Chanon Wangtrirat

11 hours ago, Phil Rhodes said:

The thing with SSI, and all the TM-x stuff, is that it defines similarity to some ideal. Yes, that's more or less what all of them have to do inasmuch as they're targeting cameras which are targeting our eyes which are targeting sunlight. Still, I suspect that people will always want a quality number that isn't relative to a variable. People like CRI because high numbers are good, and interpreting an SSI requires a lot more interpretation, understanding, and frankly subject area expertise.

You get my point ^^

17 hours ago, Albion Hockney said:

All I'm saying is Quasar Bulbs for example have a high TLCI, but they always look a little magenta. Skypanels in tungsten match damn near perfect to tungsten lights on camera. I know you hate skypanels and I agree they are overpriced, they are too heavy, and the numbers don't look that great - but like all arri products they work. 

That’s what I encountered too. 

3 hours ago, Phil Rhodes said:

Skypanel is not that great.

It was presumably engineered to suit Alexa, given the corporate situation surrounding it.

The colour is only one issue; they're also amazingly heavy, a bit flimsy and incredibly expensive.

It introduced at the time that LED was not that mature, mediocre efficiency and less predictable color back then. I don’t surprise why Arri choose that route back then. It’s more predictable than most light there plus get function like changing color. So yeah it tuned toward alexa. 
it’s  overdue for arri to come up with something better than skypanel(form) fixture.

fact is Mike Wagner who do marketing for ARRI lighting moved to Litegear as CEO.

Also my DIY panel already copy form of litemat for my usage because it’s so lightweight.

BTW, i cant find SSI of skypanel. Could someone measured that?

13 hours ago, Albion Hockney said:

I think all of these numbers can only be so helpful. Every camera is different - Everyone talks about how bad skypanels are, but they look really great on the alexa. 

Yeah that’s exactly why I average this and called “average light quality index” as if you use solely SSI to evaluate it that will render skypanel crappy.

31 minutes ago, Phil Rhodes said:

Hmm, I could measure some of the stuff I have here.

But I think in the end we can keep coming up with different numbers forever. It is hard for something to have a good TLCI, but be really bad at R9 or R12.

P

I find it easy to differentiate such product apart from "97 TLCI" "97CRI" some of them get these numbers but end up in 80~ range ALQI

Surprisingly Aputure P300C got best here on daylight.

Hi! I'm Thomas. 

After I used CRI to measure such thing so long ago with those flos now LEDs. 

I found these measurement of SSI, TLCI, TM-30 too complicated to evaluate light when you just want to buy one, It's too much data use to judge how good light is.

True that SSI has been invented but we are too far from getting such technology like LED to get perfect score or even touch 85 in daylight without using UV chip instead of blue. Which mak SSI score look awkward especially when compare to conventional lightsource like HMI or tungsten. 

BTW SSI invented to use to match 2 lightsources not directly use to evaluated such light quality. It could be use but honestly look awkward to me.

So I came up with something more simple of single number defined quality of light 

Do you have any updates on this project?

The Original apollo is aging. There is some problem about connections between LED strips. Because it's diy project.

 

There are new panels I made have same formfactor like litegear litemat. Which use aluminum sheet as structure have really clean surface for strip and adhesive to stick on. Use 5 colors RGB+CCT in 1 chip. Produced clean white from 2800K to 10000K same as before but has more consistant in brightness.

 

Photo of my friend lit using Kino Flo KF55 on the right and new Apollo LT on the left.

All using same control as before on Blynk with now running local server on NANOPI duo in powerbank. Not scared of no internet connection anymore.

The latest test pave for second panel I will made. Experiment for decision of strip.

This is the most accurate combination I've done.

RGB + White 6000K CRI 95 + Warm White 2850K CRI95. The most expensive one also.

 

Achieve CRI Ra 96.7

 

Calibrated using Sekonic C-700 with MG+-0 @5500K

 

How are you powering it?

 

 

I used Meanwell 24V 400W Power Supply. Sightly under voltage to 22.7V lower the heat stress on panel.

Very well done! Is this just a personal project or are you looking to expand into a business?

 

Maybe but not for now as I have to test it myself first.

Why did you choose a Warm White led strip over a Cool White led?

 

The Phosphor on warm white has stronger yellow-orange-red distribution which easily compliment by add blue spectrum by Blue LED and add more red by Red LED

 

The Blue chip in white LED have so much power in itself opposed to red which is lower energy than blue. Red LEDs are also inefficient compared to blue.

Amazing work, man.

 

I did something similar to you, but i used a RGB, WW and CW led strip to achieve the disered color temperature.

 

How did you find the numbers for each color temperature? Trial and error with a spectrometer?

 

I planning in build a larger scale of my first led panel, but this time in a smarter way.

 

Thanks man!

 

 

Yes, Trial and error.

 

Firstly, I use i1Pro2 spectrometer to measure the CRI CQS above. but later I borrow Sekonic C-700 for better evaluation.

Aha, PWM. I assume the frequency is pretty high to ensure no flicker?

 

P

Yes, more than 20000Hz. There is sightly banding occur in >1/400 shutter speed but most of the time I shot just 1/50-1/100 so it performed flawless even dim down to 5%

Nice work.

 

How are you allowing the ESP to control the brightness?

I used ledc function on esp32 which can be easily coded.

 

Blynk virtual wire just send variable from slider on smartphone.

 

 

ledcSetup(3, 12000, 12); // 12 kHz PWM, 12-bit resolution

 

And then

 

BLYNK_WRITE(V0)

{

PWM_LED1 = param.asInt();

ledcWrite(1, PWM_LED1);

}

 

 

This is example from others but my code runs at 13 bit resolution (8192 level) to control brightness.

Hi! guys. I'm again on Lighting topic.

 

 

 

 

Some of you might have read my post http://www.cinematography.com/index.php?showtopic=66862&do=findComment&comment=427134

 

That one was really old and now decommission for upgrade and refurbish with some exciting new feature that I got.

 

I've made LED called : XineLED Apollo for evaluating LED to be use instead of Fluorescent in my arsenal.

 

This one made from home depot softpanel LED troffer which you can. find in various size.

 

 

 

Here is what I wonder to have

 

- High CRI over 90-95

- Color Tunable

- High brightness

- Battery Powered

- Large Aperture

- Smart Function

- No flicker dimming

 

This have to be replace all LEDs inside the unit to new one which denser and have RGBW Strip.

 

Nominal Power consumption : 280W

 

 

 

 

Last time I posted the topic. I said, It would have to use CRI 95 LED for White LED to have best color rendition. WRONG!!!!!

 

This is due to red primary which use in CRI 90-95 LEDs is really Red spectrum opposed to Orange-Red in normal YAG CRI 80 LEDs.

 

Here's measurement I take with i1Pro2 Spectrophotometer with CRI CQS and various standard results.

 

This is CRI 95 WarmWhite + CRI 86 Warmwhite (for Spectral continuity) + RGB

 

 

This is Normal Panasonic 3W Warmwhite LED CRI 85 + RGB. The result is really clear.

 

 

Main things need to go right with RGBW

 

- White LED need to have continuous spectrum but being cut off at some spectral for RGB to add later

- Blue LED in RGB LED have to fairly deep blue. Likely to cause Metamerism for both eyes and Camera sensor, Royal blue 460-470nm for optimum color rendition and light source match.

- Controlling software need to be totally HSL color manage solution (for professional equipment)

 

SKINTONE test with Standard Sony photo

Tungsten 3200K mode

 

 

I'm pleasedwith how it's render skintone.

 

 

CONTROLLING :

 

I use ESP32 for controlling the brightness in each channels, It's amazingly cheap but yet fully functional for me.

Control using Blynk platform on Mobile phone. As I have no skill in writing code on OLED/LCD menu interface.

DRAWBACK : Blynk application needs to connect to the internet. which make me have to turn on hotspot on phone when I use outside range of my home wifi.

 

Here's current functions I've already written the code.

 

White Mode (K)

- 3200

- 4100

- 5000

- 5500

- 6000

- 6500

- 7500

- 8500

- 10000

This is preset being calibrated with both Spectrometer and camera sensor. (Canon, Sony)

 

Effect lighting

- Fire

- Police

- Lightning

- Party fade

 

RGBW custom mode

 

All being control on smartphone.

 

 

 

It also have small OELD Screen for status message.

 

 

 

 

 

Many feature will be add later as I find out how to write the code ( i'm so noob on this.)

 

Video just test on Hue mode

 

 

 

 

UPDATE Coming !

Thanks for sharing this...it reminded me of this product, which I found very interesting: https://www.bhphotovideo.com/c/product/1237741-REG/luxli_orc_viola_5_viola_5_multi_color_on_camera.html

 

I wonder if you're able to get your solution to this level but with a larger panel and softer light source of course. Please keep us informed of your progress.

 

I know a bit about coding, I wonder if there is something I can help with.

 

 

Oh! Thank you. Really cute little light with such tons of functions. I afraid that modding it to drive larger panel don't really work the way it should because it have to be calibrated. But at least I get some idea for the interface. ESP32 inside for sure.

 

 

Usually this sort of thing is quite simple to scale. Chanon?

 

P

 

 

Yes. But not that easy.

Nicely done. Where's that constant-current strip come from?

 

The first resistor current limiting strip I used is here. https://www.aliexpress.com/item/Superbright-2835SMD-LED-22-24lm-led-2800-6500K-120leds-m-5m-per-reel-5mm-PCB-CRI/32673461159.html?spm=a2g0s.9042311.0.0.CX7hVw

 

The Constant current I bought from here. https://www.aliexpress.com/item/5m-reel-DC22-27V-112-leds-m-2835SMD-High-CRI-90Ra-Constant-Current-IP20-non-Waterproof/32547434392.html?spm=a2g0s.9042311.0.0.CX7hVw

 

I have already test the constant current one. 90' CRI LED. It's somewhat have little greenish tint to it. Which I think easier for me to correct with RGB LED, Green one have such poor power compare to blue or white LED.

 

And the RGB I found here. But I haven't bought it yet. It just very new in market. http://www.addressable-led.com/Products/Constant-Current-RGB-LED-Strip.html

 

There're some spec that I'm concern. The RGB one is the most serious one. The Apollo made out of RGB strip which have not deep enough blue LED. It's fall in light blue region. Made it after calibrate to the camera seem to be greenish on eyes. I will test one before purchase multi-meters for making another one.

 

I have plan to buy C-700 for evaluation soon. for Scientific measurement.

Hi! There I'm Thomas.

 

In the past 2 years I have made Dive like LED fixture posted here http://www.cinematography.com/index.php?showtopic=66862&do=findComment&comment=427134

 

 

It's made out of Off-shelf LED strip contained Horrible warm-white LED and RGB LED. The output is low because I have to use Red LED to correct there poor color rendition to such just usable but it's very versatile and fast to use because of RGBW type ability to change color. As it is made before any of Celeb or Skypanel debut at that time. I can see the potential of this kind of fixture features.

 

Now it's 2017. The High CRI LED comes surprisingly cheap. All LED products are freaking cheap.

 

This is XineLED Apollo I named it because it's pretty bright. Consumed 288w (opposed to Skypanel nominal 400w I won't going to fight). It brighter even 4x4 bank and most of Celeb and Select.

 

 

 

It's Light head + PSU 360W from Meanwell 24V

 

I use Arduino DUE to control the intensities of multiple channels, RGBW.

 

 

the Arduino DUE can control PWM modulation in 12bit resolution at 20KHz which plenty of room to prevent the flicker problem. It may cause some trouble with High-speed more than 300fps but hey! 288w won't suite for that job right?

 

12 bit pwm dimming means precise control of intensity up to 0-4095 level of brightness.

I write the code which set the white balance in 12bit of 4 channels and then rotary encoder will feed the dimming level and calculate level output each channel should lit. So it's more precise than 8bit 0-255. But this is for internal calculation only.

 

The Code is not fully complete I just barely use just 5500K preset as I can't write the menu system for LED/OLED yet. I will hire someone to program it.

 

The Panel consists of 30m of LED Strip ( 15m for Warm-white and 15m for RGB).

 

The housing made of plastic and aluminum bezel with Very good quality diffuser It's all from Troffer LED panel which I bough from home depot here in Thailand.

 

one with LED at the back of it not the edge. Then I removed all crap original LEDs.

 

 

 

 

I planned to go for Constant current version of strip soon because normal one use resistor to drive the chip, It's inefficient.

 

 

Here some pics from testing it on bench.

 

 

Housing finished the side with durable plastic document folder and big cable tiles. It's pretty durable. Flexible conduit I used large copper cable instead.

 

 

Some scene put to use with 3200K preset (just write analogwrite code on arduino).

 

 

 

Aquablue with few lines of code make straightout sci-fi look.

 

 

 

Here is short video test the Magic lantern RAW at 3K featured XineLED Apollo in the room with Aqua blue. It's ability tune the color because it's RGBW.

 

 

Thanks for all comments. I would love to see your appreciation.

Looks pretty good to me.

 

 

I have experiment with Arduino Mega which can drive at 32KHz which produced no flicker at all within 1/2000.

 

It's now being develop with Digital circuit.

 

It may get more efficient than old circuit which used current limiting transistor.

Last time I posted Just made my self LED fixture. Tuneable, Color lighting. http://www.cinematography.com/index.php?showtopic=65114

 

The photo in post broken link.

 

 

 

 

I just have opportunity to really testing it in real set and with professional camera like RED One ( really the first RED One sir!!)

 

The power get cut on that day and I can't use any Kinos. So interior is very tight space and mostly closed. Reflector was use to bounce the light from outside through the door right side and left of female actor.

 

I put LED on C-stand and get it above actor to simulate windows light from above which tend to get warmer because it mostly will get yellow tint. So I adjust it to 5300K preset. Power? it was tap from RED v mount battery.

 

It's RGB ribbon + WW LEDs ribbon mixed together and calibrated on histogram, vector scope and some eye on Magic Lantern.

I installed micro prism sheet which similar to Litepanel Nanopticlens used on Helios line directing the light to 60 degree.

 

 

 

Here are samples screenshot

 

 

 

Here LED was lit from above the man.

 

 

 

There are a lot to be done with it. It's consume too much power supposed that because it's RGBW and not efficient LED chip. But hence it's RGBW it have much flexibility as it design to use.

This is really interesting. The idea to program Arduino to control DMX 512 on this module.

I've been looking for a way to control those cheap buck converters from a microprocessor for a while. I don't think most digipots will do it with sufficient resolution - most of them are only 7-bit, although you could bridge them out with fixed resistances to permit trimming in a smaller range.

 

Digitally programmable LED drivers are usually PWM, which, as you correctly say, is not the right way to go. It's possible to build digitally-controlled switching power supplies, but it's not something I've ever felt like getting into.

 

P

 

Exactly a problem I am facing too.

 

I wonder how those LEDs in the market that have digitised control system doing with controlling the power. Find it at cheapest method is the hardest thing.

The 2014 Nobel Prize in physics was awarded to Isamu Akasaki, Hiroshi Amano and Shuji Nakamura today, in recognition of their work in the development of the blue LED.

 

P

 

Yes, Blue LED is the game changer. We can't walk to this point if there is no Blue LED!!!!

 

Another technology that I think it's innovative is the UV violet LEDs. Manufacturer called Yuji Beiling international make this. Instead of Blue chip excited phosphor to make white light. They used Violet UV LED to excited Red-Green-Blue phosphor resulting in more controlled color spectral output.

 

 

.

 

http://www.yujiintl.com/high-tlci-led

 

I have called them and guess what???? one of the customer of them is Kino-Flo. Using there LED to make Celeb LED light.

 

As seen in Kino-Flo Celeb video

 

 

For now I can't effort to buy those LEDs. It's too expensive to make just this one but I think if it made up. With CRI and TLCI up to 97. Ridiculously excellent color rendition.

 

They make conventional 5mm LED used on Litepanel 1X1 but with Blue LED CRI up to 90 too. Also Ribbon 5050 LEDs with High CRI / TLCI Violet Chip LEDs.