Gyroscope Bias Estimation - Unnecessary Max Operation

[alainschoebi]

Summary
This pull request removes an unnecessary std::max operation in the gyroscope bias estimation code. This improves code readability and efficiency without impacting the functionality of the algorithm.

Description
The targeted code block checks if the gyroscope sample's norm exceeds a threshold (kGyroscopeForBiasThreshold). If it does, the function exits early; otherwise, it computes an update weight.

Here is the code block:

  if (gyroscope_sample_norm2 >= kGyroscopeForBiasThreshold) {
    return false;
  }

  float update_weight = std::max(
      0.0f, 1.0f - gyroscope_sample_norm2 / kGyroscopeForBiasThreshold);
  update_weight *= update_weight;

For the std::max operation to pick 0.0f over 1.0f - gyroscope_sample_norm2 / kGyroscopeForBiasThreshold, we need

0 >= 1 - gyroscope_sample_norm2 / kGyroscopeForBiasThreshold

which implies that

gyroscope_sample_norm2 >= kGyroscopeForBiasThreshold

However, this cannot occur due to the previous if-statement if (gyroscope_sample_norm2 >= kGyroscopeForBiasThreshold), which would exit the function before reaching this point.

This suggests that the max operation is unnecessary here and can be omitted for code readability and efficiency.

[arilow]

Thanks @alainschoebi for your pull request! Changes LGTM.

I ticket has been created to test your changes and merge as soon as we're able to do it internally.

[alainschoebi]

Hi,

Thanks, @arilow. Do you have any updates on this? It appears that the merge has failed.

Best regards,
Alain

[lneumarkt]

Your proposed changes have been already internally merged and will be included in the next release. Thanks for contributing to Cardboard!

[xinyunh0929]

This was merged in v1.26.0.