Why the phase noise of a U-Blox GPS module at 10 MHz is terrible

Why the phase noise of a U-Blox GPS module at 10 MHz is terrible.

There was a YouTube video showing the poor phase noise of a U-Blox GNSS module programmable pulse output at 10 MHz but cleaner at 8 MHz. It suggested that it was because the module does an integer division of its disciplined clock at 24 MHz. So the 8 MHz is clean and the 10 MHz is not.

https://www.youtube.com/watch?v=M6SmdqMA-Dg&t=1611s "Using a GPS module as variable frequency reference standard"

In a U-Blox application note, I found that the presenter is correct in that it uses an integer division, but the clock is 48 MHz. The non-integer pulses seem to be made up of two pulses with a delay between them to give the correct frequency pulse but with many artifacts and consequent phase noise, as shown in the image from the paper.

https://content.u-blox.com/sites/default/files/products/documents/Timing_AppNote_%28GPS.G6-X-11007%29.pdf  "GPS-based Timing Considerations with  u-blox 6 GPS receivers"