From my personal research and development of DSP, I would say that digital can sound better than analog, with good algorithms. If digital "is worse" then it is just a bad algorithm. And unfortunately, there are a lot of people who write code, who do not understand what people are looking for in
a plugin.
I am both a musician and a developer though. So I did a lot of development that takes into account the musical side.
Regarding "tubes". There seems to be a lot of confusion out there, what really "tubes" are, in what context they are used, and how that makes them sound.
First of all, tubes, the popular ones, say 1950s are not at all hifi. There is a lot of dynamics in them, that makes them saturate, and add tonality/harmonics to the audio. And a lot of the time, they are in a configuration (push/pull etc) that gives 3rd order overtone. So overall, one is looking at some dynamics, and some third order distortion. Third-order distortion gives psychoacoustically less intermodulation distortion, so it can be used to saturate tops of guitar, without getting a typically "distorted" guitar sound. I think this is what a lot of people like about it.
Also in vintage limiters, mics, etc, transistors, there is again the third order overtone, that gets louder as signal approaches clipping.
And then there usually is frequency responses involved, in components or speaker elements etc. Atleast highpassing.
So it really just comes down to eq (dynamics if you want) softclip eq.
Analog is just analog math, with analog imprecisions.
Great to see more people who understand this. The analog debate is sometimes going wildly out of proportions, and poor engineering no doubt has contributed to that, on the digital side. Where lesser engineers measure frequency responses and fail to see the simplicity of the design. And make DSP solutions that have much more complex and cpu-wasting code, that don´t really even hit the target.
And there is a tube-design that has a particular more of a second order sound, for instance, old tube radios. I saw an engineer that converted these to small amps, and said they sounded like tubeamps, and in that context, they really sound different than the third-order configurations. But a look at the design, and what it does, should probably reveal a good DSP approximation. It typically comes down to a shaper, and dynamics, and eq, and this time, maybe more of a second order thing. I don´t know what exactly does that, atm, since I have never really been interested in the sound. If anyone wants the sound, I can actually try to implement a mode in my softclipper though, it should be easy.
Peace Be With You.