I must admit I'm not entirely clear what the answer is.
For quite some time I had subscribed to what your referenced article states. However, in practical situations through personal experience with various sensor formats, it is the angle of view that changes perspective. I'm tempted to say irrespective of format but this is where the theory gets a little fuzzy.

Bear with me for a second.
Assuming a subject that is 1m from your camera and is 2m wide. You frame your image to just incorporate the entire subject.
So with various sensor/film formats, you'd need different focal length lenses to achieve the same framing. The smaller the sensor/film format, the wider the focal length.
But you achieve the same perspective even with different focal length lenses regardless of sensor/film format because the angle of view hasn't changed, or has it?
I can only assume the angle of view is a measurement taken from the edge of the scene to a point in the centre of the sensor/film in which case the angle of view stays the same regardless of sensor format. This is possibly a wrong assumption.
But in reality the scene is not captured this way. The edge of the scene will be captured at the edge of the sensor/film whilst the centre of the scene is captured at the centre of the sensor/film. So the angle of view as measured from the edge of the sensor/film to the edge of the scene does change as your sensor/film size increases.

I was always taught to extrapolate to a more extreme scenarios to help illustrate a point.
So lets assume we now compare the same example scene captured with a miniscule digital sensor, something you might find in a phone camera. Same camera and subject position but using different lens of different focal lengths to achieve the same framing to just incorporate the entire subject within the frame.
Compare that to a gigantic camera where the film size is 2m across. Ie. The film is the same size as our subject and the image will be captured at 1:1 on this theoretical giant camera. Its not hard to imagine on this giant camera the light rays are almost parallel at all parts of the scene resulting in theoretically zero perspective distortion.
And in the tiny camera scenario the light rays in the centre will vary greatly with the light rays at the edge.

So what's my conclusion, I'm tempted to say format does make a difference but you'd be hard pressed to see a difference unless the difference in formats are very large. Will there be a difference between FF and APS-C, probably hardly noticeable.
How about FF and 8X10 large format - then absolutely. Between FF and 645, probably not much but likely when you go to 6X7.
You'd probably notice much bigger differences in DOF and less in terms of perspective changes.

Anyways, that's just me applying some layman geometry to try and get my head around the conundrum. Feel free to poke holes in my theory.

In the mean time, try the Brenizer technique if you don't mind stitching in post processing. Its akin to using a much larger sensor and will be ideal on a stationary object but live subjects are also very doable as the man himself demonstrates.