In Germany absolutely, there are sometimes events by users in the Taschenlampen-Forum (TLF). I was at several meet-ups of flashlight enthusiasts (for example in Cologne). 
But here in BLF more complicated if everyone lives in another country
Looks like I should register in TLF someday… Moved to Italy a couple of months ago, but it’s not that far.
If you want to come by to a meet-up in Cologne, would be cool I think
One of the OppleMod chat members (Victor) compared X-Rite ColorMunki Photo measurements with different versions of the Opple apps.
Hi-Res mode was used for ColorMunki, so it measured 380-730nm spectrum with 3.3nm step (n = 106).
Ra is absent for Measure 1 since he started to record it only from Lamp 1 - Measure 2.
LM Smart 3.2.2 should actually say 3.3.2.
This chart proves that there’s no difference between Smart Lighting 3.3.2, 3.5.1 and 3.7.1 (modded or not).
And also you can see that Home app 3.5.1 produces worse results than Home 3.3.1.
If we remember complaints about how Home 3.5.1 behaves with low-CRI lights (see here and here) and the fact that Smart app can’t show negative R9 values (see here and here), it becomes clear that OPPLE Home 3.3.1 is indeed the recommended app for now. So I’ll focus on it for future mods and abandon Smart Lighting apps completely.
5 Thanks
Seeing how you’re on android it might not help me much.
On iOS there is no “Home 3.3.1” as it jumps straight from 3.2.1 to 3.5.1 for “Home” and 3.2.2 to 3.4.3 for “Smart”, which would you suggest?
To summarize, I got the following R9 measurements from T10 6000K LED.
SMART version=3.7.1, R9=83 (this measurement was similar to earlier version of HOME before 3.3.1)
HOME version=3.3.1, R9=60
HOME version=3.5.1, R9=40
According to @koef3, who I believe has proper measurement equipment, the R9 of my LED in TS10 should be around 67. See here for the full review. This measurement was published a few weeks after my post above, so I was not sure which version was the most accurate at the time of writing, although I had a feeling that 3.3.1 was the most accurate from my skin test.
In addition to accuracy, I found version 3.3.1 to be the most useful because it could display Ra and R9 values for low CRI, high CCT LEDs.
I appreciate Steve’s efforts to improve the algorithm, but I believe the recent alleged improvement came at the expense of accuracy and usefulness when measuring some of my LEDs, especially those with high CCT.
By the way, is there a way to install Home version 3.3.1?
wait for spectral reconstruction version,
so called algorithm improvements are actually still old algorithm central, no fundamental improvement.
actually old algorithm has inherent problem, I explained this before. its result is not explainable and behave kinda randomly. what I care about now actually are variables like CCT/duv/cx,cy, are those variables reliable?
CRI accuracy must be improved through spectral reconstruction.
3 Thanks
my Windows app GUI as above, currently only serial port connection allowed.
working on bluetooth support:-)
very preliminary, want to copycat ams osram’s color calculator but not easy 
5 Thanks
if color calculator someday is open sourced, it would be great news.
we can just feed the reconstructed spectrum into color calculator and all metrics are calculated. It would be like an IS system, kinda similar to Everfine’s IS system but in a extremely small form factor(of course due to sparse resolution/channels, accuracy would not be comparable to a real IS ).
Please note that Steve’s latest developments have not yet been published. We’ll have to wait until the next update rolls out so we can try his spectral reconstruction implementation. You can find the Android APK file of version 3.3.1 in @Cod3r’s OppleMod chat.
@stevechang from the chart above (made by Victor) CCT measurements look quite accurate. We would need more testing to confirm but it’s promising. This should indicate that CIE xy are accurate too. Sadly, I don’t have a reference meter so I think I can’t help on this subject.
Btw I got my LM4 last week 
I’ll try to connect using a serial adapter as soon as I have some free time, then figure out how to implement your library.
2 Thanks
I’d say go with Home 3.2.1 since it looks like only version 3.5.1 produces some strange results.
Argyll CMS seems to be open source, but I’m not sure if it would be of any help to you: Download Argyll CMS Source
Though maybe it has some useful formulas?
(suggested by Victor who made the ColorMunki chart above)
2 Thanks
although I haven’t verified CCT/CIE xy because it was provided by another colleague, but basically I think it’s OK. those metrics are derived by XYZ calibration, it’s way easy to do than spectral calibration. so I think there should be no big issue with those metrics.
basically LM4’s metrics are divided into 3 groups:
1.XYZ calibration derived/calculated metrics, for example CIE xy/CCT etc.
2.CRI/cs/eml etc, those were originally derived using ai method, but I found inherent issues inheriting ai method directly from LM3. so I’d propose spectral calibration method.
3.flicker related metrics, those are basically kept same as LM3’s.
6 Thanks
I finally got a LM4.
After quick testing with modded application Opple home 3.3.1 it seems that the results doesn’t look too bad. There are deviations for sure, but overall (and only for LED lighting!) it could be used for a very rough estimate for light quality. For other lights the results are messy (sunset with 12000 K with only 77 CRI and duv over 0.007, wait what), but this device is not made for that.
Unfortunately I don’t have the coating for my sphere ready but I think this is not a problem, since I can compare the results of my spectrometer directly with the LM4.
7 Thanks
this is the measurement of sunlight, all those metrics are calculated using the reconstructed spectrum.
most metrics are good enough(XYZ calibration should be used in conjunction with spectral reconstruction, XYZ for metrics like CIE sx/sy and cct and so on, while spectral calibration for color rendering index).
the spectral graph drawing is kinda distorted now, I guess the spectral curve is too smooth because the sensor can not detect the minor details in the real sun spectrum(I have not compare with cl-500a measured spectral curve yet, but I guess there could be some zigzags in the real sunlight curve ). below is the sun spectra curves I measured long ago(totally different weather condition compared with above graph).
1 Thank
for sunlight, currently LM4’s CRI is totally not reliable. as for duv, you mean it is also not accurate? I don’t know how my XYZ matrix would behave under 12000K sunlight.
I look forward to your comparisons…
Im interested in seeing how the Opple 4 w stock firmware, compares to your spectrometer, when testing a 4500K 519a…
compare column H and column L, we can see Ra accuracy is boosted.
column H used old algorithm, while column L uses new algorithm.
5 Thanks
This looks really good, both CCT and RA are hugely better now. Congrats! Amazing work!
Any idea when we will see the new algo in our Android apps?
2 Thanks
actually I don’t know, depends on Shanghai app team now.
2 Thanks
Do you know which app will get your new algorithm? Home or Smart Lighting? Or both?
Glad I bought my LM4, the new algorithm looks like it can easily be used for some proper measurements for flashlight reviews. It may not be on the level of professional gear, but just to tell people CCA, CRI and R9 of a light It should now be perfect.