This EMFIT Review looks specifically at the QS EMFIT version for overnight sleep tracking
EMFIT have 3 essentially-identical products to provide sleep analysis; one each for athletes, the elderly and children prone to night seizures. EMFIT QS is the variant for athletes.
On paper the EMFIT QS looks like the dream-product for us athletes, filling a major gap in our analysis of our fatigue and recovery states.
Detailed beat-by-beat analysis of heart rates can offer MUCH more insight for an athlete than ‘just’ what zone they are training in. MUCH more. I’ve had a several year-long quest to try to measure some of the things I thought I ought to be measuring. Essentially that quest was MOSTLY resolved with Garmin’s HRM-TRI and BIOFORCE HRV/Elite HRV/ithlete.
What was NOT resolved was a better understanding of my night’s sleep. I had pretty much given up hope on a practical way of achieving that: wrist based activity trackers are not sufficiently accurate; wrist based optical HR technology is not yet good enough (Oct 2017); and a chest strap is simply not practical – this rules out EVERY SINGLE body-worn device on the market currently claiming to monitor sleep. Most of them are inadequate for athletes in any case.
Why bother? Most of us pretty much know how well our training has gone that day. But most of us have absolutely no idea how our body is adapting to the stimulus that is caused by the exercise. The adaptation is the bit WHERE YOUR FITNESS IMPROVES – pretty key. Furthermore if, like me, you’ve been slavishly using waking HRV for the last 18 months then you will know that up to 5 minutes each morning faffing about to get a reading can get a bit tiresome AND you are never quite sure if that single-figure reading you get is correct.
EMFIT QS is a product with ZERO pieces attached to your body. It all goes under the mattress, blasting your readings up to the net via wi-fi and on to TrainingPeaks and elsewhere. If it works this will be perfect. I would imagine that any serious athlete will need something like this. We are not talking MARGINAL gains we are talking MATERIAL gains…potentially. Let’s see.
Unboxing & Configuration
The box contains a user guide, the device itself, and a power supply. You’ll need a bed and a home WIFI.
The box is pretty good quality and the strap+cord part of the device is of an excellent quality.
IMPORTANT: In order to work, the app first requires your device’s unique code to be entered and registered over the internet. Do this FIRST. ie before anything else…now! OK? Got it? NOW! 🙂
I quite liked the idea of not using Bluetooth on a smartphone for once. However, whilst not that complex, the setup process took me quite a long time. I don’t have WPS and my WIFI password is fiendishly long and prone to keying in errors. It took me a looooong time to set this up. I almost gave up. Still, you only do it once. There is a laptop-based setup process or a smartphone SETUP app (Android-only, iOS to follow).
IMPORTANT: If it goes belly-up during setup then you will need to reset the circular pod. On the reverse of that is a tiny hole. Turn it off. Stick a pin in it. turn it on again. once the red light stops flashing then take the pin out.
TIP 1: My first configuration tip is that when you get to the end of the process your smartphone will be connected to the EMFIT QS device, the latter being configured as a WIFI hotspot. So you probably do not have an internet connection at that point. When the app tells you to verify all is correct over the internet, switch over to your smartphone’s WIFI settings and triple-check that you are connected to the internet (I wasn’t…on two devices I tried).
TIP 2: Secondly the device IS FUSSY about your WIFI/router connection. Mine was forced to go faster from 270 to 300Mhz and was a N router (as in bgn). Putting it back to 270 may have enabled me to connect
TIP 3: I use an unusual IP address range at home. If yours is 10.0.0.X or 192.168.1.X then you will find setup much easier than I initially did.
Configuration – Some thoughts
Using WIFI rather than Bluetooth is a neat idea. There are issues around Bluetooth compatibility over different versions of Android and, if you think about it, there will probably be pretty similar numbers of people with Bluetooth as WIFI, especially those who would want to buy this kind of product.
Supporting Bluetooth requires apps on iOS, Windowsphone and various versions of Android. So this avoid writing lots of smartphone apps as well as a PC/MAC app for those who don’t have smartphones.
However the interface to view the data is a smartphone-friendly, PC-friendly website. it works well on ALL the devices I accessed it on.
On the downside some people don’t like a WIFI device next to their bed, permanently on all night. I’m not sure if their concerns are medically justified ie I don’t know.
aka sticking it under the mattress.
The heart sensing strip is approximately half the width of a double bed.
Obviously if you share your half of the bed with a partner then there will be two hearts to read. You might need to experiment with the positioning in such cases. I was concerned about interference with the data from two beating hearts however I just stuck it under the mattress roughly where I guessed my heart would be above and it worked fine first time.
I used their image as it is hard to lift a mattress up and simultaneously take a photo.
The wifi device can be placed a metre or so away from your sleeping position using the >1m long power cable.
How It Works
Just like scientists can detect the rumblings of a star in a galaxy far, far away. Then so can the EMFIT detect the beat of your heart and your tosses and turns from its location under your mattress. It uses an extremely sensitive motion sensor. We don’t really need to know exactly how it works. But it does. Quite well, actually.
OK so you want to know! It’s NOT an ECG.
It is a BCG device – ballistocardiograph. A highly sensitive, foil-based compression sensor.
Note: It’s not a medical-grade device for diagnosis or monitoring but it uses medical-grade components.
Interestingly EMFIT, founded in the early 1990s, did try body-based sensors based on the same technology many years ago but there were too many artefacts of motion. So they focussed on bedroom-related activity.
Sleep data recording is automatic. Just go to bed like you always have.
Get out of bed like you always have. No need to do special readings. Just get up and go. The data is streamed up to the net by the time your feet touch the ground. In fact you can even look at the data for the night before whilst still in bed, it’s essentially near real-time from what I can gather – although some of the analyses can take 15 minutes to show up.
Just to be clear: Once it’s linked online to the EMFIT server and stuck under your mattress you don’t have to do anything else at all other than sleep and look at the data at a time convenient to you.
The following are measured/recorded:
- Heart Beat
- Respiration Rate
- Physical Activity – turning, exiting bed.
- Whole-night HRV based on frequent, multiple 3-minute sample
The following are derived:
- Sleep Classification
- Deep – physical recovery
- REM – psychological recovery
- LF/HF Ratio -Whole night autonomic nervous system balance
- Recovery Indices
- Recovery Ratio – efficiency of recovery
- Total Recovery – efficiency of recovery
- Integrated Recovery
- Evening RMSSD
- Waking RMSSD – readiness to train
That’s a lot of data and some information. Here comes the “So What?” factor. Your RMSSD is 32…so what? The real benefit to athletes comes when we can contextualise the data and we know to do something about it.
And that is where EMFIT QS does quite well.
The high level information is displayed in a dashboard of key areas of your sleep whre you can see recent trends, visual exceptions/deviations and some interpretation provided by the dashboard. As shown here it seems comprehensive:
INTERPRETATION NOTE: I always go to bed after midnight. So Tuesday’s sleep might be from 00:30 to 08:00 on Tuesday. ie Tuesday = Tuesday A.M. sleep
It is a little complex to begin with but most of us can easily spot where one particular night seems to be out of line with the trend of other nights. You can see that on the above dashboard. I had a HM race on Sunday. Sunday’s sleep (before the race) was short due to getting up early and Monday’s sleep (after the race) should be characterised by recovery. Friday and Saturday’s sleep should be characterised by tapering adaptations.
Perhaps you might also glean an insight into how many days after a race it really does take YOU to recover back to normal?
HRV RMSSD Data
HRV data is special data that virtually no other ‘wearable’ will give you. It enables very special insights into sleep patterns and the state of your body’s nervous systems.
EMFIT gives the ability to drill down into the detail behind the dashboard summaries. As shown above and then the nightly detail below.
So looking at the image above we see a post-race drop to a low RMSSD of 34 from a pre-race high of 38 (high is good).
We can drill down first into Saturday’s taper-sleep as shown below with the RMSSD detail:
This graph is me after a day off training and having a lie in ahead of another day off training before a race.
As graphs go, that’s pretty awesome. It shows a clear upwards trend in “readiness to train” and it also shows my HRrest, at times, getting down to more sensible sub-50bpm levels.
You might also read into the data that the lie-in sleep after my 8am wee-stop was probably not so beneficial.
It confirms and quantifies how I felt and re-assures me that my taper may be going well.
The graph below shows the same RMSSD data but for the night AFTER the race. It looks much more subdued and ‘flatter’ although the white recovery trend-line is still upwards (good).
A quick word on the navigation before we continue to other aspects of the dashboard
On the dashboard the arrows drill to more detail (as just shown above)
Whereas the “i” gives you more information about the data on the part of the dashboard you are looking at, like this:
Back to the dashboard !
A low ANS Balance figure MIGHT indicate that your body is adapting well (literature is not conclusive from what I can gather).
The low Saturday figure is possibly good with expected values being in the 25-to-75 range. However drilling down into the same night’s detail for that information we see the following:
So that shows quite a lot of variability. I confess to not knowing quite what to make of that insight. Does this show that the post 8am lie in was extra beneficial??
Heart Rate & Respiration
The heart rate and respiration tabs both show the absolute average for the night as well as the range of averages over shorter durations. As before these can drill down into the minute by minute detail.
The detail is interesting but I suspect that the variability shown below is probably more reflective of the relevant sleep phase, which is shown separately within EMFIT.
Minute-by-minute changes, or even week-by-week changes, to heart rates or respiration rates are not a sign of getting fitter. Both should FALL as you get fitter but rather over longer timescales.
I would imagine this particular graph is of even more use in some of EMFIT’s other products such as those that detect episodic sleep problems eg sleep apnea, where breathing stops.
The sleep classes do not take a view on my total volume of sleep but rather on the breakdown of sleep types. The green box shows the ideal percentage range. So on the Thursday night (right) my LIGHT sleep was compromised whereas on Sunday night my REM sleep was compromised (REM=mental adjustment!).
My initial thought here was that such a high level summary would be enough. However, drilling down into my post 8am Saturday lie-in, then I did actually get all 3 types of sleep but in quite disparate chunks. Are 6x periods of Deep sleep better than 3x? Or is it the total percentage amount that’s important? Probably the latter.
Another word for recovery is adaptation. IE you did all of that exercise how much is being taken on board by your body in its adaptation?
Pretty fundamental really.
Of course this won’t be a true precise measure of adaptation but it is probably one indicator.
And I think that is one of the keys when using EMFIT. There is no one number or ratio to hit. You’re looking for exceptions and outliers to trends. Something out of the normal. You’re also looking for good trends to keep going in the right direction.
In addition to the dashboard view and the drill-downs from there, there are other views that I’ll briefly cover – these are TRENDs and TIMELINE
This is the exact same data as before. Just that you can see all the previous bits of data for the month-to-date rather than being limited to the week-to-date. You get a nice trend line in some cases too. So this is a neat presentation of longer term trends.
I couldn’t get the timeline view to work in Internet Explorer but it was fine in chrome, as shown below. It is what it is and I ‘m not quite sure how useful that is for me.
I tracked my smartphone app’s view of waking HRV over the same time period, my smartphone app is Bioforce HRV. Bioforce re-bases RMSSD to an index of 100 by using the formula 20lnRMSSD. So I will re-base the EMFIT figures to that same value – as the calculation is simple enough to do. Both also give a measure for HRrest. The smartphone app bases its results entirely on 3 minutes. EMFIT takes a view on the complete night.
With this table of data clearly we are not comparing like with like as the data is arrived at by quite different calculations. But they are supposed to indicate a similar thing.
With the HRV differences I could suggest that the figures are broadly in line except for 23/Mar and 21/Mar and I could further suggest that perhaps on those days Bioforce was affected by the vagaries of only taking one 3 minute reading. Or you could argue that EMFITs data on those nights were corrupted by my partner in the same bed. I might also suggest that normally if I got an unusual Bioforce reading I would take another immediately afterwards; I deliberately didn’t do that for this period of dual measurement.
If I doubted a value supplied by EMFIT I could drill down to the minute-by-minute detail and see if it looked ‘about right’. Therefore all that can be faulted with EMFIT is the nature of the recording technology (the BCG from before). Does it really work? It seems to.
I’m not intending to show you a scientific comparison of two products. Intuitively it feels to me that EMFIT’s measurements are probably more CONSISTENTLY accurate over many months than those I’ve been using with Bioforce over the last couple of years. Sure, Bioforce and the Polar H7 is probably sometimes more precise and accurate for one-off readings but sometimes it isn’t – as demonstrated by taking a reading 3 minutes later with Bioforce and sometimes getting markedly different readings.
Will I stop using Bioforce and save myself 5 minutes every morning faffing around???
Actually yes I will. Hello EMFIT !
Now having just said that I’m going to switch products here comes the BIG caveat.
EMFIT is SO great in one respect. It saves 5 minutes of faffing around every morning. I just get up and get on with my life. I sometimes save 10 minutes when my Polar H7 battery is on the blink.
5 minutes x 365 days = a lot
But I’ve found that I’m MUCH LESS inclined to look at the EMFIT data. There’s too much of it. Look at the following diagram, EMFIT is showing data/information NOT knowledge/Insight and Wisdom.
EMFIT does not present the KEY exceptions to me very well. Sure I can spot them on a graph, maybe…if I take the time.
The great thing with Bioforce (ithlete and other products) is that you get a traffic light green=train, red= no train, amber=think about it. You compare that with how you FEEL and you pretty much know what you are going to do that day. That decision is made in an instant on seeing the result of the 3 minute test.
With EMFIT the recording is done automatically but the interpretation takes LONGER than 3 minutes a day.
With EMFIT I will have to pore through the data on my smartphone going to work or analysing it at some other point of the day. in reality that isn’t going to happen every day perhaps, more likely, only on days when I feel a bit funny and in need of data-reassurance.
So because EMFIT requires no manual interaction with the product whatsoever on a daily basis EMFIT need to automate the exception/alerting process either to SMS me or to email me. I suppose they could develop an alerting app but that would defeat the whole point of the delivery platform they have chosen.
Sharing the Data
There are: team features (below); a link to Under Armour that I could not get to work; and the ability to export CSV files. None of these readily allow sharing with other platforms that endurance athletes might use. Unfortunately endurance athletes use several disparate bits of technology.
Another piece of tech kit, like EMFIT QS, or another software app is ‘interesting’ … until you get bored with it.
It’s better to have the raw data piped into Training Peaks or SportTracks or Golden Cheetah or some other product/’portal’ that you may be more likely to visit more often for analysing trends in sleep alongside trends in performance.
However the problem with sending the data elsewhere is that the ‘elsewhere’ places are not really geared up for this kind of data yet (sporttracks made an announcement this week actually). Sure those software platforms have places to stick the data but not that much else in terms of providing a comparable analysis to what EMFIT can already do.
The sporttracks environment (desktop) lends itself to another plugin to analyse this type of functionality; training peaks is a great place to hit one of the key target markets (ithelte may have got there first with their recent announcement); Golden Cheetah could be a great place to get the right kind of endurance athlete clients on a growing platform of the future.
Apparently there is a facility for a coach to easily look through his team’s stats but I have not seen that.
I could only assume that EVERY professional sports team should have this sort of facility to monitor their highly paid athletes. It would be negligent for them not to IMO.
Certainly triathlon coaches might find it useful, although convincing every athlete to invest in one might be tricky at the price.
But even with such team or coaching functionality the analysis of lots of data is passed from the athlete to the coach. Again, this takes time for a coach to analyse. I’m not convinced they would do it all the time.
The practical upside is that, once installed, you have ZERO work to do to collect all your data.
If you are away from home for whatever reason then you would have to take the device with you and re-pair the device to a new WIFI connection.
I suppose you could create a WIFI hotspot on your mobile phone if you really wanted to use your mobile data all night to continuously stream the EMFIT data.
Whilst Mr Triathlete might only do 10 races a year he might also go on training camps and spend several days away from home around competition venues. You could imagine other athletes in other sports spending even longer periods away from home.
Athletes also, apparently, go on holiday aka ‘training somewhere else’.
Some form of caching seems to be needed. Perhaps tied with a smartphone-based link of some sorts.
As well as caching data, another nice refinement of the product would be to be able to remember more than one WIFI connection. Otherwise you are at the whim of the hotel WIFI or a family/friends WIFI.
I currently use a manually taken waking HRV value with Bioforce HRV.
As said earlier: Because I take readings manually I am ALWAYS instantly aware what the app’s recommendations is for me that day. With the EMFIT, because you don’t have to do anything to record the data, then you may get into the habit of not checking your data each day.
So what might be a handy addition to EMFIT QS is some form of alert feature by email or SMS – this could be for exceptions in your data OR it could be a summary of the metrics that interest you.
This is a great product which far exceeded my expectations.
It contains all the base info in a probably more ‘correct’ form than any other sensible alternative that I have seen using smartphones as a recording device (I’ve seen a few). As far as I can tell it has it all the data that I need and more besides.
It collects all that data in a totally unobtrusive manner. Totally seamlessly without interrupting your life. I’d border on saying perfectly so.
I have had zero problems with interference from my partner’s heart rate at night. You will have to factor this in as a possible source of error. OR buy a large bed OR fall out with your partner!
The relative lack of portability, lack of caching AND the lack of ‘actionable exceptions’ in the presentation layer are where my criticisms would come in. Data presentation certainly looks good but it is missing a trick with these exceptions and alerts – on the positive side the developers could add that relatively easily if they so wished.
Edit: After the review was written EMFIT quickly introduced caching. After some initial glitches, this appears to work fine and should fully allay the concerns of those worried about a WIFI transmitter next to their bed all night. I am not yet clear if this enables the devices to be moved to other locations for several nights on end with the ability to sync when back ‘home’.
The website IS computer/tablet/smartphone-friendly and more than does the job well of presenting the analyses on multiple platforms. I had a minor issue with TIMELINES on IE11 but that really was the limit of the problem.
So: I’m impressed. I’m going to change how I do things with HRV. I’m going to get downstairs for my coffee 5 minutes earlier every day and only consult the EMFIT Oracle when I feel a little fatigued for that all-important TRAIN or NO-TRAIN decision.
EDIT: 18 months later in October 2017 and I still use this. Awesome insights. There is a slightly revamped interface with better metrics and caching.
Price, Availability & Alternatives
The best 3 sleep trackers are EMFIT, Beddit v3 and the OURA Ring.
BEDDIT v3.0 is an alternative, do NOT buy BEDDIT v1.0 or v2.0. As of October 2017 BEDDIT 3 is no longer ‘properly’ on sale. Either the product is about to be discontinued or a version 4 is due. BEDDIT were bought by Apple and it is possible they bought the product just to get the algorithms/metrics that they will then incorporate into the Apple Watch.
The active Amazon prices of BEDDIT and EMFIT are shown in the Amazon links, below. Please click to buy from there to help support this blog. Thank you.