I got a good explanation for the difference. It’s not resolution, as one might expect: that’s always 1080p. And it’s also not frame rate, which is a constant 10fps. What it is: higher (SD) or lower (HD) compression of images. Pretty much the same thing as compression level on a JPEG image. The more compression, the blockier and more pixilated the image appears. You can see the effect if you save a JPEG from an image editor in both low and high quality. The file will be both smaller and more pixilated in the low quality image.
This pretty much explains what I see from WyzeCam video. If you look at an area of the image that contains a smooth gradient of color and zoom in on that, you can see in the SD mode that there are more distance steps (pixelation) as you look across the gradient. It also explains why sometimes you don’t see much of a difference in the data rate between the two modes… Some images don’t lend themselves to much compression (lack of large areas of the same color/intensity). If the camera is seeing such an image, then it makes sense that the data rate would be closer together in both modes.
Also, if the image contains huge areas of the same color, the overall data rate will be lower in both modes because the image can be better compressed. Try this: cover the camera with a black cloth, so the image is just total black. You will see, I believe, the data rate fall dramatically as the camera takes advantage of the greater compression that is possible.