Projector light output measurement

jfinnie

Distinguished Member
A recent review on this site by @Phil Hinton of the JVC DLA-N5 projector raised the odd(1) eyebrow(2) due to the seemingly very high measured nits. Some(1) discussion(2) ensued(3), and at request of some(1) folk(2) in the review thread I've taken the discussion of these numbers and put them in a dedicated thread, where perhaps we can discuss these numbers, the technique that lead to them, suggestions of measurements and presentation styles interested persons may find useful, and other matters.

The numbers in question:
Phil Hinton said:
We measured 460nits in the best D65 capable HDR mode (Frame Adapt) and contrast measured in at 47,734:1 (460/0.01) with no zoom and 47,874:1 (304.9/0.006) at full zoom with the manual iris wide open. Results like these will vary from projector to projector as they are bulb based and each age differently, but the contrast results are very close to those advertised by JVC.
For folk who regularly measure projectors in situ, the headline 460 nits is particularly surprising... On typically sized home theatre screens you'd more usually expect to see real world measurements in the order of perhaps 75-200 nits for high lamp mode depending on screen gain, screen size and throw for a JVC lamp projector with an appropriate (for the projector) sized screen. This nit measurement, being candelas per metre squared as the full SI units definition, would be taken facing the screen surface. However this isn't how the the review measurements are done it seems
Measurements are taken from the lens at 8ft to remove screen surface and room effects, we are measuring performance direct from the projector (the Klein K10-A is fitted with a filter for this task). All our projector reviews in this room follow the same controls and should be comparable.
My understanding is that's problematic it would seem for a couple of reasons.

Firstly, the Klein doesn't report Nits when it is used facing the lens (inspite of what calibration software may say). With the diffuser in place it would appear the correct usage would be to flip the meter profile in the Klein to Lux mode which should give approximate corrected Lux readings (Klein don't calibrate the meter and diffusers together, unlike high-end spectros like Jeti, so there is some error margin in using it for irradiance as it uses an approximate reference diffuser for the diffuser losses and characteristics). You would then take that Lux reading and perform a calculation (lux/pi=nits for a unity gain display) and then present that if you wanted as nits at that given screen size, assuming unity gain. (Klein FAQ on diffuser usage here). A better tool than the Klein to use would be either a calibrated light meter or a spectroradiometer with a calibrated diffuser and irradiance setting.

Alternatively, if you work out the projected screen area in metres squared at the given throw and zoom, and multiply that by the measured lux you will have an approximate measure of the lumens of the projector. This is approximate because the correct method for ANSI lumens measures the irradiance at various positions around the screen in a grid, to account for the fall off in light at the screen edges. In many cases though it is a good enough starting point.

Secondly, the 8ft distance used seems anomalous if this is fixed. As @ask4me2 pointed out, 8ft throw on these projectors produces almost comically small screen sizes (38" to 77" diagonal). Even once lux vs nits is taken care of, not many folk will be interested to know how bright a 38" image the projector can manage (even 77" would be right at the edge where hardly anyone bothers with projection these days).

If 8ft meter position is regularly used as a house standard, it would penalise projectors with shorter throw lenses significantly, as at 8ft their measured irradiance will be much lower due to the lens setup. Measurements for similar projectors would be broadly comparable at a relative (not absolute) level as JVC, Sony and Epson have similar (but not identical) throw setups.

What would be the best way to present information in reviews?


From my own point of view, it seems the best way of presenting how bright a projector is would be to present the number of calibrated lumens achieved at the two ends of the throw and across the different lamp modes, and perhaps give an indication of the effect of filter losses if available. It is very easy to work out how many nits you will get on a screen of a given size once you have the number of lumens that the projector is actually capable of putting out; there are lots of calculators on the web that can do it for you.

IMHO the "gold standard" of this kind of measurement was probably achieved by Ekki from Cine4Home here where the full range of effects on light output of iris, lamp, filter and throw were catalogued.
1586025319902.png

I'm not suggesting AVF go down that route fully as it is a lot of information, but the nominal 4 corners mentioned above would be a good start. A lot of folk don't want to or can't use high lamp, for instance, so the useable calibrated low lamp output is very useful to know, and often folk can't control their mounting position relative to the screen easily so an idea of the effect of throw on the output is also very useful.

Additionally, in the interest of user friendliness it might be nice to provide some ready standard ready reckoner to end users; eg: "On a 100" unity gain screen at 6m throw with high lamp, no filter and iris fully open and you might expect to achieve around 150 nits, and at 3m throw that would increase to 170 nits". (numbers are fictional, don't put them into a calculator!)

Key there would be to settle on a standard theoretical screen size for all reviews and perform the measurements and calculations to allow that screen size to be offered as a way of comparing really world achievable nits between projectors with different lens setups, which the measurements being made at the moment would seem not to allow.

Look forward to thoughts from my learned fellow forum members on the subject.

I will try and use the next post to provide some useful links to materials around the subject for those interested.
 

Attachments

  • 1586025300425.png
    1586025300425.png
    31.7 KB · Views: 29
Last edited:

jfinnie

Distinguished Member
Some useful links and info:

Notes by Dr. Optoglass: The Units of Photometry

Useful formulas:
Lumens = screen area at given throw in m2 x measured illuminance in lux at throw
radiance Nits (cd/m2) for unity gain screen = illuminance in lux measured / pi (3.1415)
Lumens = screen area in m2 x nits x pi (3.1415) (for unity gain screen)

You can get screen areas in m2 from here if your geometry has seen better days:
Screen Size Calculator for various aspects
 
Last edited:

jfinnie

Distinguished Member
Measurement technique for ANSI lumens (belt and braces / gold standard of light output measurement).
Pinched from here ANSI Lumen as you'd have to buy the official standard from ANSI:

ANSI is an acronym for American National Standards Institute. Inside of it, a special committee has developed a set of standards to measure the performance of picture display. These standards are contained in the ANSI IT7.215 document, which deals with various aspects of image performance. We’ll focus mainly on measuring the emitted light, with a simplified explanation of how the emitted light is measured according to ANSI.

The projector must be optically and electrically arranged for optimum performance on the entire screen area.The projector must be stabilized for at least 15 minutes before measurement at room temperature (25 ° C), with no other light source. Any light that falls on the screen as a result of reflections and refractions should be less than 1% of the main incident light.

ansi.jpg


Figure 1 Figure 2

The test pattern shown in Figure 1 was developed by ANSI to set a projector and prepare it for ANSI measurements. It consists of a completely white field with 6 blocks. Each block represents 5% of the total area. The brightness of the projector is adjusted so that the 5% block is clearly distinguishable between the 0% block and the 10% block. The contrast of the projector is increased until the 95% block is still discernible between the 90% and 100% blocks. The brightness and contrast are adjusted to obtain these conditions.

Once the projector is adjusted and stable, a complete white field similar to Figure 1 (but without the 6 blocks) is used for actual measurements of light. The light emission is measured in 9 points on the screen. The positions of these 9 points are represented in Figure 2 by the circles at the centre of each block. The measurements are then averaged and the screen size can be decomposed to give a reading of Lumens.
 

Stridsvognen

Well-known Member
If you look at the contrast measurements its odd that there has not been ringing a bell, no JVC have near the same contrast in min and max zoom, and if you calculate the nr for black and white he listed they also give verry different results than mentioned in the review.
 
Last edited:

jfinnie

Distinguished Member
I must say I didn't understand that either. Ignoring the questions of technique, the more you look at those numbers the more they start to just look like a very bad day at the office; especially on the premise that they're supposed to be comparable to previous reviews yet show almost twice the light output of the NX9 in its review. Though those measurements it appears were done at 13ft from a screen surface it seems, and not 8ft from lens. I'm not sure the consistency claim really holds here.
 

Stridsvognen

Well-known Member
Guess we are still waiting?
 

bandyka

Member
Having measurements in nits is quite important for some applications, such as madvr tone mapping that we are collectively testing and developing on AVS, knowing the precise actual light output of my JVC X7900 was key to achieve the best results and we are at a point where I am astound by the image quality.
The key is, exact measurements are very important for calibration purposes of any kind and for installation, screen selection, seating placement, contrast requirements, iris position determination.

And I am talking actual on-screen performance not off the lens.

My 7900 measures around 130nits on a 120" screen.
 

Stridsvognen

Well-known Member
Having measurements in nits is quite important for some applications, such as madvr tone mapping that we are collectively testing and developing on AVS, knowing the precise actual light output of my JVC X7900 was key to achieve the best results and we are at a point where I am astound by the image quality.
The key is, exact measurements are very important for calibration purposes of any kind and for installation, screen selection, seating placement, contrast requirements, iris position determination.

And I am talking actual on-screen performance not off the lens.

My 7900 measures around 130nits on a 120" screen.
What do you use to measure you on screen light levels with?
 

The latest video from AVForums

Podcast: Sony A90J Thoughts, madVR Processor & Sonus faber Lumina II Reviews + Movie & TV roundup
Subscribe to our YouTube channel
Support AVForums with Patreon

Top Bottom