I own a pair of these speakers which I have modified slightly, but even when they were stock I didn't find them bright. I don't see any brightness in the frequency response measured, either:
https://www.audiosciencereview.com/...ma-frequency-response-measurements-png.65017/
There is a broad excess of energy in the range from about 900 Hz to 2.2 kHz, and a bit of a dip at 3kHz but otherwise these measure as exceptionally well balanced speakers for the price. To my ears, they sound that way as well.
I set out to improve this speaker, which Dennis Murphy pointed me to as being a great value even stock. I was prepared to generate a whole new crossover, but ultimately found that the biggest issue was diffraction, which caused the elevated response in the 1-2kHz range. I addressed that with an anti-diffraction frame, which also lowered the baffle step frequency contributing to better spectral balance.
I measured the speaker drivers and started simulating (they have an interesting series crossover between the tweeter and supertweeter, BTW). I figured I would need a more complicated crossover, but it actually needed a less complicated one. Shorting out one of the resistors in the woofer circuit gave increased attenuation at high frequencies, and an increased Q, which was also needed.
All of this was documented on this site:
https://www.avsforum.com/forum/89-speakers/3068362-improving-sony-ss-cs5.html
I would be happy to send the frames I made (and 2 wires for shorting that resistor) if Amir would be interested in giving his impressions of the mod, along with measurements. I honestly think the on and off-axis measurements of the modified version are exceptional, along with the sound (as long as they aren't pushed too loud.)
I think it would make an interesting experiment to show how the diffraction frame changes dispersion, and possibly perception.