Heart-rate tracking accuracy - Sense vs Charge 4

[FCS_1]

Hello, is there a general consensus among users whether the Sense has a better HR-tracking accuracy than the Charge 4? I don't really care much about the Sense's ECG tech but I do hope to get a good HR tracking at rest.

[thatssoketo]

The FitBit Sense is my 4th FitBit... by far the worst at tracking heart rate. 

Numerous complaints on this forum, just do a search.

[Giampi71]

The general consensus is that Charge 4 has much better HR tracking than Sense/Versa 3

[Murray67]

I think the Sense is decent at measuring HR while at rest, but it's completely useless for anything beyond very mild exertion. It seems to have a real problem getting stuck around the 105 BPM level and not wanting to move above that, even if your heart is racing.

Some have speculated that this reluctance to move upwards may be designed into the firmware, in order to prevent false HR spikes at low activity levels. This was something that my Charge 2 and Versa did on occasion, things like momentarily spiking to 160 BPM at the beginning of a low-intensity walk. But if so, they overdid it to a ridiculous degree. If it's a mistake in the firmware, they can fix it, but the fact that they've allowed the problem to continue for so long doesn't give much reason for hope.

[Giampi71]

I suggest to watch the videos made by The Quantified Scientist on Youtube.

This is the HR accuracy test for Charge 4

This is the HR accuracy test for Sense

I think it is not possible to dispute such clear objective analysis that confirm the feeling (and more) of the users on Sense.

[IgorR11]

That's what we see in this test, the sense heart rate monitoring has a very high margin of error, this harms those who follow up for sports purposes. I'm honestly disappointed how fitbit is dealing with this problem, I'll be back to my apple watch until next month!

[Giampi71]

The picture above posted by @IgorR11 has been taken from this just released Youtube video

It is once again another confirmation that Pure pulse 2.0 is just a random number generator.

And no, it's not a case of a faulty device. I did a test with Sense and Versa 3, that share the same sensors, and they are both equally embarrassing. 

[Giampi71]

In the video above the author correctly stated that the test was just personal and result might be different from others.

I therefore want to share my tests

First test 

Comparison: Fitbit Sense (left wrist), Fitbit Charge 4 (wright wrist), Polar Vantage V (right wrist), Pixel phone reading Polar H10 chest band 

As you can clearly see the Sense measurement is nonsense (no pun intended). Charge 4 for the first 35 mins was actually spot on but then started to show anomalous values.

Polar optical sensor matched the H10 chest belt. 

Second test 

Comparison: Fitbit Sense (left wrist), Fitbit Versa 3 (wright wrist), Polar Vantage V in my hand paired with H10 (the setup allowed to wear V3 and Sense in the optimal positions in the wrists)

In this test Versa 3 this time is completely wrong while Sense sometimes is not too bad but far away to be considered precise. Then, like the Charge 4 in the previous test, at 35 mins mark it went nonsense.

Third test 

Date: November 3rd

Comparison: Fitbit Sense (left wrist), Polar Vantage V (optical sensor, right wrist)

So, the average HR of Fitbit Sense is higher than the peak value measured by Polar. 

Just for sake of clarification the HR measured by Fitbit is completely wrong as this was a very slow little run at an average pace of 6'40"/ km

NOTE: the tests were done at the end of 2020. Afterwards Sense/Versa 3 received a couple of updates but, as per changelog, none of the updates concerned the HR sensor accuracy.

Fitbit support/mods often states that it is not correct to compare different brand sensors. I want to answer as following:

- we are measuring a clear physical data, it is definitely correct to compare different measuring system as the data is very clear

- in the first test I compared a chest band (the H10 that is universally considered the best way, consumer wise, to measure HR) with the optical sensor installed in the Polar Vantage V. Two different technologies give identical results. So yes, it is correct to compare 2 different measuring systems.

[t.parker]

@Giampi71 what I find surprising is that without doing anything special but using the watch as usual it actually can provide good enough readings from time to time. Most of the times, I get garbage like this (which I documented here plenty of times) (ORANGE - Sense, BLUE - Garmin + PolarH10):

But then there are times Sense actually can make sense:

 

True, that the charts above are not perfect and there are areas Sense couldn't figure out the HR but if we are talking about wrist HR sensor and two trail runs (variety of surfaces, unstable weather, etc.) so I am quite pleased with such result. I'd say - close enough. I still wouldn't use it for HR zones-oriented training but if I just wanted to know some approximation then this is acceptable. Interesting that sometimes it's almost doing well even with higher intensity training, like this:

The training above shows 7 intervals, each one has to go into anaerobic, then cool down into recovery/endurance. Sense apparently has issues with registering higher HR. What I expect here is probably some aggressive smoothing algorithm that filters out high-freq signals. If only the spikes were reached, that would be quite decent reading (let alone the beginning, again I don't expect wrist HR to be always spot-on).

And it would be alright if not that it's like 1 in 10 times that I get a decent reading. As I said, they come out of nowhere, I don't do anything differently. Sometimes exactly the same run (route, pace, controlled HR) gives different results. Sense can do a good job from time to time. Moreover, sometimes I just put it on, don't care how I wear it, whether it's in the correct place, tight or not - just let it be and surprisingly, it comes out alright. There is nothing really that may help to figure out what was done differently and why 1 in 10 times the readings are good enough. That's why I don't think it is fixable (maybe except for the spikes not being registered - if it's an algorithm, it's fixable).

[KB9WNS]

I encountered the same issue with accuracy.  Not much better then a random number generator.  My Sense has been "measuring" my heart rate even when it's off of my wrist.