor those lamps, which don't have a full spectrum.

It is very easy to test any LED flashlight for lack of red light, by directing it to any rear reflector of a vehicle. With the bad ones, it remains dark, but with a tri-phosphor lamp they light up bright red.

I haven't tried any dimmable ones, but I use a set of Northlight Loke for tabletop shooting.
http://www.clasohlson.com/uk/Loke-LE...ht/Pr182702000

[chauncey]

The CF bulbs that were used in that test were these http://www.amazon.com/Cree-18-Watt-2...LED+Light+Bulb
Both types are dimmable.