Inconsistent waveform chart update

[EMCCi]

But with the min/max decimation, there are two decimations per pixel, and two points (a min and a max) for each decimation. So, there are 4 points for each pixel (obviously only one gets displayed, but they are required because LV needs to choose which to plot in order to plot the most significant one, at least this is what I understand after reading the paper and the example code). So, to obtain a correlation between pixel and timestamp, isn't enough.

[thols]

You are right. Actually, the last time I did this, I did not make four points but three. I decimated only if I had more points than three times the pixel-width, and split the data into the number of pixels I had, did a min, max and mean on that data and plotted all those three points for all pixels. That should be correct, but I can't find a reference for that now.

So I would suggest to use NI's suggested way to be safe. What you do is to calculate and plot all four points for each pixel. So actually, all four points gets displayed, and in some cases you will get two or more vertical pixels filled for one horizontal pixel, in order to present the plot "lossless".

[EMCCi]

The problem is that the NI's suggested way is for waveform plots without X axis units, therefore the time when the decimated point has taken place is not important. The only important thing for them is to order the Y min/max pairs in the same way they have taken place, but there is not X value for them. They only work with the Y array.

For an XY plot with X and Y arrays, I have done the following code, that takes as X value, the time when the decimated pairs min/max have taken place. But there is no NI's suggested way here.