Silver Fire

[mechman]

A few weeks back I started working on a mix called Silver Fire. Silver Fire is a mixture of numerous mica based paints, polyurethane, and a white latex paint. This is my second time mixing up the formula - the first time was one of the original mixes and it had a bad blue/red push. This new mix uses Liquitex artist acrylics. The second time around produced different results from the first via the spectrophotometer. According to the color temp calculated by CalMAN, silver fire comes in 100 points away from D65. But as we've learned over time, this only paints part of the picture for some mixes. The more complicated the application, the more data needed to come to a conclusion. One needs only look at the dnp Supernova review to see what an application that does not have a good spectrophotometer reading can do with a projected image. The Supernova is ISF certified and rightly so. But we're talking about Silver fire. During the process I sampled the base components, the color components, Silver Fire lite (less color component added), and Silver Fire.

Here is the color component



Here is the base



Here is Silver Fire lite



Here is Silver Fire



Here is a reference spectro reading from an X-Rite N5 gray



To go along with this I also took calibration readings from my Silver Fire panel. These readings consist of 0-100IRE grayscale images in increments of 10. It also consists of gamut readings which are 75% white, red, green, blue, cyan, magenta, yellow, and 100% white. The software used for both these readings and the spectro readings above is CalMAN Enthusiast. My spectro is an X-Rite i1pro which has been factory calibrated each fall since I've owned it. For the calibration readings, I use a Colormunki Create colorimeter. Supposedly this colorimeter is much better on the lower end than my i1. But I haven't seen enough to be convinced. I do know that it is much faster at getting the readings.

So without further ado, here is the calibration reading from the silver fire panel



For comparison purposes, here is the reference reading directly from the projector (BenQ w5000)



The things that stick out are that it appears to have a red push to it and the CIE chart shows all of the colors pulling towards red. We also went from a spectro reading that was 100 points higher than D65 to an average color point reading of 70 points less.

[mechman]

Empirical data used will be pictures. Camera is an Olympus Evolt 500 mounted on a tripod at 12 feet back. All images are color balanced using an X-Rite ColorChecker Grey Card. The process used to do this is to get a reference shot using the intended light source (projector) and then processing all of the images from within Adobe Bridge. Having a known source of gray in the image determines the color of the rest of the image. White and black levels were adjusted to the silver fire panel. Camera was on a tripod ~12 feet back at roughly head level while sitting.

While I was working on SF, I was also working on several other things - a 2:1 Black Widow, a C&S panel, and a new project of Bill's. The comparison photos will have BW2:1 in some and C&S in some. I also included some with the Vutec Silverstar in them so that you could see how SF compares to a high gain commercial solution.

BW2:1 is on the bottom SF on top.



That first image was the color balance image. And you can already see the, what I like to call, 'noise' being generated by too many reflective particles in a mix. This is the case with both panels but not with the background BW screen.

In this next image you can clearly see that SF is brighter on the bottom of the panel. If you look near the top though, you can see that it starts to darken a bit near the top. I'd like to reiterate at this point that I adjusted the white and black levels for this comparison, as well as the others, for the Silver Fire panel. This is an on axis shot.



Same spot with some ambient lighting



Ok so the SF panel has the brightest area on it. Yet it is not uniform. It also does not have the ambient capabilities of the background BW screen, as the blacks are clearly much better on the background BW screen. So now lets shuffle the panels a bit to either side. To the right with SF still on top and BW2:1 on bottom.



And with ambient light



From the last two photos it shows that SF quickly loses its edge as it moves to the perimeter areas. Whites are equal to the background BW screen with blacks going to BW.

Now to the left





One final thing I did with these two panels was to get some shots at roughly 18 degrees off axis. This is in line with where my wife sits. Previously I had estimated it at 20 degrees but I've since acquired a laser (it's a middle of the road construction laser) that will give me these degrees and is easy enough to line up with my screen. So I measured the seat at 18 degrees. Going from right to left with ambient light.







Pretty much identical to the right side. At this point in the photo shoot I was frustrated with the BW2:1 panel (yes - in my book it's a failure) so I changed it out with the C&S panel. C&S was made using the VUPE base. During these shots I took some with each on the bottom and some with each on the top. I figured C&S would be a good way to gauge SF's whites in the middle portion of the screen.

In this set C&S is on the bottom and SF is on the top. Ambient lighting...







These next shots have SF on the bottom with C&S on top. Some are dark room shots and some have ambient light.













Clearly SF has better blacks in the ambient lighting. But the black level crown would have to got to BW in the background. And the whites are good on SF on axis. Yet they fade quickly as you move off axis. C&S wins the white crown.

Some quick shots of a Vutec Silverstar alongside SF. The Silverstar is on top.







SF isn't even in the same building as the Silverstar when it comes to whites. It does appear that the Silver star hot spots, similar to SF. But brightness and contrast are set for Silver Fire.

[mechman]

I have a Sekonic L758C Digital Master light meter which I use for both spot meter readings as well as fL and lux readings. I figured I would take readings of the middle of the SF panel from the middle, left and right sides. What use is this information? Well it will tell us if the panel is in fact brighter in the middle than the sides.

These readings were taken using a 50IRE gray scale image and are measured in foot lamberts.

Top Left - 1.3
Top Middle - 1.7
Top Right - 1.3
Middle Left - 1.4
Middle Middle - 2.3
Middle Right - 1.5
Bottom Left - 1.3
Bottom Middle - 1.6
Bottom Right - 1.4

For comparison purposes, Black Widow yields these results.

Top Left - 1.8
Top Middle - 1.9
Top Right - 1.9
Middle Left - 1.9
Middle Middle - 2.1
Middle Right - 2.0
Bottom Left - 1.8
Bottom Middle - 2.0
Bottom Right - 2.0


Gain readings

Not done yet, but my preliminary on axis reading was .95833. May be a week or two on these.

[mechman]

In a nutshell, Silver Fire hot spots. And it's color reproduction is a bit confusing as the spectro reading shows 100 points higher than D65 and yet the calibration temp shows 70 points lower. I've seen funky things happen with regards to these two measures before - like I said earlier, the Supernova. But these were with layered technology which is way more advanced than Silver Fire could ever dream of. Unless one wanted to examine the aspects of Silver Fire under the microscope, I think it's safe to say that the shifting is due to the massive quantities of mica mixed within a large amount of polyurethane in an attempt to mimic 'pop' yet only results in fizzle. It's like the firework that goes off in the launcher, cool for a second or two, but quickly disappointing. I'm a bit unsure where 'pop' became hot spot with added graininess from the paint itself.

Silver Fire has also been touted as an ambient light screen paint. I tested the darkest mix and Silver Fire comes no where near Black Widow with regard to black levels in ambient lighting. And as for the brightness of Silver Fire, sure it's bright in the hot spot. Beyond that, it's on par with Black Widow with regard to levels.

One final thing to mention is the amount of tedium to this mix. Parts of it are measured in ounces, parts in milliliters requiring different measuring tools. And the milliliter measures are of the thick artist acrylics - of which you use less than a tenth of the bottle for both blue and yellow. It's a pain to mix and when you're done you'll have enough to do several screens as the total ounce count lands at 104 ounces. Which would be enough to do several screens.

This is my second Silver Fire mix. It ended up better than the first. But it's still not a usable product.

I'll end this review in noting that I have been in contact, again, with one of the originators of this mix and have, again, requested a sample directly from them. I am still waiting....

[mechman]

Spotmeter readings posted.

[dradius]

Wow. Great review, thank you for all the work.

[mechman]

I've been asked recently if I had ever compared Black Widow to Silver Fire. The answer to that would be yes. In all of the pictures above, the background screen is the Black Widow. Some of the conclusive things that can be determined from the screen shots via comparison is that Black Widow has much better blacks and off axis the whites are fairly equal. Why are the whites equal off axis? Look closely at the checkerboard pattern shots. Silver Fire's whites vary from one part of the white box to another part. What is the cause behind this? It's more than likely either the polyurethane or the massive amount of refractive mica used to create something which it cannot be - whiter and blacker at the same time. In the hot spot honey hole of Silver Fire, the whites are better than Black Widow's whites. Yet across the board, the blacks are never darker.

I'm in the process of painting up several new panels, a lot actually. These include polyurethane + paint mixtures, satins, eggshells, semi gloss and several others. You can expect more comparisons to Silver Fire in the future when all of this is complete. I'm thinking it will be around the Olympic break for the NHL when I finally get a chance to accomplish all of this. Keep in mind that I do have a life outside of all of this and my family is in high demand lately with practice's, girl scouts, etc.

Personally, I've kind of come full circle in all of this and lean towards what Smokey Joe taught me. Grayscale readings and spectro readings reveal most all of a product's ills. Pictures are just that... pictures. You see some of the pictures passed off elsewhere as 'conclusive data'.

[wbassett]

mech I know the list is already huge, there are a couple more that really need to be added.

One test is SF where the 'colorant' is a color match of the complex color mix that the application calls for. I am saying nobody will be able to tell the difference between the difficult and messy recommended mix, and one that is color matched. Neither will be neutral, I understand that but the point is that a person wouldn't be able to tell the difference.

Then the next logical test would be to replace the colorant with an actual neutral gray and then test it against the original SF and see which one performs better.

As far as tests about grays made with RGB vs other pigments or just black and white... We've already done that by testing all these other mixes and we can see that the spectral curve looks messy, the balance is bad, and the temp is off. When it comes to neutral grays, we've tested them to death too and I still maintain that if it meets neutral specifications then it doesn't matter how it was made. RIT agrees with that and everyone in the industry does too so I tend to think they know what they are talking about. Even at that though, we've seen it ourselves that the end result is all that matters.

[mechman]

wbassett wrote:

mech I know the list is already huge, there are a couple more that really need to be added.

One test is SF where the 'colorant' is a color match of the complex color mix that the application calls for. I am saying nobody will be able to tell the difference between the difficult and messy recommended mix, and one that is color matched. Neither will be neutral, I understand that but the point is that a person wouldn't be able to tell the difference.

Then the next logical test would be to replace the colorant with an actual neutral gray and then test it against the original SF and see which one performs better.

As far as tests about grays made with RGB vs other pigments or just black and white... We've already done that by testing all these other mixes and we can see that the spectral curve looks messy, the balance is bad, and the temp is off. When it comes to neutral grays, we've tested them to death too and I still maintain that if it meets neutral specifications then it doesn't matter how it was made. RIT agrees with that and everyone in the industry does too so I tend to think they know what they are talking about. Even at that though, we've seen it ourselves that the end result is all that matters.

I always planned on someday getting the "color component" from HD or Lowes in one of their sample cans. I will try and get it in with the rest in a few weeks. I don't think a neutral gray would work. It might, but I'm thinking all the additional mica paints will shift the results.

The problem with SF isn't only it's color component, it's the massive amounts of mica and polyurethane. It hot spots and has a pronounced grainy image. I think that is the 'pop' that folks who have seen nothing but SF refer to. It's what? 75oz of mica and polyurethane to something like 13oz of actual paint.

And not too mention the negatives of having a screen this dark with slightly above 1 gain. But if folks know how to interpret the data, it's all up there.

[jinjuku]

I wanted to post to the 'tedium' of mixing the colorant/base/ viscosity.

It all 'depends' on what you view as tedious. I did the SF Lite (I also have the components for BW and plan on shooting a screen with it this weekend).

All the measurements can be taken as Fluid Ounces. That is 29.57 (round to 30) ML's per fluid ounce. for $14 you can get a set of syringes. So there is no real conversion worry. The official SF thread should be more transparent about that though. As a side note: Blame the Imperial/US f'd up measurement system. If we were metric (Liter and Gram) this type of confusion wouldn't exist. Gawd I hate the U.S system of measures.

Is SF more 'work' than some other mixes: Yes. Is it a lot of work or tedious in the grand scheme of what it takes to get a projector and screen going? No.

As to hot spotting, I can only post from my experience: with a 100" diagonal 16:9 screen (will crop to 96" most likely with border) I have no hot spotting that I can tell.