This is just my opinion and I'm no means the final authority.... so take my opinions and verify them with your own experience.
With compressors, the sound characteristics doesn't come from whether it uses tubes or not. You see, the tube is mainly used as a gain stage. After compressing the signal, you have reduced level, so you need some amplifying device to boost the levels up again before coming out of the equipment. This is where vacuum tubes and solid states come in.
Granted, they can *color* the sound somewhat... but coloration has little to do with whether there are tubes or not... it's got more to do with audio transformers in the signal path. With modern solid state design, they've eliminated the audio transformers for COST reason. In my DIY, transformers add about $300 minimum in parts alone to a stereo channel. Add the cost of manual labor to solder the transformer wires (because you can't automate these process) and prices go even higher.
With TRUE vacuum tube designs, because of the high voltage present, they need some way of isolating the high voltage DC from the audio signal output, and so the need for audio transformers is a MUST... not an option. Transformers also provide impedance matching and signal isolation.
Note there are solid-state design that still use audio transformers, (example: SSL, API, Neve... ever heard of those names?) and they are excellent designs. Of course, excellent prices too.
With compressors, the more important question is the method they use to detect, and achieve their compression. Some use optical methods to detect loud material, some use solid-state method. And then, that is further subdivided whether they used JFET, or transistors, or ICs to reduce the level of the material.
Optical compression have a soft-knee response. With optical compression, you have a light that glows brightly in sync with the program material, which affects an LDR (Light Dependent Resistor) that varies the signal coming out. The famous and most revered Urei LA-2A uses optical compression.. i.e. the JBL T4B optical unit alone costs about $150. An LA-2A costs about $2300 per channel. This unit produces sound that is nice and fat... almost every pro studio have one or more of these units!
BTW, the LA-2A uses audio transformers, vacuum tubes, and optical compression! Yay! ... no wonder it sounds great. And yes it is vintage design, if you open it up, you'll just see wires criss crossing around. No PCB.
The other way of doing compression is using an IC chip (commonly called a VCA). It's usually an all-in-one unit chip that does everything. For example, THAT corp makes these chips that are used by our most loved RNC compressor. Presonus ACP-8 ($900 for 8 channels) also uses the same IC chips from THAT Corp. The RNC is pretty transparent using it's Super Nice mode. $180 for stereo channel. These units use a VCA (Voltage Controlled Amplifier) to reduce levels of the signal. The IC chips are cheap... about $2.55 per piece. HINT: if you want to build your own, using the datasheet from THAT Corp, you could be saving a lot of money.
There is also the JFET compression used by the famous 1176 compressor. The JFET in this case is used as a variable resistor to vary the gain of the device. I've never heard of an 1176 (I'm currently building one), but most pro studios have them as well in their arsenal. Cost: $1800.
All 3 methods will produce different characteristic compression. You can't say X method is better than Y... They're all good, just depends on what will sound better on the material and track you're compressing and the effect/sound character you want to achieve.