Uh... Is that saw's blade supposed to be rotating? If so, it seems that the sparks are going the wrong way.
A demo saw like that would have a blade that rotated top to front to bottom, so that if cutting a vertical object, the rotation of the blade would help support the weight of the saw, and if cutting downwards on an object in front of the user, it would kick away rather than towards the user.
So, sparks are emitted at a tangent to the edge of the blade, in the direction of rotation. Most sparks are emitted at the point of contact with the job, but some are emitted further around the blade. Sparks are *not* emitted normal to the blade.