Last Updated on May 27, 2022 by pg@petergamma.org

And for which application is it the most accurate? And how much accuracy do we need? Do we suffer from a burnout or stress, and do not know what to do to make our heartrate drop? Do we swim, bike and run, and use Garmin Connect as our analysis tool? Are whe doctors which want to check whether our patient is still alive, or whether he is near death?

The group of Dr. Milind Y. Desai, cardiologist in Cleveland, Ohio which is affiliated with Cleveland Clinic published papers in which they investigated the accuracy of heart rate monitors for a specific application. They used a Cardionics ECG device as a reference device.

They published papers with a ranking, and numerical and statistical data about the devices they investigated. They found out, that the Polar H7 chest strap is the closest to the Cardionics device, followed by the Apple watch.

These studies came out before the Polar OH1 and and the Verity sense came onto the market. DC Rainmaker and 5KRunner tested the accuracy of the Polar OH1, by comparing the data in Garmin connect to a chest strap. They concluded that the Polar OH1 is a very accurate heart rate sensor, if not the most accurate optical heart rate sensor on the market.

An Australian group tested the accuracy of the Polar OH1 and compared the data to a Nautilus ECG device, and presented numerical and statistical data. They concluded in a paper in Plos One, that the Polar OH1 can be used instead of an ECG device for their specific application. Other papers tested the accuracy of the Polar OH1 with the Polar H10 as a reference device. They confimed the high accuracy of the Polar OH1.

The successor of the Polar OH1 Verity Sense came onto the market. Who expects that a successor model of a sports sensor is less accurate then his predecessor? Nobody knows this, it needs to be confirmed.

Then Polar increased the battery life of the Verity Sense about 100% by only a firmware update. Does this not affect it s accuracy? DC Rainmaker stated, that he investitated this topic in Garmin Connect, and that he saw zero deviation from a Verity Sense before the firmware update and after the firmware update, which he said was confirmed by countless other testers.

How was this miracle be achieved? Did Polar use novel algorithms to reduce noise in noisy PPG signals?

HeartPy – Python Heart Rate Analysis Toolkit for noisy PPG signals

And can the not changing accuray after the firmware update be confirmed by scientific studies with numerical and statistical data? No such study has been done yet to confirm this, as far as whe know.

Recently, Rob ter Horst PhD, from The Quantified Scientist published a Scientific Test of 50 Smartwatches on his Youtube channel:

Rob s reference device where two Polar H10 ECG chest straps. In Rob s test, the latest Apple watches where the closest to his two Polar H10 chest straps, as far as accuracy is concerned, for different applications, for one subject.

Rob confirms findings of Dr. Milind Y. Desai s group, that Apple watches are among the most accurate devices.

He tested also the Polar OH1 and the Verity Sense, but these devices where not among the most accurate onces. But are these accuracy differences relevant for your specific application?

Unfortunately, Rob ter Horst does not systematically deliver all numerical data for all watches in his tests, with a ranking, he publishes not his algortithms, which could be reproduced by other people, to obtain a ranking with absolute values and statistical data. Standard scientific tests from reviewers around the world which test the quality of optical heart rate sensors would be highly desirable.

Now, it is also possible to stream sensor data from Fitbit, Garmin and Polar devices into Home Assitant, but only for wellness data in an easy way. To make further tests in tools like Home Assistant or other analysis tools, it is highly desirable to have more data, to make good decisions.

Another question is which reference device to choose. Athletics usually use a chest strap, which is their most accurate reference device.

Scientists and doctors usually use ECG devices. But which one? Milind Y. Desai group used Cardionics devices, the Australian group who tested the Polar OH1 used a Nautilus device. Bitano pubished a paper where they compared the accuracy of Bitalino ECG devices to the «Gold Standard Biopac»:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6498399/

But Biopac is also a Gold Standard as far as the costs of their devices is concerned, which starts around 4000 USD for entry devices, as far as whe know.

And is a Gold Standard device necessary for your specific validation? Is not also an iWork IX-214R recorder sufficient as a reference device for your study, which can be bought on ebay starting from 300 USD, but has been used for many years by many scientists, and has a long list of accepted papers in which the device was used?

There is no controversy about the accuracy of ECG devices, as long their sampling rate is 16 bit, as far as whe know. The accuracy of the Bitalino Biomedical Toolkit, which has a sampling rate between 8 and 10 bit has been confirmed in a paper with a Biopac device as a reference device.

Whe had a discussion with Wibble in the Pine64 forum, about using a 24 bit sound card device for digitalizing analog ECG signals from Scott Hardens Sound ECG devices.

Wibble was against it, since this would not make any sense. He said he was working for a long time in this field. According to his view, a 16 bit sound card would be sufficient for digitalizing ECG signals from a ECG chip.