After listening to the 9H and Siberia Elite last year on USB, I was expecting RAW Prism to feature a polite volume sound; something that wasn't too loud. I was wrong. RAW Prism is loud, really loud. As for the signature, I'm somewhat surprised. Let me explain: I had these expectations, of a loud & muddy bass with a V-Shaped sound. What I got was a flat, cold sound, with average imaging and detail. The soundstage, unfortunately, is tiny.
You can customize it (a little); SteelSeries gives you the chance to sculpt the audio with a 5-band equalizer (via the SteelSeries Engine), which does a good job of clearing some of my issues with the overall signature.
The Engine also allows you to tweak the lighting options circling the ear-cups, and enable (or disable) microphone optimization (more on that in a second).
The headset delivers a proper punch, with a surprisingly tight and fast bass. There's not much sub-bass rumble (which takes away from its body), but the mid-bass has a pleasant impact and manages to keep up even with some of my fastest tracks like Hajnal or Brianstorm.
The bass transition nicely into a a relatively clear, but dry midrange. It's got a nice level of detail, with a fairly decent timbre and body.
This would have been a far better sound if the treble was more laid-back. As is, it's agressive, hard and at times uncomfortable. It becomes fatiguing after several hours of listening. The cymbals are splashy, and there's a nice dose of sparkle here, adding some much-needed clarity and detail.
What Siberia RAW Prism noticeably lacks is a good soundstage. There's very little depth and width, less than G230 and on-par to the P11. Separation and imaging are better, though not by much.
If you use the headset without tweaking the signature at all, Siberia RAW Prism will not sound all that impressive. The soundstage is tiny, while separation and imaging are decent (at best), which means that you won't be able to use the headset to get a better sense of space and positioning.
When you turn the SteelSeries Engine 3, the situation changes somewhat. The headset still sounds good after the EQ has been played with, and you even have pre-made profiles (like MMO, FPS, Voice) that sound good.
The left ear-cup houses a small plastic nub that also serves as a omni-directional microphone. Because of its positioning, there's noticeable sound-bleed from the drivers, which means everyone around on the other end will hear themselves through the headset. Not only that, but due to the mic's nature, it also picks up surrounding noise, quite clearly I might add. From breathing, to keyboard typing, cars outside, everything was picked up and accurately reproduced. I also use a mechanical keyboard for typing and gaming, a Cherry-MX Brown model, and the constant clatter made the whole experience unbearable. If I would've been playing a public-game, any public game, I would've driven my teammates on the verge of insanity.
Here's where SteelSeries worked their magic. The Engine comes with a Mic Auto Optimization setting, which filters a lot of the surrounding noise. It takes care of the bass bleed, and dampens some of the surrounding commotion. The trade-off is that it makes it sound like I took a few steps away from the microphone. Funnily enough, neither the keyboard sound, nor my smooth breathing were dulled too much by the adjustment.