Eiosis e2deesser

De-esser plugin

Eiosis e2deesser

Review by Russell Cottier

Eiosis was the brainchild of renowned plugin designer Fabrice Gabriel, who set up the company before teaming up with Steven Slate to co-found Slate Digital. Gabriel recently explained that Eiosis had to take a hiatus, but that over the last year or so Eiosis has hit the market again with some new plugin versions. The e2deesser is not actually a new concept, Eiosis were launching the similarly named E2 deesser in 2008 which had a similar concept at its core but the e2deesser seems to take advantage of the exponential growth in CPU processing power and presents us with a much more advanced and intuitive interface. The unique concept behind the plugin is that there is a split signal path offering a multi functional workflow for de-essing tasks. Gabriel himself boldly claims that the plugin takes ten seconds to set up and only ten seconds to fine-tune though perhaps that might be said for many other de-essers.

In terms of signal flow the input is split into three paths, Detection, Voiced Processing and Sibilant Processing. The Detection algorithm identifies the presence or lack of sibilance in the input signal and switches the output source from either a Voiced Processing stage, where EQ is applied, or the Sibilance Processing stage which implements EQ, wideband gain and the Auto and Smooth algorithms. These two algorithms are useful for producing a smooth accurate de-essing result with minimal effort.

The GUI is classic and uncluttered, it has a brushed metal look with metal styled controls. Additional metering information is given as buttons illuminate and a corona around each knob glows white indicating processing and levels at the related stage of the algorithm. At the top of the plugin there is a preset bar, similar in functionality to that of the major-name plugin houses. Functions include undo and redo, preset loading and saving. There is the obligatory A/B comparison option — of which the settings are stored in the DAW project — not the preset itself. The included presets are actually very usable, and there are various options available from standard de-essing to more advanced functionality. There are even settings for those alternative de-esser uses such as drum overheads, and a rather effective guitar string squeak removal tool. Navigation of these is a breeze with the sub-menu structuring.

Next along the top bar there is a rather curious play icon pointing into a clock. This is the sync option menu for the Spectrogram view. Various options for the timebase displayed are available. A further button allows toggling of the visibility of the Spectrogram. Finally there is a general configuration menu icon that gives access to the lookahead time. The processing load of this plugin is fairly light due to the idle mode (explained below) and it is quite viable to use a 15ms lookahead without impacting CPU usage. The increased look ahead does seem to allow the plugin to catch “T” and “Ess” sounds slightly quicker.

Below the menu bar there is a Spectrogram that can be synced to the transport and the timeline in a number of manners, including the default loop view, and fixed view lengths of 2, 5 or 10 seconds. The spectrogram is drawn out as the audio plays and under loop selection mode this is automatically scaled to fit in the entire loop on screen. However — annoyingly — the loop selection does not work correctly in Pro Tools when Pre/Post Roll is active. The Spectrogram is useful as it shows intensity of the input signal in a blue colour but changes to a copper colour for sections where sibilance is detected. Also when the EQ is being operated in the lower part of the plugin GUI a set of red lines appear in the Spectrogram to indicate the centre of bell curves and cutoff frequencies. This is actually very useful for quick selection of frequencies as the Spectrogram makes spotting the fundamentals very easy.

Below the Spectrogram on the left of the plugin we see a set of metal-look controls, a Mode control allows selection of 8 options. Solo Vocal, is the normal option for de-essing mono vocal sources. In Background Vocals mode the left and right channel are processed independently for sibilance. The Voice Over mode is a little more sensitive than the Solo Vocal mode. After this come the non-conventional de-esser usage modes of which there are several. Guitar Squeaks which looks for gaps in the guitar playing and identifies string squeak in those sections.

It should be noted that this is not a spectral repair squeak removal tool, as you might find in the likes of iZotope RX, however it is very effective for highly compressed acoustic guitar, for example. The Overheads mode is intended for removing peaky resonances from cymbals in drum overhead tracks. There are also two mastering modes, Stereo Mastering and Mid Mastering. The Stereo Mastering mode allows taming of high frequency annoyances that might be popping out of the mix. The Mid Mastering setting solves a very common problem in that it is designed to look for centred vocals in a mix and notch out sibilance from these without influencing the rest of the mix. It is useful that this option is available in the form of a single plugin as attempting this process with multiple plugins can be somewhat time consuming. There is a bypass button in this section that bypasses the entire plugin processing and two large knobs, Sensitivity and Amount. The Sensitivity is basically a threshold control and the Amount scales the severity of the sibilant processing stages.

Underneath this there are four fine controls. Auto can be cranked to dynamically adjust the amount of frequency peaks that are removed from the sibilant sections. These offending frequencies are dependant on the plugin mode selected. Smooth applies a saturation to the Sibilance Processing, and as such, clips it at a given level, reducing the amount that the offending sibilance pops out. Output Gain and Dry/Wet controls are also usefully provided. Finally, there is black screen area to the right that has metering which allows setting of an Idle mode, which disengages processing when the input drops below a threshold to save on CPU usage. This can be automatic or manually set.

The EQ section to the right of this displays three curves, a sibilant wideband gain offsets the red sibilant processing curve which can have up to four bands of EQ applied with Bell, Shelf, LP and Air modes available. Likewise there is a blue Voiced EQ curve that can have four bands applied to the non-sibilant audio. Finally a white Dynamic Curve bounces between the two curves to visualise the current processing state. There is a clever Auto Zoom option to keep the display resolution as useful as possible and a set of listen modes that allow soloing of the Voiced channel, Sibilance channel and a Sibilance Focus mode, which leaves the Voiced processing content in, but with some attenuation.

Overall the plugin is simple to use and an effective workhorse. It leaves little in the sense of artefacts that we are used to in de-essing, and judicious use of the Smooth control reduces the need for further compression down the line for even the spikiest of vocals. At USD $99 (regular price USD $149) it sits slightly above the cost of many other de-essers, but the mastering and guitar modes are certainly worth the extra, and you do get two licences for your iLok account. Despite a couple if operational annoyances the plugin sounds great and works well, there is a free 15 day trial at so there is really no reason not to try it.