Wtf?

[Steve Schell]

Big shout out to GM on this one. These seem to be classic radial horns, as invented by John Volkmann of RCA. Mr. Volkmann's understudy A.J. May once explained the radial horn to me (sure wish I had tape running) something like this: The radial horn achieves the excellent directivity of the conical (straight sided) horn in one axis, usually the horizontal, while achieving the exponential loading at the sacrifice of directivity in the other axis. In other words, the horizontal pattern has excellent directivity, while in the less important vertical axis directivity is sacrificed and doubly poor... a reasonable compromise for real world performance in a horn of reasonable dimensions.

One strange aspect of most radial horns is that the curved sides become a bit closer in the initial throat flare. This is necessary to maintain the exponential expansion due to the rapid expansion in the other axis due to the straight walls. The cross section of these horns can be visualized as being curved as a radius in the horizontal while being flat in the vertical.

[Steve Schell]

To add just a bit, the calculated horizontal pattern is curved while the vertical pattern is straight. This concept was validated by Altec's adoption by 1952 or so, and almost everyone else's within a few years. The excellent directivity provided by the straight walls was, then as now, very desirable.