Emfit QS vs. Oura Ring - which is more accurate?

A couple nights ago I happened to talk on the phone for 30 minutes while lying in bed. While we know that sleep trackers can confuse sleeping with not moving, I was surprised to see that even the Oura Ring (worn, I think, on the hand I was holding the phone with), thought I was in light sleep:

Other than that, the stages were somewhat aligned, though nothing like in the polysomnography comparison that the Oura team produced (which could mean that the Emfit just isn’t that accurate). Also, the Emfit QS seems to over-detect deep sleep. I was hoping that the Oura Ring would under-detect deep sleep, but given that a phone call was read as “deep sleep” by the Emfit and light sleep by the ring, that’s likely not the case.

More on my blog - “Which is the more accurate sleep tracker - Emfit QS or Oura Ring?”.

Has anyone used other sleep monitors that had better accuracy (except the discontinued Zeo)?

2 Likes

We have a Resmed S+ sleep tracker hanging around the office here in Berkeley if you want to compare this one - I’d be interested in what you find out.

Hey, does your Resmed show you actual graphs of ambient noise + temperature correlated with sleep activity (motion, restlessness, breath, etc.)? I’m trying to figure out if elevated light and temp levels early morning are causing me to wake up (way too) early, but if all the app gives you is some “sleep depth” graph, and I can’t actually see a plot of the temp and light data, it’s not very useful.

Hi there. Im very interested in what your current findings have been on the emfit qs VS the Oura ring. Im interested in one of these for tracking sleep, and want the most accurate one possible.

I do appreciate the easier userface and app content for the ring, and wouldnt mond a smart alarm feature. But at the end of the day I need whichever is the best atm.

Please let me know. Thank you!

I’d like to wait for the new Oura Ring and test it vs. the old ring on my other hand. I was surprised by how often and how much QS and Oura differed on the recorded deep sleep. Both claim “sleep lab grade precision” but one had to be blatantly wrong.

If the new vs. old Oura ring also give vastly different results, that would be pretty funny (and sad).

If you need to decide before the new Oura ring ships, then look at factors other than accuracy. Emfit is a passive monitor. No need to remember to charge or wear it. Oura is portable - you can quantify your sleep easily in different environments (I found that my best deep sleep ever was in a very cold tent at a festival, after walking the whole day). Oura isn’t influenced by having a sleep partner, QS Emfit might. (On the other hand, I found that having a sleep partner is the #1 cause of poorer sleep quality.)

1 Like

Thank you for your response. I dont sleep with a partner atm so im not worried about that.

I can imagine the differences between the emfit qs and the oura ring to be due to the fact that one is a wearable on your finger. I cant imagine how a ring of that size, just on your finger, is able to track movements for your entire body. The emfit qs to me would definitely represent the more accurate tracker if that is the case. I do not think the newer ring will have night and day accuracy differences to the old one, and im not interested in tracking throughout the day anyway.

However, in your personal opinion, would you say the oura ring feels more accurate?

Also, for purely sleep tracking, Emfit qs appears to be the best, or do you know of something that rivals it

Is the Second Generation Oura ring any more accurate?

I’ve been wearing both the old and the new Oura rings on each hand at night, for about a month (since Oct 12, skipping some days). The two rings generally agree on the bed start and stop time, but I also only put them on during sleep, so it’s not that hard to detect the transition from no finger to finger. They also mostly agree on deep sleep duration, but on several occasions one indicated twice as much sleep as the other. (If there’s interest, I can plot the data).

What concerts me a lot though, is that REM and light sleep detection is still clearly very inaccurate. As many other folks reported, Oura detects them asleep while they’re watching TV or reading in bed. Here’s the most recent instance of that, from yesterday morning:

It’s well established that not remembering wakeups is quite common. That incorrect sleep detection is about 90 minutes off. Can you really trust that the device would accurately detect you waking up for a few minutes due to sleep apnea, or something else bothering you, if you fall back asleep quickly and have no recollection of it?
I can’t.

I’ve raised this problem with the Oura team and they’ve acknowledged it, and said that deep sleep detection is a lot more reliable. I would have to trust them on that, but the two rings showing double/half deep sleep on some nights doesn’t give me much confidence.

Not saying Oura is useless. I did notice long-term correlations - e.g. when I sleep alone, I tend to record more ~45 minutes of deep sleep vs. 0-15, but not always. However, any sort of short-term conclusions, or “insights” about micro-wakeups, seem unwise. For example:

  • basing your day on what the ring says about your “Readiness”
  • deciding to buy the same mattress as a friend’s after you’ve slept in on it for a night
  • thinking that your pet walking on you during sleep isn’t a problem

In the end, after ~5 years of sleep tracking with the QS Emfit, Beddit, Oura Gen 1 and Oura Gen 2 and a bunch of phone apps, all I can say in my case is that, for me the best sleep is 7+ hours, solo, on a consistent schedule (same sleep and wakeup times), and with earplugs. Cold vs. warm, dark vs. light bedroom, didn’t make much of a difference for me (I know they do for others).

So did I really need a fancy sleep monitor to figure that out? The best judges have been my performance at the gym, and ability to focus. Also, the effects of bad sleep can be delayed by one or more days, so I can have a crappy night of sleep and still perform well, if I’ve had good sleep the few nights before. Likewise, after a good night’s sleep, I can yawn all day and feel tired (probably because I had bad sleep some nights before that). This complicates drawing correlations, and sleep monitors further confuse the problem when they display readings like “High readiness, see how far you can push yourself today”.

2 Likes

Forget about your friend’s mattress; according to my Fitbit, the harder the surface, the better I sleep :slight_smile:

Having a record of nothing more than when you went to bed and when you got up can still be valuable, e.g. to keep an eye on how regular your sleep schedule is, or to provide some context to other data like heart rate or blood sugar measurements.

Sure, but you don’t need a $300 - $1000 ring for that.

Worse, the ring can confuse things. Last night I wore the two again, and one reported a sleep start time one hour later than the other, and half the deep sleep.

1 Like

Agreed; I currently keeping track of my sleep times by pressing the button on an old Fitbit One that refuses to die (or be lost…) when I go to bed and when I get up… It’s less tedious than having to open an app to start/stop tracking, and much less tedious than having to review and correct auto-detected sleep periods every day!

1 Like

Been using the new Oura for a month. Got a little skeptical of the sleep stages. I’m disappointed to read about these accuracy issues. My stages are somewhat erratic, leading me to wonder why I’m not getting enough deep sleep. I even tried melatonin, kava, and kratom to see if any improvement would happen. Kratom (2 grams) produced the worst results. Melatonin better. But all of this is mere speculation if the device isn’t accurate. My top goal was getting more restorative deep sleep.

1 Like

I’ve been trying to improve my deep sleep for the past ~2 months by throwing everything I could at it:

  • same bed, sleepwear and finger ring
  • no partner or pets
  • A/C set to 18C
  • daily resistance training
  • grey-out curtain + eye mask
  • nasal irrigation before sleep
  • 33dB NRR earplugs
  • weather (I’ve been living in Medellin, “the city of eternal spring”)
  • no daily stimulants; only rare social drinking
  • f.lux or equivalent on laptop and smartphone
  • occasional intervention: Spectra479 blue blocking glasses (didn’t seem to make any reliable difference)
  • dinner intervention: timing, and with or without carbs (didn’t seem to make any reliable difference)

The average deep sleep duration has been 48 minutes, but I got variations from 0 and 10 to 82 and 89 minutes, without doing or not doing anything special I could put my finger on.

Clear correlations are rare and not that useful:

  • alcohol reduces sleep quality (duh); but not always… some Sangria at dinner knocked my deep sleep to 23 minutes, while a few glasses of white wine + 2 gin&tonics + 1 beer were followed by 52 minutes of deep sleep. Offset by club dancing that night? Who knows.
  • being sick (cold or stomach flu) reduces sleep quality
  • sleeping on couches sucks
  • just sleeping more tends to result in more deep sleep and higher readiness (duh), but again, not always

What’s more frequent is the complete lack of correlation. Average or lowest heart rate don’t correlate with deep sleep. Got 89 minutes once, with AVG HR 63 and lowest HR 57. Another night, AVG HR 61, lowest 50, deep sleep only 53. Another, 62/56, deep sleep 38 minutes. I don’t know how to control the HR anyway. All those nights were following resistance training days including 30 minutes of walking (10 uphill).

My last hope is sleep timing. I’ve only gone to sleep (way) after midnight, but even 2am vs. 6am doesn’t correlate with anything:

After getting only 10 minutes of deep sleep one night when I have kept all the variables I could think of as constant as I could, I wrote this in my sleep log:

10 minutes deep sleep and 50 the night before? What the FUCK have I done wrong?!

So color me a skeptic. If you’re looking for a more enthusiastic crowd of Oura users, head on to the Facebook group.

5 Likes

@Dan_Dascalescu - thanks for this detailed post. I think there’s a pattern in QS reports about sleep stage tracking: initial hope and enthusiasm, and then somebody does a deep dive into the data and finds it very hard to get a useful signal, despite seemingly confident reports from the tool. This goes all the way back to ZEO, which many QS folks loved. The deeper you got into thinking about the data on sleep stages, the less confidence resulted.

3 Likes

Soak ear plugs in canning wax for +44dB of sound reduction. Likely causes tinnitus and ear health problems. Any sleep trackers with formal FDA certification? good enough for hospital use considering liability?

I can’t see the blog post you linked!

I’m the most interested in the HRV from the Emfit QS or Oura ring.
It seems like the Oura has pretty accurate HRV, but what about the Emfit QS?

Not so much of a surprise - Dr. Rhonda Patrick also noticed that while she was relaxed nursing her son, the ring calculated her as being in REM sleep, multiple times.

1 Like

There seems to be a tradeoff between sensitivity and accuracy.

I never have issues with my Garmin, but I’d categorize it as a “lower sensitivity” device. It doesn’t detect naps for example and I don’t think there’s a way to input them even if you wanted to.

Other devices, like Biostrap, Oura, and less so Whoop (I’ve been impressed by their sleep tracking) try to be more sensitive, and therefore have more false positives.

Hi Lars, I retweeted from the QS account. Great write up. I typically silently erase repeats of the same post across multiple topics here on the Forum, since many people follow all topics. In your case, since you’re a regular, can I leave this for you to do? Just choose the one you like best and delete the others.
Gary

Here is my detailed comparison:

So… I’ve been interested in sleep tracking a long time. I used to have apnea but I’ma fitness freak and I no longer do. Sleep studies are expensive, even with insurance, so I’ve been looking for alternatives.

I’ve landed on three devices to get the job done. Any one would likely work if you aren’t looking for quality data in every domain, but I like looking at them all together.

First, best and most sophisticated is the dreem headband. It’s the only device I know of FSA approved in this space and it uses eeg data to know when you’re actually asleep and when you’re really in each phase. Most other devices use algorithms to guess at your sleep stages. For instance, my other two devices will credit me for sleep while I’m tossing and turning or watching TV. Dreem is tough. I don’t sleep nearly as well as garmin or go2sleep tell me I do. The reports and analysis are top shelf. And you can get your raw data. Yay.

Although dreem is awesome and uses oximetry to get your pulse, they don’t provide that data, probably because it’s from your forehead and thus not accurate for saturation purposes. My garmin provides this data, but from this wrist it tends to err in the low side and during sleep registers very low. Not wanting to ignore possible apnea events, I looked into other devices. There are plenty of pulse ox machines but fewer for all night and most don’t have sleep really in mind. Enter go2sleep.

By default this is a ring but rings aren’t very accurate either. They have a fingertip model now. That’s what I bought. Testing it while awake it appears to be much more accurate than the watch. It’s a very Chinese import device that has English speakers as an afterthought but I made it with and started using it a couple weeks ago. It’s pretty good. Comfortable at night, I’ve been very compliant using it. It provides an AHI score each night and a graph of pulse ox. I do compare the occasional dips with the garmin… theory being since the watch seems to trend low a sudden drop on the fingertip device I check against the garmin… I doubt that the garmin would lag so badly to miss a big drop… but I generally trust the go2sleep over the garmin and despite a few questionable low readings (you get that if you move your arm at the hospital too) the score I get implies I’m normal, and I trust that should that change it would register.

Next I plan to do a professional at home sleep study and compare my devices findings.

I realize I might seem obsessed but apnea is no joke and I really don’t want a cpap if I can help it. However, should I need one I will get one…

1 Like