Interesting info. So looks like unfortunately the best colors are sold by specialty hobby companies (which means those LEDs are probably way overpriced). Personally I have graduated to using SMD LEDs in most of my models, so this info is of limited use to me.
I also question you powering all the LEDs in parallel. These are all different LEDs, different brands and types. So, they don't all have the same electrical specs (like Vf). That means that each LED is most likely conducting different current through it (resulting in some being dimmer and some brighter). That also affects the LED'sl color balance. I think the test would have been better if each LED was powered through its own current limiting resistor, with those resistors chosen to send the same current through each LED. That way, you could have for example 10mA passing through each LED.