Haven’t gotten my hands on LM4 yet, after reading about the various bugs.
(I do have the LM3 for several months now.)
Question regarding the LM4, it seems like many have mentioned bogus R9 values for low-CRI lights.
Is it possible that because low-CRI lights tend to have a negative R9 value, and the LM4 software does not show negative result, it has a software bug that just shows “R9 = 100” (maybe software bug for negative R9)?
It is a possibility.
However, I have observed R9 values of 100 in the following cases as well:
Wurkkos TS25, 519A 5000K R9080, output level 6/7 and above
Sofirn SP10 Pro, LH351D 5000K high CRI, output level 2/7 and above
Convoy M21C-U, XHP70.2 5000K low CRI, turbo
Wurkkos TS30S, SBT90.2 5800K, output level 6/7 and above
(I adjusted the distance between the light and LM4 so the lux levels fall below 10k.)
Therefore, I believe that certain conditions trigger R9 values of 100, and how the LEDs are driven appears to play a role. I have noticed low readings of R9 in SBT90.2 when used in low output, but it reads 100 in higher output.
Thank You!
data pairs:
line 1: Xrite
line 2: Opple 4
3013K, Duv 0.0012, Ra 92.0, R9 58.7
2939K, Duv 0.0015 (+25%), Ra 93.7, R9 65.9 (+12%)
4110K , Duv 0.0008, Ra 92.8, R9 64.3
4003K, Duv 0.0013 (+63%), Ra 93.7, R9 67.1
3937K, Duv 0.0009, Ra 94.5, R9 69.7
3837K, Duv 0.0018 (+100%), Ra 95.4, R9 73.8
5904K, Duv 0.0074, Ra 69.8, R9 -35.7
5430K, Duv 0.0037 (-50%), Ra 69.3, R9 —
6322K, Duv 0.0035, Ra 66.2, R9 -39.5
5898K, Duv -0.0012 (-133%), Ra 65.2, R9 —
6018K, Duv 0.0072, Ra 69.3, R9 -38.0
5523K, Duv 0.0032 (-44%), Ra 68.3, R9 —
4966K, Duv 0.0035, Ra 91.2, R9 69.4
4743K, Duv 0.0036, Ra 91.8, R9 78.6 (+13%)
5365K, Duv 0.0003, Ra 92.7, R9 68.0
5075K, Duv 0.0012 (+300%), Ra 93.3, R9 65.9
4621K, Duv -0.0060, Ra 92.9, R9 72.6
4393K, Duv -0.0052, Ra 91.1, R9 69.3
4143K, Duv -0.0046, Ra 96.9, R9 84.3
4017K, Duv -0.0013 (+72%), Ra 96.4, R9 75.6 (-10%)
4907K, Duv 0.0059, Ra 89.6, R9 45.5
4587K, Duv 0.0053, Ra 90.8, R9 65.9 (+45%)
5603K, Duv 0.0144, Ra 67.9, R9 -52.6
5234K, Duv 0.0106 (-26%), Ra 68.9, R9 —
2810K, Duv -0.0023, Ra 93.5, R9 72.3
2826K, Duv -0.0020, Ra 96.9, R9 78.1 (+8%)
============
seems the Opple 4 DUV is unreliable
in some cases it reads high, in other cases it reads low
and seems the R9 does not work for Low CRI… on my Opple 4 it gives a wrong positive number… on McBob’s it gives no number just a dash —
On one sample the Opple 4 gives an R9 that is 45% higher than the Xrite…
I wonder if there was an update of the app that just clips R9 below a certain value?
not sure if it is an App difference or a Hardware difference
I use the Iphone App…
for LEDs that should have negative R9 it gives a positive R9 number, not a dash…
maybe the dash is from the Android App?..
in both cases it is clear that Opple 4 cannot show negative R9 values…
Don’t know how’s the accuracy of the used sensor and
what’s the reason of the devices behaviour.
Maybe no calibration, firmware bugs, the electric circuit,
the sensor itself (maybe a Chinese fake).
But your relative numbers can be misleading, too.
If we add the LM3, the conclusion might be similar.
I would look at the actual error rather than the percent error. A Duv measurement within 0.0005 or even 0.0010 is probably fine for most uses. CCT within maybe 100-150K. There’s not really any meaningful difference between Duv 0.0008 and 0.0013, for example. I think more formal testing would be required to determine the specific limitations of the Opple 4, but it certainly seems to struggle with low CRI Duv. It does seem to do a respectable job on all measurements with most high CRI lights.
This is some older data I have for the Opple 3. In general it doesn’t have the same issue with Duv as the Opple 4, but it often doesn’t accurately measure cool white CCTs.
| X-Rite i1Studio | Opple 3 |
|---|---|
| 3444K, Duv -0.0019, Ra 96.9 | 3481K, Duv -0.0022, Ra 95.8 |
| 4055K, Duv 0.0028, Ra 92.1 | 4017K, Duv 0.0020, Ra 92.7 |
| 4305K, Duv 0.0033, Ra 83.3 | 4205K, Duv 0.0033, Ra 81.7 |
| 3253K, Duv -0.0023, Ra 97.6 | 3520K, Duv -0.0032, Ra 96.0 |
| 3369K, Duv -0.0007, Ra 95.4 | 3395K, Duv -0.0002, Ra 95.3 |
| 3086K, Duv -0.0030, Ra 98.1 | 3138K, Duv -0.0029, Ra 96.7 |
| 3118K, Duv -0.0022, Ra 97.5 | 3144K, Duv -0.0018, Ra 96.3 |
| 3706K, Duv 0.0022, Ra 84.7 | 3817K, Duv 0.0015, Ra 83.9 |
| 3827K, Duv 0.0026, Ra 83.0 | 3771K, Duv 0.0028, Ra 82.0 |
| 5371K, Duv 0.0144, Ra 68.7 | 5247K, Duv 0.0121, Ra 65.9 |
| 5361K, Duv 0.0111, Ra 67.5 | 5354K, Duv 0.0110, Ra 64.7 |
| 6066K, Duv 0.0039, Ra 66.6 | 5708K, Duv 0.0058, Ra 68.0 |
| 5876K, Duv 0.0055, Ra 69.6 | 5571K, Duv 0.0060, Ra 69.7 |
| 5706K, Duv 0.0110, Ra 67.8 | 5448K, Duv 0.0100, Ra 65.5 |
| 6180K, Duv 0.0029, Ra 72.1 | 5849K, Duv 0.0040, Ra 71.9 |
I agree,
and thanks for the opple 3 data also
the Opple 4 in that test reads 0.0038 lower than the Xrite:
5603K, Duv 0.0144, Ra 67.9, R9 -52.6 5234K, Duv 0.0106, Ra 68.9, R9 —
the Opple 3 in this test, reads 0.0023 lower than the Xrite
5371K, Duv 0.0144, Ra 68.7 5247K, Duv 0.0121, Ra 65.9
from my own Opple 4 to Opple 3 comparisons… it seems the Opple 4 reads about 0.0030 Higher than Opple 3
at the end of the day, neither Opple is as accurate as an Xrite, but my overall impression is that the Opple 3 DUV is closer to the Xrite than the Opple 4 is to the Xrite…
otoh, I am encouraged by your R9 data… it seems the Opple 4 is in the ballpark
thanks for taking the time…
I wonder how the duv can be so good on the LM3 when the CCT is quite off. I mean, I knew the CCT is not very precise, but always thought that this would mean that x,y has a more random than systematic error. Very interesting!
thanks for the great charts!
my first impressions are that CCT and Ra are pretty close to Xrite, for both Opple 4 and Opple 3.
for DUV the Opple 3 seems closer to the Xrite, with the Opple 4 reading as much as 0.0030 lower
for R9 the Opple 4 reads as much as 20 points higher than the Xrite, and cannot display negative R9…
Note that McBob’s readings for lights w negative R9, his Opple 4 displays no number, but on my iPhone app, Opple 4 displays a relatively large positive number (above 65)… have not confirmed if this is just a difference in Android vs iPhone app, or if the actual Opple 4 he and I have, are set up differently at the factory…
Whoops, first chart used different colors than the others for the sensors, quick edit so now the X-Rite is gray on all charts
Thank you so much
Maybe only tweaking the app is how Opple intends to try to fix things. After all, the readings we see are simply the product of how the software interprets the output (through the firmware) from the sensors. That, and I have a feeling that they did not design the device to be firmware upgradeable.
I have a spectrometer I can test with.
Not sure when I can do it. I already did a quick test in my sphere (which combines all spectrums in the beam together, and therefore gives a more stable measurement) instead of directly into the beam.
So what you guys suggest?
I would like to see a test using one spectrometer, one Opple 3 and one Opple 4 comparing readings from warm, neutral and cool, all High CRI
for examples
I would suggest you test the following LEDs
we already learned the Opple 4 does not provide accurate R9 with Low CRI… so no need to retest low cri leds… imho
In reality we need a test across each sensor (and ideally multiple sensors each) that changes each variable and across each other.
Low output
Mid output
High output
Medium Low cri (say 70?)
Mid cri
High cri
Low duv
High duv
Low ra
High ra
Low r9
Mid r9
High r9
But each variation needs to be tested against other variations too. So we need a low cri light at low, mid, and high output, then a mid cri light also at low, mid, and high output, etc.
We don’t know if say the Opple 4 (or 3 for that matter) corrupts duv at different outputs, or if a low CRI corrupts the DUV measurement, or R9…etc
Sadly we also don’t know if the Opple itself is consistent across different devices. Conducting ALL the tests across two, better three, different Opple 3 and also O4s would let us see variation.
That data can then be statistically analyzed…
Feel like conducting 108 tests two or three times? LOL! Assuming you have 13 different lights covering each variation…
(My numbers are a little off, it’s not quite that bad. Forgive me I’m eating dinner)
Thinking about it, well one, any tests you do will be good.
Two, If folks have both a good high quality sensor and an Opple, or an Opple 3 and an Opple 4, we could do a Google doc and compare multiple results across multiple uses. Less precise statistically you enough data helps improve that
It obviously has bug in application. I have for example such measurement
CCT=3470 Ra=94 Duv=-0.0047
R1 = 100 R2 = 98 R3 = 94 R4 = 88 R5 = 100 R6 = 96 R7 = 34
R8 = 86 …
As you can check the mean of R1-R8 is 87 not 94 
I received my unit today but is not connecting to the app (on 2 phones).
Any ideas?
