Huawei Watch Science Test: GT 2 Pro Review By The Quantified Scientist

Hello everyone, my name, is robbed, and I'm a postdoctoral scientist based in Vienna, Austria in this video. I do a comprehensive set of scientific tests on the Huawei watch, gt2 pro first I'll test the quality of the sleep tracking against a scientific EEG monitor. Second, I will test the heart rate measurements and finally, I will evaluate the accuracy of the step counter on my channel. I've tested many sleep apps and sleep trackers. So far, I've not tested a Huawei device with the true sleep algorithm. Yet so this analysis of the gt2 pro will be the first look at how that algorithm performs.

Additionally, when it comes to the heart rate measurements, the gt2 pro uses the true scene 4.0 plus technology, and it will be interesting to see how this performs. As always, I do not want to waste your time, so timestamps are in the description below and also on the timeline before getting to the test. I would like to provide the most important background information on the Huawei watch. Gt2 pro the gt2 pro is along the higher end of Huawei's watches, with a very premium build quality. I see it as the premium version of the gt2 and gt2e, in contrast to the latter 2.

The gt2 pro is made of titanium, sapphire, glass and ceramic and comes with wireless charging. It has a 40-day battery life with typical use, which is pretty good. I also own the gt2e, the cheapest version in the Huawei gt2 lineup, and I'll release a review of this device in a couple of weeks and see how it compares the Huawei watch. Gt2 pro has several sensors, including an accelerometer, a gyroscope, a geomagnetic sensor, an optical heart rate sensor and an air pressure sensor at the end of the day. This means that the watch continuously measures, your heart rate, oxygen, saturation, steps, sleep stages and stress levels that most of these are measured continuously is pretty cool enough.

On the background, let's get to the test, and I want to start off with the sleep test for the sleep comparison, I wore the gt2 pro Tibet for three nights. At the same time, I also wore this portable scientific EEG device, and I recorded myself using an infrared camera, the EEG device measures, brain waves and muscle movements. It's called the hypnotized z-max and is being used by several of my colleagues in scientific studies. If you're interested in this device for scientific studies, I will link it below. I manually went through the recording of the EEG and scored each part of the night for the different sleep stages.

I also exported my data from the Huawei servers and wrote some code to extract the data in a usable format with the infrared recording. Furthermore, I can actually check what my movements were like and see if the gt2 pro correctly predicts when I'm awake a quick but potentially important side note before I show the results. The gt2 pro uses the Huawei true sleep 2.0 technology now Huawei says it conducted a joint study with the center for dynamical biomarkers at Harvard medical school to validate their sleep tracking algorithm. I found an abstract from a conference that describes this study, which I will link below, but this provided very little details. Also.

It seems that they did not compare the sleep tracking of the Huawei algorithm to scientific EEG measurements, but instead to measurements based on heart rate and breathing rate. So I'm unsure how much trust to put in this study. Now, let's get to the results I obtained. Let's first have a look at the accuracy over the three individual nights, after which I will do a statistical overview analysis. Here we see the first night I recorded on top.

You see the sleep stages as they were recorded using the EEG device on the horizontal axis. We have the time of night and, as you can see, I went to bed quite late, a little after 1 pm on the vertical axis. We have the different sleep stages, that being deep sleep, light sleep, rem, sleep and awake. Now these sleep stages are plotted in the order that are usually also displayed in research on the bottom. You can see a similar plot, but now for the sleep stages as they were recorded using the gt2 pro.

If we first look at deep sleep according to the EEG, which is marked here in purple, we see only a partial overlap between the gt2 pro and the EEG device. In the beginning of the night, the match is ok, but later in the night, the gt2 pro detects a lot of deep sleep that was not really there. Next, let's look at ram sleeve, which are marked here in red. Here we see that the overlap between the gt2 pro with rem sleep is not that great a lot of the ram sleep was actually detected as light sleep by the gt2 pro to see the sleep cycles. I added non-rem sleep in blue and again marked rem, sleep in red.

Each sleep cycle starts with a combination of deep sleep and light sleep together called non-ram, and they always end in rem. Again, non-ram is marked in blue and ram in red. Looking at this knight of data, I would say it's difficult to pick up on the sleep cycles. Just looking at the gt2 pro. There are some parts of the knight that hold promise, but I will need to look at more nights to make my final verdict.

Next. Looking at the times, I woke up during the night marked here in green, the gt2 pro performed. Okay, it picked up on the longer awakening, but not the two shorter ones. The shorter awakenings were actually detected as rem sleep by the gt2 pro. When it comes to sleep start and sleep end, we see that the gt2 pro performs very well.

It was very accurate at detecting sleep on set and the moment I woke up now. Let's look at the second night of sleep data which I displayed here. Looking again at deep sleep first marked here in purple, we see a similar situation to before the deep sleep detection right at the beginning of the night is okay, but later in the night, it detects deep sleep when there was no deep sleep. Looking at the general patterns for the gt2 pro I see for this night, they only partially make sense. In my normal sleep cycles, I would have more deep sleep at the beginning of the night and more rem sleep at the end.

The gt2 pro actually displays a lot of deep sleep at the end as well, which is not typical in general, and also not for me. Next, looking at rem, sleep marked here in red. This looks a bit better for this knight. It seems to have a bit too much ram sleep, and it misses this section right here. However, overall, it looks slightly better than we had for the first night, looking at the sleep cycles, this also means that this looks a bit better than it did for the first night.

However, personally, I would still have difficulty picking up on the sleep cycles when looking just at the data from the gt2 pro looking at awakenings. Again, we see that the gt2 pro picked up on the longer awakening, but not the shorter one. Finally, looking again at the moment, I fell asleep and when I woke up, this appears to be where the watch shines, since it quite accurately picked up on these moments. This is the final night, and for this knight we see that the deep sleep accuracy is pretty poor. There is very little overlap with what was actually deep sleep and the watch in general predicted way too much deep sleep, especially in the middle and end of the night ram sleep tracking for this night is mediocre somewhere in between the quality of the first two nights we looked at again.

The sleep cycles are only partially visible, though it's not terrible. It definitely performs much worse compared to, for instance, the Fitbit inspired 2. I looked at last week. Finally, looking at awake detection, the gt2 pro is again adequate, even though it misses some shorter awakenings. It does pick up on the longer ones again, looking at sleep on set and the time of awakening the gt2 pro performs really well.

Based on these results, I would say that the gt2 pro is perfect at detecting when you fall asleep at night and when you wake up in the morning. Additionally, it detects awakenings during the night quite well, detecting the actual sleep stages appears to be less good. However, to get an even more objective view of the results, let's calculate some statistics regarding the consistency between the Huawei watch, gt2 pro and the EEG device. First, let's look at the total percentage of each sleep stage that the EEG device here on the left and the gt2 pro on the right predicted. Overall, these percentages are not far off.

We see that the gt2 pro predicts a little too much deep sleep, which is consistent with what we saw in the individual nights to compensate. It predicts too little light sleep. The total percentage of rem, sleep and the wake time, on the other hand, are pretty accurate, more important, even than checking the total percentages is checking if the gt2 pro predicts the correct sleep stages at the right time. That is what I displayed here. On top, we have the sleep stages according to the EEG device and on the left, the sleep stages according to the gt2 pro now, each column here sums to 100, meaning that we can see what percentage of each of the actual sleep stages was recorded as each sleep stage by the gt2 pro.

If we first look at deep sleep, we see that about 40 percent of what was actually deep. Sleep was also predicted as deep sleep by the gt2 pro most of the rest was actually predicted as light sleep. If we look at light sleep. On the other hand, we see that a little over half of what was actually light. Sleep was also predicted as light sleep, whereas the rest of it was about evenly distributed between deep sleep and rem sleep.

Next, only a quarter of what was actually rem. Sleep was also predicted as rem sleep most rem. Sleep was actually tracked as light sleep by the gt2 pro, though only very little of it was confused with deep sleep. Finally, most awake time was correctly detected as awake time, though a significant portion of it was still tracked as rem sleep overall. By adding this new information to the individual night, sleep analysis we looked at before my conclusions.

Don't change much. The gt2 pro is perfect at detecting when you fall asleep and when you wake up in the morning, and it detects clear awakenings during the night quite well, detecting the actual sleep stages is not the best feature of the sleep tracking of the gt2 pro. To put that into perspective, let's look at the results of the gt2 pro and the Fitbit inspire 2. I looked at last week side by side on the left. Here we have the results of the gt2 pro we just saw and on the right results for to inspire 2.

As you can see, to inspire 2 had a much higher accuracy for all the different sleep stages, always predicting the bulk of the stage correctly specifically for sleep stage prediction. I would therefore recommend a Fitbit product like the Fitbit sense charge 4 or to inspire 2. We see here over the gt2 pro for the next set of tests. Let's take a look at the heart rate, accuracy of the gt2 pro to test the heart rate accuracy. I will compare it against the polar h10 ECG chest strap, which is generally considered to be one of the most accurate consumer devices available for heart rate measurements.

I wore both the gt2 pro and the polar h10 chest, strap for 4 spinning sessions and 3 weight lifting sessions. That way, I can check my heart rate at different heart rate ranges. Let's take a look at those results here I displayed an overview of the heart rate accuracy. Each dot here is a single heart rate measurement within the horizontal axis, the value according to the polar h10 chest strap and on the vertical axis, the value according to the gt2 pro, the blue line, indicates perfect agreement, so any value along this line had roughly the same value for the polar h10 and the gt2 pro the red line indicates those measurements where the value according to the gt2 pro is half that of the actual value according to the polar h10. The reason I added this line is because, in the past I've seen that many devices measure about half the actual heart rate when they make a mistake, the more measurements there are in a certain area, the darker black, the color, as you can see overall, there's a pretty good agreement between the gt2 pro and the ECG chest strap, since most of the points are along the blue line.

Only here in the lower heart rate ranges do we see significant deviations between the gt2 pro and the polar h10, since the points are below the blue line. This means that the gt2 pro detected a too low heart rate in these moments. Let's have a look at the individual training sessions to see if we can find the underlying cause. Here you see the first spinning session along the horizontal axis. We have the time and my heart rate is along the vertical axis in red.

I plotted my heart rate according to the polar h10 chest, strap and in blue is my heart rate. According to the gt2 pro, as you can see, I took three short breaks in the spinning session, where my heart rate would dip for this first spinning session, we see pretty good agreement. The blue line almost perfectly overlaps, the red line only for the first two segments. We see a slight delay in detecting the increase in my heart rate. Let's have a look if we see this type of artifact in the other training sessions as well.

This is the second training session, and here we see almost perfect agreement between the gt2 pro and the polar h10. So that's perfect for the third training session. Here we see some slight deviations, but nothing major. So again, this bodes well for the fourth spinning session. We see some more deviations now, I'm not sure why this training session was different.

The positioning of the watch on my arm might have been off slightly overall, though, since the other spinning sessions were pretty good, I'm willing to give the gt2 pro the benefit of the doubt for now. Next, let's take a look at the weightlifting sessions for this first session. We see something that is typical of many of the optical heart rate sensors. I've looked at before. Basically, we see that the gt2 pro can track the overall patterns in my heart rate changes, but not the sudden spikes in heart rate that accompany each set.

I did because during weight lifting I flex the muscles and tendons near my wrist, this makes it difficult for the watch to accurately detect the sudden changes in heart rate for the second weightlifting session. We see here we see a very similar pattern. Again, it misses most of the peaks, and the same is true for the last weight lifting session right here it seems only after I've finished my set and my muscles are relaxed again. Does the gt2 pro suddenly pick up on my increased heart rate, as you can see at the end of each spike? In my heart rate, overall, the gt2 pro performs well enough when it comes to heart rate measurements, it had a few small artifacts when doing cardio workouts and one spinning session was a bit more off, but overall it performed decently during weight. Lifting it did have more serious issues.

So I wouldn't buy it for that purpose. However, please keep in mind that this is true for almost all wrist worn wearables. In a future video, I will test our heart rate accuracy using many more days of workout data to confirm these results. The Huawei watch gt2 pro also features a step counter to see if this counts. My steps accurately, I went out and took exactly 4 000 steps in segments of 1000 steps now to get an accurate reference step count.

I manually counted my steps using this tell counter. Let's take a look at the results, as I mentioned, for the step counting test, I went out and took four times exactly one thousand steps I wore the gt2 pro on my left arm and I alternated holding the step counter in my left and right hand for each set of 1000 steps, which is what the left and right labels refer to. These are the actual steps counted by the gt2 pro. As you can see, they were very close to the actual 1000 steps I took for each of the four segments. It was only 32 steps off maximum, but it always over counted slightly just to put that into perspective.

Here are the steps as counted by the smart band 5 I wore at the same time as you can see, the mi band 5 was slightly more accurate, and it did not over count as much overall, though the step counting was very accurate in a future video. I will do an even more detailed step counting test in that video. I also want to test if the gt2 pro gives any false, positive steps. With that I mean, does it count any steps when it's not supposed to count steps, for instance, when I'm typing or cycling? I also want to see how it performs on other people. Overall, I think the gt2 pro performed reasonably well when it comes to sleep tracking, it performed very well at detecting the moment.

I fell asleep and the moment I woke up, and it was also reasonably good at detecting awakenings. However, detecting the specific sleep stages was something it performed less well at. If you're interested in tracking your sleep stages using a wrist, worn wearable, I would recommend a Fitbit product like the Fitbit charge, 4 the Fitbit sense or the Fitbit inspire 2. I looked at last week when it comes to heart rate measurements. I was quite pleasantly surprised.

The gt2 pro performed well during cardio workouts, most of the time. However, if your main sport involves lifting weights, I would recommend buying an ECG chest strap, since this is the only way to get reliable measurements during that type of workout. When checking the step counting, I found the gt2 pro to be good. It counted the correct number of steps while walking. However, I still need to test if it counts any steps when it's not supposed to count steps, for instance, while cycling or exercising on my home trainer, would I recommend, buying the Huawei gt2 pro? Yes, and no, it really depends on what you're looking for, if you're looking for a smartwatch with great build quality at a significantly lower price than an apple or Samsung product, and your focus is not on sleep stages.

Then yes go for it. However, if you don't care so much about build quality, the Huawei watch gt2 without the pro or the gt2e likely performs similarly but cost less. If you're, mostly interested in sleep stage tracking, then I would recommend a Fitbit product instead, like the Fitbit charge 4. Finally, I should mention some limitations of the data that I presented here. First, I just tested the Huawei watch gt2 pro on me, and it will be interesting to see how it performs on others.

Second, I only tested it for a limited amount of days and I will do larger tests for the sleep tracking heart rate, monitor and step counting in later videos. Finally, to do a full sleep comparison, it would be good to also test the gt2 pro against the full scientific polysomnography setup. I'm actually building my own polysomnography device using open, BCI components as we speak, and the first components have already been assembled. This way, I will not have to rely on sleep labs for my testing, which is especially difficult in these times of corona. In my videos, I do scientific tests on different devices like the during, the Fitbit and the scan watch and, in the end I hope to use tracking to improve my life.

So if you like that subject and like this video consider subscribing to my channel and also consider giving it a thumbs up, because it makes it easier for other people to find my videos. Thank you so much for watching and see you in the next video you.

Source : The Quantified Scientist