Many limiters are designed to be as transparent as possible — but Inward Connections' TSL3 is built to sprinkle valve magic over your music.
One of the most tedious questions that is still being asked far too often on mix and mastering Web boards is: "What is your favourite limiter?” Decoded, this means: what relatively inexpensive look‑ahead, digital brick‑walling plug‑in can I buy which will most easily allow me to attain really quite out-of-whack levels on my mixes with the least possible damage? The aim here is obviously not sonic enhancement, it is simply brute loudness, with the least sonic impairment.
The same question asked specifically of classic analogue limiters decodes very differently. Although they too were originally designed for one particular technical function — fast limiting the dynamic range of an audio signal, for example to accommodate broadcast or recording requirements, vinyl cutting and so on — that original purpose was soon complemented by an additional aesthetic function.
At least since Geoff Emerick started pushing his Fairchild 660s on Beatles recordings back in the mid‑'60s, engineers have been asking: "How much better can this limiter make the material sound?” It's not nostalgia that means decent analogue limiters are in demand: it's simply the rational desire to seek maximum musical enhancement in your audio engineering rather than minimal, but still basically unnecessary, impairment.
The device under review here — the Inward Connections TSL3 VacRac — is a vacuum-tube analogue limiter designed to make certain mixing tasks slightly easier, and your mixes sound substantially better. I can say straight away that it meets both of these design imperatives extremely well: this is a very, very, good‑sounding box.
Inward Connections, based in California and headed by designer Steve Firlotte, have been producing vacuum‑tube signal processors since 1987, and an earlier version of their limiter was available over a decade ago as part of the VacRac modular system comprising mic preamp, EQ and limiter, and as such was reviewed by Hugh Robjohns in SOS March 1998. Things have moved on since then, and the limiter has separated from its modular beginnings, grown a very highly specified valve power supply and — in the unit supplied for review, and as an available option for all units — sprouted a pair of transformer-balanced outputs.
At 3U high, with classy dark‑brown bodywork, a pair of huge three‑inch 'retro' VU meters, two pairs of large steel switches, and two pairs of large Bakelite brown control knobs, the TSL3 makes a strong and attractive visual statement, and it drew admiring looks and comment from every visitor to the facility while it was in my rack. As the existence of just two pairs of control knobs — one pair labelled 'Reduction', the other 'Gain Makeup' — indicate, the user‑controlled variables for this limiter are very limited, which at least makes operation exceedingly simple.
As there are no threshold or time‑constant controls, you simply turn one knob to bring the limiter to bear on the signal, and you turn the other to make up the lost gain, and beyond. One pair of switches sets the VU meters to display either the line output level or the amount of gain reduction being instigated. The other pair bypasses the limiting circuitry — though, alas, not the gain make-up amplifier, nor the gain controls, making A/B of the processed and unprocessed signals rather difficult — and there is a single 'Link' switch which makes the TSL3 operate in stereo rather than dual‑mono mode.
Although the main focus of my own business is mastering, I quite like recent developments in the music-production process where the boundaries between mixing and mastering are sometimes blurred. The majority of my clients still send me work and say 'do your thing' — something which could be anywhere from the clarifying and smoothing techniques I like to think is our forte, to a simple kiss of EQ and a lick of limiting . But I am also occasionally involved in more collaborative kinds of work, where the changes brought about by the mastering processing provide insight for beneficial retrospective changes in the mix. This is not identifying problems in the mix, as such — the quality of the original mixes for this kind of work are generally excellent — but allowing the mixer to identify previously unnoticed possible optimisations. Remixes are sometimes to standard stereo, but also to stereo‑minus a particular remix element. And it was on a latter such project that I first used the TSL3.
The track was a texturally rich rock song with a number of acoustic instruments weaving in and out, playing percussive cross‑rhythms which really drove the track along. In particular there was a 'chorus' of two or three acoustic guitars which were being brought in and out as a single element in this mix interplay. During the original mastering I had tried to use some compression to solidify this 'chorus' and though it hadn't worked as well as I had hoped it would, it did raise the possibility in the mixer's mind of separating that element from the mix, treating it differently and then remixing it. So he brought the mix files across to the mastering facility and we re‑ran the acoustic guitar tracks through the TSL3 with a very healthy dose of compression. The VU meters were flapping in a way that made my mastering mind panic a bit — such a rate and level of deflection would normally be signalling a pretty awful sound — but in fact the result was just what we wanted: the chorus of guitars became more 'choral' but at the same time tonally more defined. It's easy enough to mash acoustic guitars and hope the glue doesn't sound too sticky, but here with the TSL3 there was control of tone, in terms of better-defined low end and so better-defined upper mids, as well as (or because of) dynamic control.
Alongside this kind of 'mixtering' work is the more normal (though much maligned, probably not fully understood, and still by no means common!) 'stem mastering', where broad elements of the mix are divided out for separate treatment, generally for problem-solving.
So the second occasion on which I used the TSL3 was actually due to an accident of circumstances which required a stem-based approach. I was asked to master a modern jazz EP where the original bass part had been seriously compromised because the acoustic double bass was unhappy with the recording studio's humidity, and so had become prone to pronounced buzz, clicks and ill‑tempered tuning. As there was actually very little I could do with what I was given, and the bass player was very unhappy with how he sounded, it was decided to re‑record the bass part under better conditions and then bring this to me to be added to the rest of the mix, which had already been heartily approved by the rest of the band and so wasn't to be touched. As I had no idea of the required levels, both the mixer and the bass player sat in while we put the two elements together. The aim was to come up with a composite stereo mix which I could then later master in the normal fashion. The new bass track had been recorded quite nakedly, with a single Schoeps mic through a Millennia preamp, and although it sounded pretty good, with decent low‑end heft, a nice smattering of percussive finger/fretboard noise and no sonic nasties, it was dynamically pretty uneven due to the bass player's movement on the mic and to the nature of the instrument itself. It was embarrassingly easy ("You get paid for this?”) to run the track through an outboard loop consisting of a Maselec EQ and the TSL3, using the former for the mildest of lower-mid cuts and upper‑mid boosts and the latter for dynamic control. And just as sometimes the best compression technique can be a little bit of EQ, so here (as with the acoustic guitars) the best EQ seemed to be a result of the compression: not only did the TSL3 smooth out the jumpiness of the performance, it added some lovely girth to the low-end tones. I hate hearing jazz bass where unsubtle use of compression makes it sound as thought the instrument has rubber bands for strings, with no attack and no genuine sustain. My ideal sound is the opening glissando growling that Danny Thompson gets at the beginning of the title track of the late John Martyn's Solid Air. The TSL3 didn't rubberise the bass, but it was a great help in getting that naturally sustaining, flowing, tone.
Another occasion when the TSL3 came into its own was during the mixing of a quasi‑classical/cabaret mezzo soprano recital. The instrumentation here was simply voice and piano, which I had recorded myself, naturalistically, as if it were a standard classical session. The mixing was minimal, basically just setting fairly static levels for piano and vocal, with a touch of subtractive EQ for both (taking some room boom out of the piano, and some harshness out of the more forceful vocal passages). The irksome issue was that just sometimes with the normal setting the vocal would still appear a bit too strident. This was a feature of both its level and its frequency content, and though one approach to the problem could be further automated EQ and gain riding for those passages, I have often found that a subtle valve compressor, such as the Pendulum OCL2, on its Fast setting an a low ratio, is the better bet. And so it was, but the producer was one of those I've already mentioned whose eye was caught by the appearance of the TSL3 in the rack, and so he asked me to substitute the one for the other.
I wasn't at all sure it would work: as I explain below, the one area the TSL3 did not perform so well was as an overall mastering compressor, and so I assumed that it would underperform here too. But we were pleasantly surprised: there was no pretending that the TSL3 was at all as transparent and neutral (nor as controllable) as the OCL2 doing the same duty; but with gain reduction only ever registering strongly on the meters during the problem passages, the colour and burnishing it gave to the vocal, along with that element of control, was pretty perfect for the material we were working on.
Finally, as well as limiting individual tracks, I also gave the TSL3 a run with a few mix‑bus sessions and in some conventional mastering sessions. In both cases the key was really minimal use, setting the gain reduction so that the VU needles were hardly moving. On most of the projects where I tried it on the mix bus, the results were deeply satisfying in all the ways you'd like a bus compressor to be, adding a touch of density and smoothness — and apparently sheer class. However, when I used the TSL3 on completed mixes that had come through for mastering — where I might normally use, for example, a Thermionic Culture Phoenix — I had less success. Maybe it was the transformers in the unit we had, maybe it was the lack of precision in the control knobs at the very low end of their reach (I found myself using the screw on the back of the control rather than the printed line on the front as my visual reference), but most of the time the TSL3 simply added too much in this application.
The last point about the unsuitability of the TSL3 as a mastering compressor should be taken in context: it is not meant to provide the transparent control that mastering requires, and it would be a bit dim to expect it to do so — it was an experiment. In the applications for which it is designed, the TSL3 VacRac is a real peach!
Unless you can afford $35,000+ for a vintage Fairchild, real‑world alternatives include the new Anamod AM670, the Thermionic Culture Phoenix Mastering Edition, the Manley Vari‑Mu and the Knif Vari‑Mu.
- Superlative sound.
- Easy to use in both tracking and mixing environments.
- Looks lovely.
- No full bypass.
- Maybe slightly too 'juicy' for some mix‑bus and mastering work (the unbalanced version might be different in this respect).
- Like everything else in this top‑notch category, it isn't cheap.
This is a look‑good, feel‑good box that can do amazing things with vocals, bass and acoustic guitars.