You are here

Celestion F12–X200

Full–range Guitar Speaker
By Dave Lockwood

Celestion F12–X200 speaker.

Struggling to make amp modelling work on stage? Something just 'not quite right' in the feel? Maybe Celestion's new speaker will make all the difference.

Celestion's F12‑X200, announced over a year ago but only recently shipping in volume, is a 12-inch speaker with a built-in high-frequency driver and passive crossover, designed to be used with guitar amp and speaker-modelling systems. The coaxial HF compression driver gives the extended high-frequency response necessary to work with a speaker-emulated signal, using the cone as a horn flare for dispersion. The cone driver itself exhibits the characteristics of a normal guitar speaker, with a ceramic magnet, two-inch voice coil and a straight-sided, lightweight paper cone. The concept here is to make it as easy as possible for a guitarist to swap out the F12‑X200 into a combo or cabinet that they might already have, and to offer as much as possible of the 'feel' of a real guitar speaker within a modelling-based rig.

Celestion F12–X200Even as a lifelong tube-amp devotee, I have to admit that the best digital guitar-amp modellers from class-leading companies like Fractal, Kemper, Atomic and many more are now capable of creating a very authentic and pleasing playing experience in the studio. So, if we can record quite happily listening to digital speaker emulation via full-range/flat-response (FRFR) studio monitors, then surely we should be equally happy on stage, monitoring via larger, high-powered, full-range speakers?

Indeed, many people are now entirely happy with that experience and have converted their live rigs to FRFR, or disabled their modelling or impulse-response based speaker stage in order to use conventional speakers with modelled amps. Others, however, including myself, continue to find this type of playing experience rather less satisfactory, and there are some reasons, beyond just unfamiliarity, why that should be so.

The first of these is one of expectation. When listening to a modeller on studio monitors whilst recording, I am replicating a situation in which my speaker is isolated in another room and I am listening solely to its miked-up sound. That is already very different to listening to your amp 'in the room' and, in itself, takes a bit of getting used to. However, I don't think I've recorded a guitar part in any other way in the last 20 years, so clearly it is something one can get very comfortable with.

The F12‑X200 features a typical 'guitar-speaker' cone.The F12‑X200 features a typical 'guitar-speaker' cone.In contrast, most players using modelling on stage are looking to replicate a situation in which much of what they are hearing of their own instrument will be coming directly from a guitar speaker somewhere behind them. Of course, if you are on in-ears or huge stages where you can only really hear wedges in front of you, your experience will be very different, but most of us trying amp/speaker modelling on stage want to feel like we are using an actual guitar amp and speaker, and PA-type speakers don't really do that very well.

Recent years have seen the launch of a number of nominally FRFR, dedicated guitar-amplification systems specifically for use with modellers, some of them from the makers of modelling systems themselves. Tech 21 were perhaps something of a pioneer in this field, with their (recently updated) Power Engine combo. Atomic's CLR products have long been considered amongst the best, whilst Line 6's PowerCab range goes a stage further into emulation and includes powerful DSP to facilitate different virtual configurations of speaker and enclosure. Some of these products use extended-range, high-headroom guitar speakers, whilst others are really just repackaged, two-way, powered PA speakers, but the best of them can certainly offer workable solutions, and bring with them all the benefits of portability and consistency that entice many players down this path. I can't say that I've tried them all — I doubt that anybody has — but I also can't say that any of the ones I have used have made me entirely 'comfortable' on stage with a modeller. That was until I got my hands on one of these...

Test Setup

The chassis and magnet depth compared to the Vintage 30, on the left.The chassis and magnet depth compared to the Vintage 30, on the left.I installed the F12 first in a closed-back 1x12, and powered it with a hefty rackmount PA power amp, using a diminutive Atomic AmpliFirebox pedal as a source. If I have to use a modeller for something where I can't use a real amp, this is the one I will most often choose, so I am very familiar with what it can do. I put a compact mixer in between the Atomic and the power amp to serve as a more convenient level control and as a post-modelling EQ if needed.

Power handling is stated as 200W RMS, which should make it OK with the kind of big, solid-state power amps users often like to employ in modelling rigs, whilst sensitivity, at 96dB (1W at 1M), is roughly equivalent to many conventional guitar speakers, so it will also be loud enough for use with any reasonably sized combo's power stage, feeding a modeller into the effects return input. I don't actually have any guitar combos with solid-state power stages, so I opted for the next best thing, powering the F12 in an open-back combo cab with the same big PA amp as used for the sealed cab.

The extra depth compared to a conventional speaker didn't prove to be a problem mounting in a compact tube combo.The extra depth compared to a conventional speaker didn't prove to be a problem mounting in a compact tube combo.

Finally, just on a 'what if' basis, I also installed it in an old Mesa Mark II combo, and patched the modeller into the effects return input, thereby using just the tube–amp power stage to drive the speaker. Most of the distortion happens in the preamp stage in those amps, and the four 6L6s tend to be really just ticking over at any tolerable volume, so I thought it might be interesting to try.

The nominal frequency response is 60Hz to 20kHz, with a crossover frequency of 3kHz. As no specific enclosure is specified for this speaker, the bottom-end response will obviously vary according to whether you've put it in a sealed or open-back box. 60Hz doesn't look very 'full-range' on paper, compared to a typical PA-type speaker, and indeed it isn't, but the first thing most of us have to do when adapting favourite modelled amp tones for live use is shave off a chunk of low end to stop the bass becoming 'woofy' and overbearing. One of the most impressive things about this speaker in action, to me, was how little I found I had to adapt presets developed for recording in order to use them with the F12‑X200 in a live context.

Comfortable & Familiar

Firing up the sealed-cab rig, the sound was surprisingly comfortable and familiar, and running through a range of presets from clean to '11' left me in no doubt that this speaker does exactly what Celestion intended it to do — it gives you the sound and the 'feel' of a real guitar speaker with a modelled guitar-amp-and-speaker source signal. The compression driver has none of the hardness that I associate with having to listen to guitars through traditional wedge monitors, and in no time at all you forget it is there and just enjoy the sound of your modeller... assuming that you do. And that is an important caveat — if you don't like the sound at source, you still won't like it through one of these!

I don't use sealed cabs much these days, and rolled off a little of the bottom end just for personal preference, but apart from that, I didn't feel like I needed to touch the EQ on the mixer across a whole range of sounds. Seriously impressive! Questioning my own enthusiasm somewhat, I switched the source to one of my less favoured modellers. Brittle, harsh, stiff — yes, it was still all of those things through the F12. To be fair, it was improved just a little, but it wasn't totally transformed, and I wouldn't want it to be. What modelling users will look for in this speaker, I think, is sufficient fidelity to accurately reproduce their carefully dialled-in modelled sounds, adding only the feel and dispersion of a guitar speaker, and my experiences during testing suggest that's exactly what the F12‑X200 does.

Running through a range of presets from clean to '11' left me in no doubt that this speaker does exactly what Celestion intended it to do — it gives you the sound and the 'feel' of a real guitar speaker with a modelled guitar-amp-and-speaker source signal.

The open-back combo box was a bit more 'room-filling' at the expense of some of the bottom end — no need for any post-modelling low-end roll-off. The rear radiation from the F12 seems pretty much like any normal guitar speaker, which makes sense when you think that precisely the same areas of cone are free to radiate rearwards. I was happiest with the sound when using small–cab emulations with the open–back cab, and larger (2x12 and 4x12) emulations when using the sealed box. This seems fairly logical, but it is hard to be definitive on something so subjective as 'feel', and maybe there's always some influence from the expectation of what a particular rig 'should' sound like. The only 'semi-official' enclosure recommendation for the F12 I've seen is for a ported, sealed cab design somewhat in the manner of a Mesa Thiele cab, which would doubtless give you the most 'thump' from a single–driver setup. There is an argument to be raised that the speaker/cabinet for modelling should always be as neutral as possible to work reliably across a range of different cab sims, but we put real guitar speakers in different cabinets just based on personal preference, and the F12‑X200 seems to happily accommodate the same flexibility.

Guitar cabs and their speakers work as a complete system — we only ever hear them in combination — so an element of the 'cabinet sound' must be incorporated within any good modelled or IR‑based cab sim. In theory, that shouldn't then be replayed through another actual guitar cab, applying the 'cab coloration' twice, but I can't say that I managed to hear anything detrimental going on. My most 'characterful' combo cab sounded precisely as honky and coloured with the F12 as it does with its normal speaker, but not more so.

The Final Piece Of The Puzzle?

If you are someone who uses an amp/speaker modelling system on stage with FRFR amplification but still secretly yearns for the familiar feel of your old analogue rig, Celestion's F12‑X200 may well be the final piece of the puzzle that you are missing. To my ears (and feel under the fingers) it does exactly what the makers designed it to do, and I would have no reservations about using one in critical applications. It seems to 'push air' like a real speaker, giving you the same sense of power and projection on stage, and if, like me, you associate compression drivers with an edgy, uncomfortable treble, you'll be pleasantly surprised that the F12 exhibits nothing of the kind. The off-axis volume drop-off closely mimics a guitar speaker, so you can hear yourself well, standing near it, without everyone else on stage feeling you are too loud. You can also easily generate feedback — the good kind, where the speaker just puts energy back into the strings — with the F12: something that has always been problematical with conventional FRFR speakers, in my experience.

It could be argued that the F12‑X200 is just the ultimate in compromise solutions, offering neither the theoretical accuracy of true FRFR, nor the integrity of a real guitar speaker. In action, it proves itself to be nothing of the kind, giving a playing experience, with a good modeller, just like a 'real' guitar speaker. It doesn't reproduce mixed programme content (backing tracks and so on) all that well — certainly compared to a 'real' FRFR speaker — exhibiting a little too much of the coloration and resonances of its enclosure, but I believe that is all part of what makes it work so well for guitar signals.

Compared to the simplicity of an integrated, self-powered system, it does leave you with the question of what to power it with. If you are putting one in a combo, you'll have to bear in mind that you will only be able to use that combo with a modelling source — unlike some two-way systems, the level of the F12's HF driver is not user-variable, and believe me, you don't want to listen to a normal guitar amp through a full-range speaker! With a modeller providing speaker emulation and the F12 also adding a touch of its own character, it would seem sensible to go as 'neutral' as possible with the power amp stage, hence my choice of a big PA amp for testing, but I did also try one of the new generation of compact, pedalboard power amps. A full-on amp emulation pedal is best avoided for this job, but something that is only giving you a little bit of 'character' in a power amp will enable you to keep your options open for using different tones from your modeller. A powered Kemper or one of the BIAS heads would also seem like a natural companion for the F12. The tube power stage in the Mark–series combo didn't work half as well as the big, solid–state power amp driving the speaker in the same enclosure, perhaps because the return input signal is injected at a stage that is slightly 'voiced' rather than neutral. Other tube amps may be different, but combo users might fare better with solid-state amps, which will come with a weight benefit, too.

Guitar tone is as much art as science, and ultimately all that matters is what works for you as a player, which may well be very different to what works for me, but for those seeking to make the most of a modelling rig in a live context, I've found that Celestion's F12‑X200 can be a real game-changer.


As I said in the main text, there are plenty of compact, powered PA cabs to choose from, a number of dedicated 'guitar modelling-amp' solutions, and some high-bandwidth, extended-response chassis speakers, but nothing else offers quite the same combination of attributes as Celestion's F12‑X200.

Why Make A Full–range Speaker?

Celestion F12‑X200 frequency response plot.Celestion F12‑X200 frequency response plot.The top-end response of a conventional guitar speaker will start to decline significantly above 5kHz or so, but there is still plenty happening above that point, just not at the same level as below. If you've ever applied a steep low-pass filter at 5kHz to a recorded guitar signal, you soon realise how important the low-level, high-frequency components are to an interesting guitar sound! It is this characteristic response that modelled speakers and speaker impulse responses (IRs) replicate. Reproducing this response through a conventional guitar speaker would result in, effectively, applying the response curve twice, diminishing the upper frequency components to inaudibility — hence the (to some) slightly counterintuitive need for a full-range system to reproduce a limited bandwidth source.

Far & Wide Or Near & Narrow?

Apart from their frequency response, another big difference between guitar speakers and most full-range, multi-driver speaker systems is their dispersion characteristics. PA speakers are usually designed to maximise their dispersion in the horizontal axis, so that their tonality isn't skewed too much for audience members who aren't directly in front of them.

A conventional guitar speaker, on the other hand, exhibits a significant high-frequency drop-off by the time you are 45 degrees or so off axis, in both the horizontal and vertical domains. In theory, the dispersion of a PA-style FRFR ought to be better option, giving you a more even coverage on the stage area and in the venue, too. But stage amplification is different —you tend to stand nearer to an amp than you would a PA cabinet, and on a two-way speaker with physically separate drivers, the HF and LF will have different dispersion characteristics through the crossover zone, resulting in some tonal unevenness compared to a coaxial configuration, which is inherently time-aligned. The often-cited limitation of coaxial drivers —narrower dispersion — in this instance actually allows the F12‑X200 to more accurately mimic the directivity of a typical guitar speaker. All of which means, in practice, that if you position yourself with the same relationship to this speaker as you would with a normal guitar speaker, you'll hear much the same tonal balance as you are used to.

For anyone contemplating using F12s in multiples, I only had one to test, but I see no reason why they wouldn't acoustically couple in much the same way as conventional speakers, dispersing more in the horizontal plane when arrayed vertically and vice versa.

Miking Up?

If you are using a modelling rig that includes speaker emulation on stage, then you already have a 'finished' signal, ready to send directly to the PA. The function of having an F12‑X200 behind you is just to make you comfortable in your performance, allowing you to hear something that behaves and feels like a normal guitar speaker. It would make no sense to mic the speaker when you can take a DI straight from the modeller. Play enough gigs, however, and you will inevitably come across a venue with the kind 'sound man' that 'knows' that "DI'ed guitars don't sound any good — you always have to use a mic." So I thought I'd find out what happens if you do put a mic in front of an F12‑X200...

As I expected, 'normal' placements don't work at all well — an SM57 halfway out to the edge of the cone sounds very dark and lifeless — whilst, somewhat counterintuitively, miking the centre of the cone is actually better. With the mic at a typical one- or two-inch close-miking distance, there's too much top and proximity-effect bottom end, without enough mid–range to balance it out. Around six inches back from the centre, however, I felt I had actually found something quite useable, so if you get stuck with having to mic one of these, I'd suggest you start there!

I tried recording a couple of different guitar tones with this setup, capturing the DI'ed signal from the modeller at the same time for comparison. To my surprise, once I'd tweaked the mic position a bit — mainly for distance; the viable positions are all around the centre — the results weren't actually that different, when I really hadn't expected this to work at all!

Here are two audio clips, with an A and a B version of each. In Clip 1, the guitar plays some clean-ish chording and then single-note lines with the same sound. Clip 2 is a bridge-pickup solo sound. After a few lines, you'll hear me kick in a drive pedal for another point of tonal comparison.

See whether you think the A or B clips are the ones with the mic.

Match EQ plot for Celestion speaker.

There's a 'match EQ' plot of the spectral adjustment needed to make the mic sound like the DI, by way of a clue. Ignoring anything above 10kHz and below 50Hz, the curve might suggest that the speaker is perhaps adding a bit of 'bloom' below 500Hz — remember that the plot shows the adjustment need to make the speaker recording sound more like the DI. That sub-500Hz lift could just be a bit of residual proximity effect, and perhaps wouldn't be there if I'd moved the mic back another couple of inches, but in the end I positioned it just where I thought the two sounded most alike. The individual wobbles in the plot are insignificant here: it is the general conformity of the overall plot that is surprising. It just goes to show, in audio, as in many other areas, you might think you know what's going to work, but you don't really know until you've tried it!

You can download the uncompressed WAV files from the Media sidebar (above, right) or from the link below:

Package icon


  • Sounds and feels like a normal guitar speaker when used with a modelling rig.
  • Straight replacement for a conventional guitar speaker in almost any cab or combo.
  • Does exactly what it was designed to do!


  • None.


If you like the sound of your digital amp/speaker modeller in the studio, but just aren't getting along with it on stage with full–range amplification systems, this hybrid guitar/full-range speaker could well be the answer.


£110 including VAT.

Celestion +44 (0)1473 835300.


Published August 2019