Last week i tested the performance of the garmin vivomove sport, which, out of all the garment, watches ive tested so far, seems to have the best sleep stage tracking, like most new garmin watches. The epix 2 also has advanced sleep stage, monitoring which stretch your sleep stages. Oxygen saturation of the blood and breathing rate can the more expensive epix 2 match the performance of the vivomove sport to check. If the epix 2 can attack my sleep stages, ill, compare to an eeg device called the dream – 2 that can actually measure my brain waves and has been shown to be relatively reliable at sleep stage tracking here i show an overview of the sleep test results for Getting an overall impression for how well the epix 2 performs the dream. 2 should likely be good enough. However, the gold standard would be polysomnography, which i would also like to try on epics 2 in the future. Now, on top are the sleep stages, as recorded by the eeg device and on the left are the sleep stages as recorded by the epics i wore both the eeg device and the epix 2 to bed for 4 nights and ill see how close the predictions of The epics 2 are to those of the eeg device, each column sums to 100, meaning that we can see what percentage of each of the sleep stages. According to the dream to eeg headband was predicted as each sleep stage by the epics 2.

. If they perfectly agree, all values on the diagonal should be 100. First of all, we see that only about 24 of what was deep sleep according to the eeg device was also deep sleep according to the epics 2., most of what the eeg device identified as being deep sleep was actually detected by the epics 2. As being light sleep, this discrepancy is mostly due to the fact that the garmin epics 2 detected very little deep sleep. We can see that by looking at the individual knights, like this example knight right here. On top, we have the sleep stage. According to the dream. 2 eeg headband, with the clock time along the horizontal axis and the sleep stages on the vertical axis on the bottom. We have a similar plot, but now for the epix 2 ive highlighted all the eeg recorded deep sleep and purple here and, as you can see what the epix 2 marked as deep sleep was also marked as deeply by the eeg. However, in general, very little deep sleep was detected thats. Also what we see for this second night in general, very little deep sleep is detected light sleep agreed okayish with the eeg device at about 61. If they did disagree, this was mostly by the epic 2 predicting ram sleep ram. Sleep also did not agree very well between the eeg device and epics 2., only 42 percent of what the eeg device marked as ram sleep was also marked as ram sleep by the epics 2.

. A larger percentage was actually classified as light sleep by the epic 2. thats, also confirmed by looking at the individual knights in red. I marked a ram sleep, as recorded with the eeg device and, as you can see, theres only marginal agreement between both devices. You go through roughly 4 to 6 sleep cycles, each night, each one, starting with light sleep and deep sleep which is marked in blue for the eeg device and each one ending in rem. Sleep marked in red for the eeg device. As you can see, i likely had one two three four complete sleep cycles this night and because the ram sleep detection by the epics 2 does not seem that great, i would judge. We cannot see the sleep cycles based on just the data from the epics 2.. Awake detection did show relatively high agreement with eeg device at over 80 percent if the epics 2 did classify awake time as something else than the eeg device. This was mostly as light sleep and ram sleep, as you can see here in green. The longer awakenings, according to the eeg device, did in the match with the epics 2, though, the epics 2 also detected some extra awake moments and we mostly see the same thing for this knight right here. Theres a partial match, but also some extra awake time is detected by the epics 2.. For those of you that watched my video on the garmin fenix 7, these results might look somewhat familiar.

The phoenix 7 also did not seem to be very good at sleep stage. Tracking the similarity between the epics, 2 and phoenix 7 makes sense, given that the design and senses of these devices are identical for the most part and they mainly differ in their screen types. Now, while i was wearing the epix 2 on my right arm, i was also wearing the phoenix 7 on my left. This gives us the opportunity to check how similar the sleep tracking of the epics, 2 and phoenix 7 actually is. First lets see what percentages of each of the sleep stages both watches predicted. Interestingly, the epics 2 predicted the way less deep sleep than the phoenix 7.. The epics 2 only detected about 5 percent deep sleep throughout the four nights, whereas the phoenix 7 predicted about 11, which is still not a lot. The percentages of light sleep are roughly in the same ballpark at a little over 60 percent. The epics 2 did detect a bit more rem, sleep and awake time than the phoenix 7.. However, how similar are the sleep stage predictions of these watches over the night? Well thats displayed here here we see for each of the 4 sleep stages predicted by the epics 2. If the phoenix 7 agreed with it. Let me explain what i mean, for instance, this 89.4 percent here means that about 89 of what was marked as deep sleep by the epics 2 was also marked as deeply by the phoenix 7, and we can, of course make this same plot using the phoenix 7 Sleep stages as a starting point: now we see that only about a third of the deep sleep detected by the phoenix 7 was also marked as deeply by the epics 2.

. Now this is linked to what we saw before, which is that the phoenix 7 detected more than double the deep sleep of the epics 2. light sleep in both cases showed moderate agreement between both devices at about 75 rem. Sleep also shows moderate agreement at best with an overlap of 53 and 68 and finally awake time, overlaps pretty poorly between both devices with a maximum of 36 percent. In both these matrices, okay, those were a lot of numbers, but what did it mean? Well, i suspect that the differences between the sleep tracking of the phoenix 7 and epix 2 is not due to the devices being different. I think its more likely due to a lack of robustness in the algorithm. Now both devices were always running the same firmware version. So they were likely using the same algorithm. However, the devices got slightly different inputs because i wore them on different arms and the slightly different heart rates and movements. Measured were enough to result in vast differences in the sleep tracking of both devices. Assuming this is the case, this implies that the algorithm used by garmin on these watches should be improved at least thats. My hypothesis, if you have any other ideas, please let us know in the comments below i still do not understand why the garmin, vivomoo spore seems to do so much better, no matter which armor word on this graph shows an overview of the agreement of different watches.

With the eeg device with the vivomoo sport right here along the horizontal axis, we have the average agreement over the four individual sleep stages and on the vertical axis we have the agreement of the worst sleep stage, the better the agreement with the eeg device, the more To the top right, the device is, as you can see, the best agreeing devices include different fitbits. In this case the fitbit sends inspire 2 and charge 5 to whoop, strap 3.0 4.0 and within sleep analyzer. As you can see, the phoenix 7 venue 2 and venue 2 plus do not have a high level of agreement with the eeg device. However, the garmin vivomove sport does quite well if we now plot the epics 2 in the same plot, we see that it is quite close to the test results. I got for my original phoenix 7 test. Both have an average agreement over the 4 stages of about 50 percent. Interestingly, if we now also plot the new results for the phoenix 7 right here and with new results, i mean for the same night, i tested the epix 2. We see that this agrees slightly. Better than the previous testing, with an average agreement of about 57, it mostly shows up as better in this graph, because none of the sleep stages have a very large disagreement, meaning that its higher on the vertical axis. I suspect, however, that this marginally better agreement is due to some random variation and a relatively low number of knights tested, but future testing will have to show if this is the case or not.

Overall, though, the sleep stage tracking of both the epics 2 and phoenix 7, does not appear to be very good. However, some people might value the simple metric of the sleep score more than the detailed sleep stages. So how similar are these between both devices? That is displayed in this table right here. In total, i slept with both devices for five nights and, as you can see, there were quite some differences in the sleep scores. I got for both devices the fifth day, for instance, the epics 2 gave me a score of 73, whereas the phoenix 7 gave me a score of 88. most nights. The phoenix 7 actually gave me a higher score. I suspected this might be due to the larger percentage of deep sleep predicted by the phoenix 7. only the first day. The phoenix 7 gave me a lower sleep score. However, this night, it said it could not attack my sleep for part of the night, so thats probably the reason, so i left this night out of all the other analysis, i showed you so far. Next lets move on to heart rate tracking. However, as you might have noticed, my reviews of garmin watches came out a lot later than that of most reviewers. That is because garmin did not. Let me the watches to review before they were launched, and i have to buy all the watches myself. If you want to help the channel get noticed by garmin a sub to the channel and pressing like on this, video would really go a long way and it will also help me get these videos out faster but of course, its totally up to you now.

We previously saw that the heart rate tracking of the phoenix 7 was decent, though not great. What about the epics 2. lets take a look at the results during spinning, cycling and weightlifting in the next test. Ill, compare the heart rate measurements of the epix 2 against the polar h10 ecg chest strap, which can generally record my heart rate very accurately, well start by looking at the easiest type of exercise for a watch to track cycling indoors. Now this involves very little movement or tension on my arms and should therefore produce less noise. Here we see an overview of that accuracy. Each dot here is a single heart rate measurement, with long the horizontal axis, the value according to the polar h10 hg chest trap and on the vertical axis, the value. According to the epics 2. theres, a pretty good agreement between the ecg chest strap and the apex 2, as most points are along the blue line. However, there are still a few points away from the blue line, both above it and below it, indicating it detected both the two high and too low heart rate. Looking at this first spinning session, we see that overall theres a pretty good agreement along the horizontal axis. We have the time and my heart rate is along the vertical axis in blue, i plotted my heart rate according to the polar h10 ecg chest, strap and then red is my heart rate. According to the epics 2.

, as you can see mostly theres, a good agreement, though sometimes theres a slight delay in the epics 2 picking up changes in my heart rate, as you can see, for instance, here here and here in this second spinning session, the delays and Picking up my decreases in heart rate were a bit bigger, as you can see, for instance, here here, but also here. However, overall, the results are not bad, as you can see, for instance, also in this training session. All in all, this is still not bad compared to many other devices. However, there are definitely a few better devices out there. Next lets take a look at cycling outside, for which the overuse is played here generally. This is more difficult for most watches since theres. Much more movement than bumpiness, as you can see, this doesnt look bad. Overall, though there are definitely moments of deviation, especially with epics 2 detecting a too low heart rate. Looking at the individual rides, this becomes even clearer. Some rides are really good where the epics 2 is able to follow, along with the chest, strap almost perfectly and thats what we see here in this example, we see mostly the same thing for this right right here, however, for some rides, this is much worse like For this ride right here, where we see quite a big deviation and also for this ride right here in the beginning, overall, the ratio of very good rides to mediocre or bad rides is about 50 50.

Based on my testing, these results are not amazing, so this is definitely something to keep in mind when using the epics 2 for cycling. It appears to be a bit hit and miss if it can follow my heart rate accurately. Next lets take a look at weight. Lifting now this is generally more difficult for watches, given the tension on my wrist and arms, and indeed we see a familiar pattern. Weve seen for many watches, the epics 2 can track my heart rate in between sets, but the moment i start a set. It cannot keep up with the peaks. In my heart rate, we mostly see the same thing for the second weightlifting session, though a few peaks were now captured by the epics 2.. Overall, though, i would not recommend the epics 2 as a heart rate monitor during weight. Lifting i actually also wore the phoenix 7 at the same time, but on the other arm, so lets see how this performed in comparison to the epics 2. lets start with cycling indoors on the left are the results for the epics 2 and on the right. The results for the phoenix 7. now i limited myself to just those exercises where i was wearing both devices at the same time as you can see, the results look very similar for both devices and they perform about equally well for spinning next, looking at cycling outside This also looks very similar for both devices. The phoenix 7 actually seems to perform quite a bit worse during this test.

Compared to the test. I did a few weeks back, which i cannot really explain. I always wear the watches well away from the wrist bone and quite tight, and i did not change anything about that part of the procedure. Interestingly, if we plot the individual rides, the two watches do not always show their mistakes at the same moment. Here i applaud the ecg chest, strap in red, the epics 2 in blue and the phoenix 7 in green and, as you can see here in the beginning, the epics 2 followed along quite well, but the phoenix 7 struggled. However, during this other ride right here, it was the phoenix 7 that did quite well here in green and the epics 2 in blue struggle a bit more. Finally, for weightlifting, we also see more or less the same result for the epics 2 and the phoenix 7., as you can see, both watches were not able to pick up on the peaks in my heart rate, during each of the sets that i did, whereas heart Rate is generally measured using green light. Red and infrared light are used to measure your oxygen saturation over the last week i measured my oxygen saturation at ground level in the morning and evening using the epix 2 and phoenix 7.. At the same time, i also recorded my oxygen saturation using a dedicated fingerprint oximeter at ground level. My oxygen saturation should be within my normal range, which is generally between 97 and 100 percent and should not fall below roughly 95 percent on the left are 27 measurements.

Taken with the epics 2 and on the right, the measurements taken with the fingerprint oximeter, as you can see, the epics 2 quite often measures quite low spo2 values. Im honestly, not sure why this is, but this does mean you might need to be more careful when you do get a low spo2 measurements not to read into it too much. Just for reference here are the results of the measurements taken with the phoenix 7 over the last week. Interestingly, a much larger percentage of the measurements is in the normal range for the phoenix 7.. Now i cannot really explain this difference, since it has an almost identical design to the epics 2.. Overall, when i likely had normal spo2 values, the epics 2 reported quite a few low measurements. This means that, at least for me, the measurements of the epics 2 do not appear to be super reliable. What about gps tracking? I tested that, while cycling to and from work, and i wanted to test two things one. How long does it take for the watch to get a gps signal and two how well the gps signals overlap when cycling the same route, multiple times that is displayed here for four times i cycle to work? I started the activity the moment i was ready to leave and i did not provide the watch with any extra time to acquire the signal. The green marker indicates the moment it connected to the gps signal and, as you can see, it almost always acquired the signal.

Almost instantly, which is good, it needs a few seconds to get a more accurate location, but it quickly locks on. If we look at the actual agreement between the signals, this is pretty good. For the most part, though, there are some moments where they deviate a bit. More all in all, theyre generally very consistent, which seems to be in line with what we previously saw for the phoenix 7, and we see the same thing for cycling back. This signal is acquired quite quickly, and the agreement between the four different signals appears to be really good, as you can also see here. For instance, there are some moments with a bit more deviation, however. Overall, it looks really good and the gps tracking is amongst the best of the devices ive tested so far, im pretty impressed with the gps tracking of the epics 2. But again, this is not unexpected, given that we saw the same thing for the phoenix 7.. However, what is still really weird is that the sleep tracking of the epics 2 and phoenix 7 appears to be so much worse than that of the garmin vivomove sport. In my testing, i actually connected the epics 2 to the same account that the viva move. Sport was connected to before this was to make sure it wasnt, due to the special settings on that account or due to the fact that the sleep tracking algorithm might have more or better data on that account.

However, i did not make any difference at all in the comments. Some people speculated that garmin, indeed, is testing a newer algorithm on the vivomoo sport. However, i couldnt find any references to this online, so this is speculation for now, so it might indeed be that firmware. Updates still bring improvements to the epics 2. for most of my testing, the firmware was at 7.20, but for the final workouts and night, this was updated to 7.24. However, on this very limited data set, i did not see clear improvements, so whos the epic stu for well. I still think that the battery life and the maps are the main selling point for this watch. If you like to go on long hikes or ski trips or go camping a lot without the possibility of charging a smartwatch, this could definitely suit you now, which should you choose in that case the epics 2 or phoenix 7? Well, the user experience of the epics 2 was significantly better for media, with the phoenix 7., its just so much nicer to have a clear, bright and responsive screen. Though the battery life does suffer a bit. Personally, i would go for the epics 2. If i had to choose between these two, regardless of which device you choose id recommend connecting an ecg chest strap if you want to track your heart rate during exercise. However, i do feel that for most people, other watches, both from garmin and other brands, are likely a better option.

Other garmin watches have many of the same health features and have similar or the same sensors at a much lower cost. The garmin venue 2 did quite well in our heart rate testing, for instance, and last week we saw that the viva moves port appears to be better at sleep tracking. Both of these devices are much cheaper than the epix 2 and ill link those videos up, as was pointed out correctly by some of my subscribers. There are some features more geared towards endurance sports having to do, for instance, with performance analysis and training load that might not be available on all garmin watches. Now, as a smartwatch and heart rate tracker, the apple watch is definitely much better than the apex 2, but keep in mind that the battery life is not very good and that you need an iphone check out those videos right here. The huawei white gt3 also has a great heart rate tracker and is about a third of the price of the epics 2, and that video is linked here. So whilst theres. Definitely a group of people for which the epics, 2 and phoenix 7 are the best choice. Id really recommend you consider if the features it adds are worth higher prices now. I hope this video provided you with some value.

https://www.youtube.com/watch?v=YJRUW0azIqk