I have no doubt that the diodes change the overload characteristics of the SVT in a way that people find desirable. That doesn't mean distortion was the original intent of the circuit. I believe some of the reason for abandoning the 6146 based output section has to do with screen related failure when the amp was pushed into hard overload. I believe the screens in 6550 based amps are able to take more abuse. The diodes could be a hold over that was originally designed as protection for the 6146, but was retained because people like the way the 6550 output section overloads with the diodes in place. Or the diodes may still be intended primarily for protection. To be clear, an intent of the original designer was to make a robust amp that would withstand the rigors of professional (ab)use.
From what I have seen most users don't play their SVTs at maximum output levels, so the diodes should not typically have significant impact on the sound. It is easy enough to take some oscilloscope readings to determine the output power levels at which the clipper circuit becomes active. Until readings are taken, it's all supposition.
The Trace Elliot V4, V6, and V8 all have clipper circuits after their master volume control. The V4 and V6 are known primarily for clean power and have relatively low gain preamps. The V8 is also known for clean power, but it has a high gain circuit (V2) that can be switched into the preamp for OD. The clipper in the V8 is between V5a and V5B. The master volume controls the gain of V5a. The amp is setup to produce high gain OD in the preamp that is independent of master volume settings and clipper activation. If distortion was the intent of the clipper it should be before the master volume. Consider V2 in Marshall JCM 800 and JMP amps introduces asymmetric clipping into the signal well before the output section begins to clip.
I have seen posts from multiple techs, who suggest the drivers in V-Types provide insufficient signal to push the output section to full potential. This could be related to the clipper circuits, however all models produce lots of clean power with no aural indication of a functioning clipper, and these amps begin to compress long before any evidence of OD becomes perceivable. Additionally, all of the bench tests I have seen where power was measured showed the amps exceeding their stated power specs.