In the next public release of the GUI, "Gain Convention" will be set to "Body uV" by default for Cyton boards. This lowers the uVrms values in the GUI. I recommend using "Body uV" instead of "Classic".
Richard, how does the Gain Convention affect recordings? Does this new setting mean that old recordings have incorrect uV values? Is there a link explaining the old and new defaults for Gain, and what they mean in terms of signal processing?
Maybe I misunderstood, but I thought that one of the previous issues was that if the scale factor was changed on a channel, there was some misinterpretation of new uVs, since scale was not taken into account. But the default scale is x24, and I thought the gain conversion was correct for that scale factor. A documentation reference would be helpful.
The old GUI just used '1' for the gain, even though it was known that the default ADS1299 channel settings gain was x24??
Or was this something that happened when we transitioned to Brainflow access vs direct serial port parsing in the GUI? Recall that Joel Murphy ran calibration runs using the GUI, with known microvolt external voltage sources. And those calibrations were correct, otherwise something would have been fixed.
Or perhaps the misinterpretation crept in, when AJ split the Cyton stream parsing out to the Hub... In any event, it MUST have been correct in the GUI when Joel did his external voltage tests. This was WAY prior to Hub days.
classic means that in equation to convert int24 to uV ADS gain is not considered(in fact its one in this formula), so it doesnt show real voltage
So 'classic' just means the scaling (gain) formula interpretation that began either with AJ's Hub, or the Brainflow based GUI, correct? The old old GUI not using the Hub, was doing it correctly for default EEG channels at x24 scale.
Richard @retiutut and Andrey, I split this discussion on Gain Conv, into a new thread for clarity. My impression is that this menu should disappear, and Body uV should be the default. I don't think many users are wedded to using incorrect channel microvolts.
It looks like we will keep this button for now, as it was explained that some legacy users may find value in "Classic" gain mode. We will add a description to this button that describes both gain modes.
If it is indeed '1' as earlier stated, does not that mean that for the default EEG case, the uV shown are 24 times what they should be? Instead of 10 uV (a reasonable value), 240 uV would be shown? That is misleading. Are you sure the 'gain' in the equation is not actually 24, as it was for the old old GUI?
Starting with GUI v5.0.2, all ADS1299 boards (Cyton, Cyton+Daisy) will use a dynamic scaler. This ensures that the scale factor, set in the Hardware Settings view, is always in sync with the numbers shown on user interface displays and recordings.
Comments
In the next public release of the GUI, "Gain Convention" will be set to "Body uV" by default for Cyton boards. This lowers the uVrms values in the GUI. I recommend using "Body uV" instead of "Classic".
Richard, how does the Gain Convention affect recordings? Does this new setting mean that old recordings have incorrect uV values? Is there a link explaining the old and new defaults for Gain, and what they mean in terms of signal processing?
Maybe I misunderstood, but I thought that one of the previous issues was that if the scale factor was changed on a channel, there was some misinterpretation of new uVs, since scale was not taken into account. But the default scale is x24, and I thought the gain conversion was correct for that scale factor. A documentation reference would be helpful.
classic means that in equation to convert int24 to uV ADS gain is not considered(in fact its one in this formula), so it doesnt show real voltage
body means that correct gain from ads was used to calc uV value
Its not only visual, it also affects values in file
I voted to dont add two options and keep only "body" cause it shows real uV value....
Andrey, thanks. So this page was wrong, as far as the actual conversion the GUI used to do:
https://docs.openbci.com/docs/02Cyton/CytonDataFormat#interpreting-the-eeg-data
The old GUI just used '1' for the gain, even though it was known that the default ADS1299 channel settings gain was x24?? Gees, that is embarrassing.
@Andrey1994 Thank you for clarifying! I agree that maybe we should keep "Body uV" and remove the button.
It would be helpful to see if any users still see value in keeping "Classic" mode.
@Andrey194, Could I get confirmation of this:
Or was this something that happened when we transitioned to Brainflow access vs direct serial port parsing in the GUI? Recall that Joel Murphy ran calibration runs using the GUI, with known microvolt external voltage sources. And those calibrations were correct, otherwise something would have been fixed.
Or perhaps the misinterpretation crept in, when AJ split the Cyton stream parsing out to the Hub... In any event, it MUST have been correct in the GUI when Joel did his external voltage tests. This was WAY prior to Hub days.
In old gui 24 was hardcoded inside this equation, it was correct as long as you dont change gain since 24 is a default for cyton
@Andrey1994, further clarification. You wrote:
So 'classic' just means the scaling (gain) formula interpretation that began either with AJ's Hub, or the Brainflow based GUI, correct? The old old GUI not using the Hub, was doing it correctly for default EEG channels at x24 scale.
And 'classic' means that in this formula,
https://docs.openbci.com/docs/02Cyton/CytonDataFormat#interpreting-the-eeg-data
'gain' is set to '1'.
Richard @retiutut and Andrey, I split this discussion on Gain Conv, into a new thread for clarity. My impression is that this menu should disappear, and Body uV should be the default. I don't think many users are wedded to using incorrect channel microvolts.
William
This change has just been made. Here is the Pull Request: https://github.com/OpenBCI/OpenBCI_GUI/pull/931
It looks like we will keep this button for now, as it was explained that some legacy users may find value in "Classic" gain mode. We will add a description to this button that describes both gain modes.
Richard or Andrey, can you clarify? With Classic gain mode, is 'gain' in this equation set to '1' or '24'?
https://docs.openbci.com/docs/02Cyton/CytonDataFormat#interpreting-the-eeg-data
If it is indeed '1' as earlier stated, does not that mean that for the default EEG case, the uV shown are 24 times what they should be? Instead of 10 uV (a reasonable value), 240 uV would be shown? That is misleading. Are you sure the 'gain' in the equation is not actually 24, as it was for the old old GUI?
Regards, William
After further internal discussion, the team has decided to remove the button and communicate this change in the docs.
Here is the GitHub Issue to clarify that a dynamic scaler will be used from GUI v5.0.2 onward:
https://github.com/OpenBCI/Documentation/issues/45
@wjcroft I think this has been resolved.