Sleep as android getting better? (New REM-detection algorithm)

Hey!

I don't find many threads regardning SAA here. I get the impression that SAA isn't popular as a lucid dreaming aid, and that's for a reason.
I know, sleep tracers using barely accelerometers or sonar aren't too acurrate, and when it comes to REM-detection, it's a completley different story and alot else to concider. Nothing beats EEG or detection of eye movements.

If you haven't heard, they have updated their algorithm for detecting REM: https://sleep.urbandroid.org/how-do-...e-your-dreams/ So this is interesting, maybe there's hope after all.

I've been using the Xiaomi Mi band 2, which is cheap and recommended by the devs of SAA, it also supports HR-monitoring which according to them also helps with REM-detection, so i should get the best possible results by using that. However, I haven't had much success with this, i rarely recall getting any lucid dreaming cues in my dreams. The best use for SAA i had so far is to wake me up after a certain time when doing WBTB.

Some examples:

This doesn't look too bad (clouds represents lucid dreaming cues, red line heart rate)

Attachment 10243

But here, there's missing cues in REM, and i don't know why. This happens alot unfortunatley.

Attachment 10244

How it used to look like, less REM, and clouds all over the place.

Attachment 10245


My guess is that SAA has become more conciderate in it's approach to accurately triggering lucid dreaming cues, rather than doing more of guesstimates as it used to. That's really good, if it's more accurate of course. But in a way, maybe it's gotten too strict? As in many of my graphs are missing cues in REM, hopefully this will get better in future updates.

The only way i can possibly find out it's accuracy is to set lucid dream cues on high volume to wake me up. Then hopefully, I can then tell if i woke up from a dream or not, but then im not always 100% certain that i did wake up from a dream. This seems too painful though.
It's like i have to wake up every night, for every and one single cue, and i might even have to do this several times for each cue to confirm it.

I think one of the biggest advantages is great support for Tasker, which makes it highly customisable, especially that you can configure lucid dreaming cues (when it think it's REM, that is) to trigger tasks. If REM detection were fairly accurate, this would be awesome.

In conclusion, is this a worthy help in getting lucid?
I would'nt resent it completley. I would say a maybe, time will tell. It doesn't hurt to try a few nights and see what this new algorithm has to offer.

As i said, there's a lot of factors to concider when trying to detect REM by using these "primitive" technologies.
Not to mention the factor of you as an individual and more "daily" factors (caffeine, supplements, medications, mental state, health conditions etc). It might work gold for some, others not.

If I do find (hopefully also recall) these audio cues entering my dreams rather frequently, then it's a success for me.
But I wouldn't hope too much and im not putting too much effort into it. I might buy a "real" lucid dreaming device in the future.

I'll get back with results.

Any thoughts?