Do you know if the same difference is also there if you move the Stereo Tool lowpass filter up to 16.4 kHz (the maximum)? Because that's also the default setting in the Omnia.9sg etc - so if this affects RDS reception we might need to change that value everywhere. It also could be that RdsSpy is affected by this but actual receivers are not - more testing needed I guess.
Here are a couple of non RDS outputs for comparison (will see if I can run RDS test again here also at 16400 and post if you like).
Ran non RDS test to show the composite signal spiking into the RDS space versus how it looks direct from ST.
Realise there is a trade-off for the IP carriage and codec - just wondered from the original post whether something like a notch filter could be used to guard the pilot or rds...
ST MPX 16400 LP no RDS.
microMPX 16400 LP no RDS
MPXTool sees more Pilot modulation at 16400 compared to running LP at say 14000/14500/15000.
Best results (no modulation over a 10min sample) was 14000 or less.
Ok. We'll look into this when Mathijs gets back from his holiday. The signal shouldn't spread out beyond the 16400 whatever you send to it - something isn't doing exactly what it should right now. The levels are pretty low though so I wouldn't expect them to actually influence anything (having audio close to the RDS area might though).
There have been processors on the market that had a signal that was way louder than what you're seeing here that went on even beyond 70 kHz.....
Edit: In the display in MicroMPX itself - which goes down to -90 dB - I don't see this at all.
Edit #2: Occasionally a small spike on the right of the L+R area. It seems to peak at around -85 dB. While I agree there should be nothing at all there (and we will look into it - it can't be that hard to fully get rid of it) I can't imagine that it affects anything in the real world. Many popular processors that are used everywhere have less than 70 dB pilot protection - and that's constantly, not sometimes an occasional spike. Both measured with Stereo Tool.