I like the response @uarte made on this, and I’d echo some of what he said with respect to changes over the years. But here are my thoughts from my personal perspective.
First, with respect to “signature sound”, I am primarily a songwriter, singer, and piano player. To the degree I have a signature sound, the key common denominator will be my voice, which can be one of those “love it or hate it” type things, with a secondary aspect being my piano-playing style, which tends to be influenced by Elton John and various gospel-flavored styles. My songwriting can actually be all over the place stylistically, sometimes even within a single album (and I’m up to six full albums of original songs plus another full album of Christmas carol covers and somewhere north of 50 singles).
I’ve also been at this a long time, starting a Tascam Model 38 reel-to-reel (in 1984, if I remember correctly), then going to that along with a Korg SQD-1 (if I’m remembering the model number correctly), then Texture running on DOS, then Passport Designs MasterTracks Pro, alongside an ADAT for the audio side, then Cakewalk Pro Audio (around 2000) and SONAR, and now Cubase (initially starting with 9.5, but only making it my main DAW at 10.5). (A few others briefly in between, too.)
At one point when I was just MTPro as a MIDI sequencer and all keyboards and MIDI modules with the ADAT, I had a fully-populated Mackie 32*8 mixer – on channel vocals, and the rest audio outputs from MIDI gear. I’ve always wanted to get a full band sound, sometimes even more than just a typical band (e.g. orchestral elements) from whatever I had. So, for example, in the hardware days, I might have a dedicated drum module (initially TR-808 then a few different Alesis MIDI modules), a dedicated piano (Roland MKS-20 or MK-88), some modules focusing on synth sounds, plus some that were meant for more imitative sounds (e.g at one one point I was doing “guitars” with the Roland U-110 and/or U-220).
But I started switching to software-based instruments around the time Nemesys GigaSampler came along. It was the first piano that worked better for me than the MKS-20. I’ve gone through a few generations on virtual drums, but I mostly use Superior Drummer and/or EZDrummer these days, with loads of SDX and EZX expansions to suit any style I might be working with at the moment. On bass, I’ve used a bunch over the years, including Spectrasonics Trilogy then Trilian, but I mostly use IK’s MODO BASS 2 these days. Pianos are another area where I’ve gone through loads of possibilities as products improved over the years. While GigaSampler was my start on virtual pianos, I went through various NI pianos (starting with their Akoustic Piano – these days I also use all of them from the KOMPLETE Standard package plus some others for whatever is right for the project), Galaxy Steinway, the Garritan Steinway, Arturia’s Piano V, and, in the last year or so I’ve started using IK’s PianoVerse MAX as my main go to for acoustic pianos (I especially like their NY Grand, which happens to be a Steinway, too). These instruments just keep getting better as time goes on, though my main acoustic piano is almost always one that emulates a Steinway, with any others I use being primarily for flavors for individual projects.
Then there are guitars, electric and acoustic, strummed, picked, and otherwise played. This has probably been one of my biggest areas of adding pieces over the years, with some of those I use quite a lot today including those from Acoustic Samples, MusicLab, NI, and UJAM. And on the amp simulation front, I’ve also got lots of options, with the ones I use the most at present being NI’s Guitar Rig and the relatively new UADx guitar amps (I especially like Woodrow, which I’ve used on my latest recordings). But I do a lot of auditioning of options when I’m “trying to find the sound”, with the idea that I’ll know it when I hear it. Not being a guitar player myself, it’s not like I have personal experience in setting up amps and pedalboards to get a signature sound. I’m just looking for what will sound best in the context of any specific recording. My latest album mostly leans toward country rock, with a few other flavors mixed in, so it is fairly heavily (virtual) guitar-oriented, almost always with a core of drums, bass, piano, guitars, and either organ synth or both.
As for synths, I really like having lots of flavors available to me. I don’t actually use synths all that heavily in most of my mainstream work, but I’ll often use them as flavors, and playing around with different flavors (I’ve got the entire Arturia V Collection along with some synths from AAS an NI) can be inspiring. And, occasionally, I do more synth-oriented recordings, especially for creative cover songs (e.g. my sort of trailer-style covers of ColdPlay’s “Fix You” and CCR’s “Bad Moon Rising”). But they can also be textures in some of my more mainstream recordings, such as for power ballads, too.
Then there are orchestral bits, brass sections, and ethnic flavors…
As for the audio processors and effects, I’ve also gone through phases on this stuff, and it tends to be about workflow at any given time. My methods have changed over time, hopefully in the services of better results. For example. last year I did a remix of my 2008 Christmas EP, adding a few additional tracks, partly because I really loved the songs from that EP but was cringing at my production quality from 26 years back (both the original version of That Time of Year and the revamp, That Time of Year Rewrapped are on the various streaming sites should anyone be curious to compare). At that time of that original, I probably had some DSP-FX plugins, whatever was in SONAR at the time, plus a few other third-party plugins. I think I had the Waves Renaissance pack of the time, probably PSP MixPack, and Autotune, but possibly not much more. I know on the remixes I heavily used both Waves plugins, some specific UADx plugins, as well as some PSP plugins, and I did a lot of experimenting to try to optimize the sound I was getting on these mixes. It’s not like you’d see the results as being a different recording artist (no less a different songwriter) from the original EP, but the results were a significant improvement on the audio quality front.
Keep in mind, too, when you talk about the Waves Diamond bundle (which I have), they call it 86 plugins, but it may well add up to three times that number due to the stereo, mono, and mono-stereo variations. If all those plugins had those variants, that would already put you a quarter of the way to 1,000 plugins, and that’s not even counting some of the other variations such as Doubler 2 and Doubler 4. Arturia V Collection has 45 instruments. Many of the MusicLab virtual guitars have multiple versions, too, such as RealStrat Elite Standard and Elite, both with normal, 2 mono , and 2 stereo variants, for a total of 6 plugins for just that one instrument.
Then there are the ones that come with some other product that I either haven’t used to date (e.g. Groove Agent SE from Cubase), or maybe have only used on some unusual projects (e.g. NI Battery and Reaktor, both of which came with KOMPLETE).
Ultimately, at any given time, I tend to have a few go to plugins for each category I use frequently, but, even there, for something like guitars, it might be one Tele, one Les Paul, a strummed acoustic (or a couple of them for different flavors), a picked acoustic, etc. And I’m sure you know from your guitar player friends that one type of guitar is never enough. Nor is two or three or … 