You're right, context is important. I didn't mean to sound like a 240 vs 220 pedant.
I sometimes call it "undisturbed" or "reasonably uniform".
Well, at dust collection velocities you need more than 10D to get fully developed flow. But 5D to 10D from a disturbance is usually good enough for both testing and not too much system effect (interaction between closely spaced components).
There's a difference between what's agreed upon for reproducible testing and what's required for actual operation. Per AMCA a 10D test pipe is used for testing DC type fans. That convention has been extended to the testing of cyclone-fronted hobbyist dust collector systems. That means you should be able to reproduce a manufacturer's test data if you're careful. It doesn't mean the system quits working if you only have 1D, just that you get different (usually not as good) results.
So, a 90 close to the input of a cyclone will almost certainly increase the loss coefficients of both the 90 and the cyclone (system effect). It will also change the collection efficiency vs particle size curve, probably spreading the transition region somewhat. I think the overall impact of that will be a lot less than a reduction in airflow or a collection bin leak. That's an opinion, I've not seen any actual data on what inlet disturbances do to collection efficiency.