Your argument isn’t logical. Recording and playback are two different things. Reference playback should impart near zero influence on the original recording. It should be a benchmark for accuracy and transparency. Even the best sounding tube gear is inherently incapable of that, especially when asked to drive the highest performance of speaker drivers. Artists and Recording engineers employ tubes to achieve a certain characteristic sound. Adding noise and distortion during playback by use of tubes alters that sound (for better or worse) from what the engineers/artists intended.
Good subjective performance does not automatically make the component(s) accurate. For example, I enjoy the sound of Magnepan speakers (some of them anyhow) but they don’t come remotely close to playing the truth of a recording. They are enjoyable but I don’t fool myself into believing they are reference (i.e. benchmark) level.
Preferring tubes in audio playback is akin to preferring the driving experience of an 80s air-cooled Porsche over a modern 911. That doesn’t make the older car the better performer.
There’s nothing wrong with preferring a “colored” system, but let’s not pretend that makes it better than state-of-the-art, nor a reference by which all others should be judged.