Geeni Lux Edison G25
Test Date: August 8th, 2024
Firmware: 1.1.1
This page may contain affiliate links. Click here to read our affiliate policy.
Specifications
Model: GN-BW230-999
Shape: G25
Base: E26
Weight: 46g
Wattage: 7.4W
Protocols: Wi-Fi, Bluetooth
Matter Enabled: No
Color Mode: Tunable White
CCT Range: 2200-7000K
BBD Score: 0.0053
Brightness: 75-700 lumens
Avg CRI (Re): 91.5
Avg TM-30 Rf: 89
TM-30 Rg: 94-101
Dimming Algorithm
White Graphs
Thermal Image
Flicker Data
Risk Level: High Risk
Frequency: 120 Hz
Flicker Depth: 14.8%
Flicker Index: 0.01
Assist Mp: 0.2
SVM: 0.1
Pst LM: 0.26
Risk Level: No Risk
Frequency: 2857 Hz
Flicker Depth: 37.5%
Flicker Index: 0.1
Assist Mp: 0.01
SVM: 0
Pst LM: 0.02
White CCT Data
2200K
Measured CCT: 2220K
Duv: 0.0004
High: 610 lumens
Low: 65 lumens
CRI (Re): 92.5
TM-30 Rf: 92
TM-30 Rg: 97
2700K
Measured CCT: 2725K
Duv: -0.0070
High: 645 lumens
Low: 70 lumens
CRI (Re): 92.9
TM-30 Rf: 92
TM-30 Rg: 101
4000K
Measured CCT: 3995K
Duv: -0.0084
High: 700 lumens
Low: 75 lumens
CRI (Re): 91.1
TM-30 Rf: 89
TM-30 Rg: 98
5500K
Measured CCT: 5465K
Duv: -0.0031
High: 715 lumens
Low: 80 lumens
CRI (Re): 90.5
TM-30 Rf: 87
TM-30 Rg: 96
6500K
Measured CCT: 6490K
Duv: 0.0014
High: 735 lumens
Low: 85 lumens
CRI (Re): 90.3
TM-30 Rf: 85
TM-30 Rg: 94
Learn More
All of this data can be confusing at first, read through this section, and I promise it won’t be any longer.
You’ll be an expert!
Lumens tells you how bright a light source is.
Most lights state they emit “800 lumens” or “1600 lumens”, however, these are seldom accurate.
Our equipment can measure lumen output, so we’ve recorded the average lumen output for whites and RGB settings.
This means you can see just how bright or dim a light can get and how vibrant the RGB colors might be.
Correlated color temperature or CCT refers to the shade of white a light source is, essentially how warm or cool a light source appears.
When you heat a metal, it goes from orange to white, to blue. This color gradient is referred to as the black body locus.
A CCT of 2700K will be more orange, while 6500K will appear more blue.
One of the most attractive features of many smart LEDs is their ability to change from warm to cool light.
However, how accurately they achieve a selected CCT isn’t disclosed, and some lights are better at this than others.
When possible, each LED is set to a selected CCT such as 2200K, after which the actual emitted CCT is recorded.
A graph showing a light’s deviation from the specified CCT is shown as follows:
One thing worth noting here is that cooler CCTs are much harder to notice large changes in than warmer ones.
Meaning, the color appearance change from 2000K to 3000K is far more drastic than 6000K to 7000K.
Color temperature isn’t the whole picture though, there’s also tint or Duv.
Duv shows you how far a light’s color deviates from the black body locus:
A negative Duv indicates that the source is “below” the blackbody locus, having a purplish tint, while a positive Duv indicates that the source is “above” the blackbody locus, having a greenish tint.
Most people seem to prefer a slightly negative Duv over a positive one.
We’ve highlighted Duv which exceeds +/- 0.002 as this is about when tint usually becomes noticeable.
For white tunable LEDs, we’ve included a blackbody deviation chart so you can see how well a light stays within the 0.002 limit across its range.
We also include a blackbody deviation score or BBD.
This uses a single number to represent how far a light deviates from perfect white across it’s entire range. The closer to 0 it gets, the better.
The oldest and most widely adopted color rendering metric is the color rendering index or CRI.
This essentially compares a reference light (such as sunlight) to the light being tested. A score of 100 indicates that the test source reflects colors identically to the reference light.
CRI leaves a LOT to be desired and has since become obsolete with the recent development of the TM-30 standard, however, since everyone is still using it, we’ve decided to include it in our measurements.
You may have noticed however that we use CRI “Re”, which includes saturated colors such as R9 or deep red. This is a much harder score to rank high in. When companies state their “CRI” score, they’re referring to CRI Ra, which excludes these color. But still, just use TM-30 Rf, CRI sucks.
The new TM-30 standard is a far more modern and useful set of color quality metrics.
TM-30 was created in 2015 and was designed to replace the antiquated CRI standard.
TM-30 gives us tons of useful information for grading the quality of a light source:
Rf or Fidelity
This is a number from 1 to 100 that measures the reflected light accuracy compared to a standard, similar to CRI.
While CRI only measures 8 to 15 colors for its average score, TM-30 uses 99 total colors.
The fidelity graph looks like this, giving you an idea of how accurately it reflects certain colors.
Rg or Gamut
This number shows how saturated a light source is on average. A number higher than 100 indicates it’s slightly over-saturated, while a number under 100 indicates under-saturation.
Color Vector Graphic
Finally, we get the CVG, which gives us a visual representation of color saturation.
We’ve included a full TM-30 report for each CCT tested.
A thermal image is taken of each bulb after 1 hour at maximum brightness:
This is mostly just for fun…
Generally speaking, smart lights dim in one of two ways:
- Logarithmic
- Linear
Here is what logarithmic dimming looks like:
And here is what linear dimming looks like:
You may think that linear dimming makes the most sense, but humans perceive light changes logarithmically, so you may prefer this style of dimming over linear.
Most lights modulate their brightness over time. This is often referred to as “flicker” although this is a generic term for the idea.
There are several ways to measure temporal light modulation and artifacts.
We measure just about all of them using a LabFlicker meter and describe them in more depth below.
Included are waveform graphs and test reports at 100% and 50% brightness levels.
Temporal Light Modulation and Artifact Measurement Standards
Here are the metrics we measure:
1. Flicker Risk
According to the IEEE’s 1789 flicker standard, flicker can be grouped into three categories:
- No Risk
- Low Risk
- High Risk
This is an oversimplification of flicker but is useful for “at a glance” judgment of a light’s flickering.
2. Frequency
Well built drivers will maintain a 0 Hz flicker across it’s brightness range, while poor drivers will almost always flicker at around 120 Hz, equal to the 120V energy system used in the US.
Many lights flicker anywhere from 120-4000 Hz and can vary considerably within this range.
3. Flicker Depth
This tells us exactly how much the light intensity changes from maximum to minimum brightness.
A flicker depth of 90% means the light source is dimming down to 10% of it’s maximum brightness.
4. Flicker Index
Flicker Index is an older metric that tries to quantify the “flicker” by calculating the area over the average divided by the total between both:
Many people consider this metric useless compared to the other newer ones, but it’s generally considered that the lower the number the better.
5. ASSIST Mp
This is another metric used to quantify flicker severity perceived by individuals under different lighting conditions.
It takes into account factors such as flicker frequency and modulation depth to assess how noticeable and potentially uncomfortable flicker may be.
We like to see this number stay under 0.1
6. SVM (Stroboscopic Effect Visibility Measure)
SVM seeks to determine how noticeable the stroboscopic effect from a light source is.
A stroboscopic effect occurs when a fast-moving object appears stationary or moves slowly under a flickering light source that flashes at a frequency matching or close to the object’s motion frequency, creating an illusion of motion distortion or stillness.
- SVM < 1: Not Visible
- SVM = 1: Just Visible
- SVM > 1: Visible
7. Pst LM (short-term light modulation)
Visible flicker typically occurs when the frequency is from 0.3-80 Hz.
A PstLM of 1 means that the average observer has a 50% probability of detecting flicker, so the lower the better.