
First of all, the drawn current from ULA pin 16. I measured this compared to another TS1000 that still has the pin 16 connected to the RF modulator. In fact, the RF modulator draws a HIGHER current from pin 16 than my 3 diode solution. So although your circuit would be better, my circuit does not load the ULA more than the original RF modulator.
Second thing is that my simple trick works very well for a cheap chinese portable DVD player but didn't work for a Sony Bravia TV. For this reason I had to modify the circuit a bit, leaving 1 diode out and replacing it with a transistor. This also buffers the signal, so there is no longer a question about ULA pin load.
Third thing is the interference, or noise. This is still present in the signal but not amplified so it is no more or less than it was before. You could clip the voltage on the base of the transistor to get rid of that, but for me it doesn't really matter so I haven't tried that.
I got some picture here with the final result on my TV.