Version 1.0.0
Spatial Convolutional Reverb
2026-08-30
build b4cc394
Spatial Convolutional Reverb
Hartley DSP created THeVerb® to allow musicians and producers to make digital instruments and recordings sound alive, with the warmth and character that comes from human performance in a real space. With today’s technology recording live isn’t always an option and, even when it is, it can be difficult, expensive or creatively limiting. THeVerb lets you bring digital sources into realistic acoustic environments, so that you can have the best of both worlds.
THeVerb is a unique spatial convolutional reverb plug-in. It allows you to place and move sound sources within a space, simulating the way sounds are affected by the distance and direction of the source from the listener, and the reverberation that occurs as sound waves reflect off the walls, ceiling and floor. Our brains are exquisitely sensitive to these spatial cues. The tiny differences they create in the sounds reaching our left and right ears create a vivid sense of space even when we are not consciously aware of them. By combining different instances of THeVerb, you can recreate the interaction of sounds from different musicians playing together at the same time, in the same room, even if they were recorded separately.
The THeVerb plugin interface
This section walks you through the process of installing and activating your THeVerb plugin. Activation is required to unlock the full functionality of the plugin, but you can also use it in Trial mode to evaluate it before purchasing.
Visit hartleydsp.com/products to download the latest installer. Run the installer and follow the prompts to install THeVerb on your system. The installer includes the plugin itself, as well as the included room library (a set of different acoustic spaces you can use with your plugin), and this user manual.
THeVerb Installer
VST3 plugins work differently from previous VST versions (which could
be installed to various locations). The VST3 standard specifies a single
location for installation:
C:\Program Files\Common Files\VST3\. Your DAW may not
operate correctly if the plug-in is installed elsewhere.
The installer will place the included room library in a folder within
your Windows user account:
%appdata%\Hartley DSP\THeVerb\IRLibrary (this typically
corresponds to a location like
C:\Users\<username>\AppData\Roaming\HartleyDSP\THeVerb\IRLibrary).
In the HartleyDSP folder we also store other information your plugin
needs, such as your settings and presets.
Your license information is included in the settings file, so you should not modify or delete it unless instructed to by Hartley DSP support (otherwise you may need to reinstall and/or reactivate your license).
When you first open THeVerb (typically adding it as an insert in your DAW) it will run in Trial mode. This mode allows you to use the plugin with some limitations to evaluate it before purchasing, to activate your license when you buy a license (removing the limitations), and to reactivate if your system changes or the license information is lost. When in Trial mode the plugin displays an orange indicator and “Trial Mode” text at the bottom of the interface.
To activate, click on the orange “Activate” button (marked X, above) at the top of the interface, just to the right of the THeVerb logo. This will open the activation dialog, where you will enter your email address and license key (below).
Activation dialog
You can find your license key in the confirmation email you received when you purchased, or in your account on our website. After entering your information, click “Activate” to complete the process. If the activation is successful, you will see a confirmation message and the plugin will be fully functional.
Your license key is tied to the email address you used to purchase, which should also be your login at hartleydsp.com. Be sure to use the same email address when activating your license. If you have any issues with activation, please contact support at support@hartleydsp.com or visit hartleydsp.com/support.
You are permitted to activate your license on up to 3 different machines at the same time. If you need to deactivate a license from a machine (for example if you are no longer using it, or if you have reached the activation limit), you can do so from your account at hartleydsp.com. This will free up one of your activations for use on another machine. If you need more than 3 activations, or if you are unable to deactivate a license, please contact support. If you lose control of the critical email address, but still have proof of purchase, we can also assist with recovery.
The plugin is fully functional in Trial mode, with the following exceptions:
Trial Mode Session Time Expiry When the session timer expires in Trial mode: Spatial features are locked (1 - distance attenuation, distance delay, binaural microdelays; 2 - XY positioning). These features can be unlocked by activating the full version (3 - activate button). To continue evaluating the plugin you can start a new session and the timer will reset.
Trial Mode Room Browser In Trial mode a limited number of rooms are unlocked. You can find them all together in the Trial Bundle section, they are also listed in the included bundles which are otherwise locked until the full version is activated.
If you are enrolled on our beta testing programme, you will be able to download a different installer, and the activation process for Beta versions requires a different license key which you can collect from the beta area of your account at hartleydsp.com.
Beta Activation dialog Note that a different license key is required for Beta versions.
First install and activate your plugin (see Installation and Activation).
Quick workflow:
When you open up the plugin in your DAW, you will see the main interface which is designed to be compact and intuitive. All the controls you need are on a single screen, a few sections expand when clicked to provide options. The main panels are labelled A-F in the figure below; later in the overview the individual controls within each panel are numbered. The Header panel (A) provides product information, settings and help. Just below that the Visualizer panel (B) shows how the plugin is working. Beneath that the central area, panels C,D and E provide the main controls used to configure the plugin, in panel E, Direct/Room Mix acts very like a wet/dry, mix blending the direct signal with reverberation (the direct signal is not strictly dry because it normally includes microdelays and attenuation). At the very bottom of the interface is the Footer panel (F) which provides information about the plugin’s operations, modes and version.
Main interface panels A - Header, B - Visualizers, C - Control Panel, D - Spatial/Room Panel, E - Output/Metering Panel, F - Footer
Hover over any control to see a tooltip with a brief description of its function.
A - Header controls
B - Visualizer panel
These show the effect of your spatial settings and convolution in real time, so you can adjust them to get the sound you want.
In the Control Panel you can adjust various parameters affecting the direct and room signals:
C - Control panel 3 - Presets, 4 - Input Mono/Stereo, 5 - Gain Match, 6 - Distance, 7 - Room Tone, 8 - Reverb and Fade Time
Use the Presets control to capture useful combinations of room, position and processing settings so you can recall them quickly.
Preset popup The first row () creates a new preset. Preset rows show
. Hover to reveal delete
, then click to remove that preset.
If a preset with the same name already exists, THeVerb asks whether to overwrite it.
Open the Presets popup and click a preset row marked .
If that preset includes a room reference, THeVerb also loads the
associated room (or User IR reference) and updates controls
accordingly.
If you change settings after loading, an asterisk (*)
appears next to the preset name to indicate the current state has been
modified.
The built-in Default preset is protected and cannot normally be deleted.
Preset changes can be sent or received by other sync’d instances of THeVerb depending on your sync mode settings. The built in - No Preset - option, leaves the current parameters unaffected but prevents them from being sent to other sync’d instances. The instance can still receive preset changes from other instances.
The Spatial/Room Panel (D) is the heart of the interface, where you choose the acoustic space and place your sound source within it:
D - Spatial/Room Panel 9 - Room Selector, 10 - Sync, 11 - XY Pad, 12 - Instance Bar
Multiple instances When the Instance Bar is active, you can see and manipulate the source positions of other instances of THeVerb in the same project.
The Room Selector is where you choose the acoustic environment that determines the room signal. Click to open the browser and find the library bundles: these are curated rooms with impulse responses modelled on real and imagined recording and performance spaces with different characteristics. Select a room by clicking on its name in the middle column. Click the magnifying glass icon to search the available rooms by name or by built-in tags describing each room’s distinctive character.
By default the room browser remains open so you can switch between rooms and audition them with your track playing. If you prefer to keep the interface uncluttered you can change this behaviour in the Settings/Interaction menu so that the browser closes automatically when you select a room.
Room Browser with Studios and Venues bundle view
If you want to use your own impulse response, switch to User IR mode in the Room Browser and load your file. In the full product, you can then enable Spatialize to map that IR into THeVerb’s spatial model, so movement on the XY pad still produces meaningful positional cues, affecting both direct and room signals, rather than a static stereo reverb. This is useful when you want the character of a custom IR but still need consistent placement control across multiple tracks. The spatialization synthesizes spatial cues based on statistics of thousands of IRs across different simulated spaces, so that your IR behaves like the more detailed and precise spatial models built into our library.
User IR mode with Spatialize Load a custom IR, then enable Spatialize to use THeVerb's spatial positioning affecting both direct and room signals.
Loading your own IR can produce unpredictable results including very loud output. User IRs are normally attenuated on import to approximate the levels used in the built-in library. However, you may still find your IR produces louder room signals than the included library IRs. To avoid unpleasant surprises start with the DAW channel fader on at a low value, and set the Direct/Room mix to a low value (e.g., 0.10). If spatialization is active, resonant frequencies can interact with XY positions, so test with representative audio and try a range of source locations.
If the room signal is too loud or quiet and you are not able to easily get the blend you want with the Direct/Room mix, you may want to carefully adjust the impulse response in an audio editor, for example applying a gain change.
E - Output Panel 13 - Direct/Room Mix, 14 - Metering Section
In the Output Panel (E) you can adjust the balance between the direct and room signals with the Direct/Room Mix control (13). The higher the value, the more room signal and less direct signal you will hear, so it acts very like a wet/dry control, although the direct signal is not strictly dry because it normally includes microdelays and attenuation (see Direct/Room Mix).
In the Metering Section (14) you can see the input, direct signal, room signal and output levels. The meters display the smoothed peak level of the signal (coloured bars) in dBFS and an indication of the peak signal (white line). From left to right the meters show:
The Gain Match options rely on the RMS level of the Input, Direct and Room signals, which are tracked by envelope followers. The current level of each signal is shown by a small triangular meter next to the relevant meter, with the currently active followers (those used to calculate gain compensation) highlighted in green.
The footer at the bottom of the interface shows information about the plugin’s status, such as whether it is in Trial mode, the convolution and graphics modes, the build name and type. These can be useful for troubleshooting and support. On the left hand side of the footer is the notification area. This displays information (such as license status), warnings (for example, if a library file cannot be found) and error messages (for example, “IR Library path not found”). It is normally closed and just displays a count of the number of messages of each type. Click on the notification area and scroll to see the messages. If a message spans several lines you can click to expand it. When a new warning or error message is generated you will see it briefly displayed to the right of the notification area, and the count of warnings or errors will increase.
Buttons at the bottom of the notification list allow you to copy the messages to the clipboard (which can be useful when seeking support) or clear the list. You can also right-click a given notification to copy a specific item.
This section explores the ideas and techniques behind THeVerb’s processing and provides practical guidance for getting the most out of the plugin. For a reference to the interface controls, see the Overview.
Convolution: at any moment the sound you hear is a combination of waves coming directly from sources (instruments, voices, speakers etc.) and reflections of earlier waves that have arrived at your ears after bouncing off surfaces in the room (walls, ceiling, floor, etc.). Mathematically, and in nature, this combination works by adding together the changing waves from the present and recent past according to a fixed pattern of thousands of “echoes” generated by the reflections. Some reflections arrive more quickly and are stronger than others. Mathematically each “echo” multiplies the sound level by different amounts according to the surfaces the waves have hit along the way; different surfaces, and the air itself, absorb different frequencies to different degrees. So, at each moment in time, we are adding together the past waves after they have been multiplied by a factor that depends on how much energy has been absorbed or reflected along the way.
Impulse Response (IR): the impulse response is the fixed pattern of echoes that occur in a given space, and it depends on the sound source’s location, the listener’s location, the geometry of the room and the acoustic properties of the surfaces. But it has a very simple real-world equivalent: the sound you hear at the listener’s location when a very short burst of noise (like a balloon popping) is emitted at the source position. By imposing this pattern on every new wave emitted by the source, we can simulate the reverberation we would hear in the room with any new sound.
Reverberation: because there are so many reflections, and they are so close together, we don’t typically hear individual echoes, but rather a gradually dying reverberation which smears the original sound over a few seconds.
“Dying” and “smearing” don’t sound very pleasant, but in fact reverb can be a very beautiful and natural sound. We never, in normal life, hear a sound without any reverb, and we miss it when it’s not there. The absence feels very cold, dry and unnatural. In many real-world settings, natural reverberation is very subtle, and, to achieve clarity in performance and recording spaces, steps are often taken to reduce it. Although we may not be aware of it, even this very subtle reverberation adds a sense of warmth, space and life to the sound. And with bigger reverb sounds (such as those we would only normally hear in a cave or a church) we become conscious of the effect; it typically adds a sense of significance or grandeur, although it can make individual notes hard to pick out. If you are mixing a slow ballad, you will probably want to add a longer reverb to the vocal, otherwise it will sound lifeless.
THeVerb processes two separate signals. The
direct signal is the sound that reaches the listener
directly from the source, without bouncing off any surfaces. The
room signal convolves the input sound with an
impulse response (IR) that captures the acoustic
characteristics of a given space, simulating the reverberation you would
hear in that room. The two signals are mixed together to create the
final output.
Using the XY pad you can place the sound source anywhere in the horizontal plane of the room, and THeVerb adjusts both signals accordingly. The direct signal is processed with delays and attenuation to simulate the effect of distance and direction. The room signal uses a different IR, or a blend of IRs, matched to the source’s position. THeVerb includes a library of impulse responses for a curated set of rooms, and you can also load your own IRs.
Both the direct and indirect signals provide spatial cues that we use to understand the location of the sound source and the size and character of the space we are in.
The direct signal is most important for understanding source location. More distant sources are quieter (distance attenuation), waves take longer to reach the listener (distance delay), and the direction of the source affects the relative timing of waves reaching our left and right ears (binaural microdelays, inter-aural time difference).
Direct signal spatial cues. Left: distance attenuation (NL left, FR right). Right: distance delay and binaural microdelays (NL left, FR right).
The indirect room signal is most important in understanding the scale and character of the space, but it can also contribute to our sense of source location because the pattern of early reflections (their level, timing and tone) depends on the proximity of the source to different surfaces. The size of the space and the way different surfaces reflect and absorb sound determine the overall duration of the reverberation; larger spaces and more reflective surfaces make it longer and as specific surface materials reflect and absorb different frequencies, these also affect its tone and the way it evolves over time.
Indirect signal spatial cues. Left: early reflections (NL left, FR right). Right: frequency-specific absorption (NL left, FR right).
Overall, the direct signal is strongly affected by the source’s location but unaffected by the properties of the environment, while the room signal is strongly affected by the environment but only relatively weakly affected by the source’s location. However our brains are very sensitive to spatial cues and the subtle changes in the room signal as the source moves around are important for a sense of immersion and realism. THeVerb ensures that the spatial cues in the room signal are aligned with those in the direct signal because they are generated a common source location. When multiple instances are used with the same room, the spatial cues are consistent across all instances, so that the sources appear to be in the same space.
The Visualizer panel (B) shows the direct and room signals in real time. Here is how to read them in detail.
B - Visualizer panel
Left click and drag on either visualizer will adjust the Direct/Room Mix control (when you increase the mix value, the room signal is louder and appears taller, and the direct signal is quieter and appears shorter).
The Direct Signal Visualizer shows how distance and
direction affect the sound that travels straight from the source to the
listener. Time runs along the horizontal axis, with the whole window
covering about 50 ms.
On the left of the plot, a tall white bar shows the relative level of the original source. To its right, a coloured bar represents the sound arriving at the listener’s ears: delayed by a few milliseconds (further to the right on the time axis), quieter (reduced in height), and - if the source is not directly ahead - very slightly earlier at one ear than the other. The narrowing grey “shadow” between source and listener indicates attenuation according to the inverse square law. Blue represents the left ear, orange the right; the slight offset between them shows the interaural time difference (ITD). Because these differences are tiny, the bars are bowed outward to make the ITD easier to see. As you move the source around the XY pad, the display updates to reflect the changing delay, attenuation and ITD.
The Direct Signal Visualizer. Time is on the horizontal axis and the whole window covers just 50ms. The tall white bar (left) shows the relative level of the original source. As the source moves further away from the listener it moves to the right (delayed by the speed of sound). It is attenuated according to an inverse square law (indicated by the narrowing grey "shadow"). The quieter sound arriving at the listener's ears is represented by the blue/orange coloured bar. Blue represents the left ear, orange represents the right ear. If the sound is not directly ahead of the listener it arrives slightly earlier at one ear than the other - this "interaural time difference" (ITD) is indicated by the relative positions of orange and blue bars. As the differences are tiny, the bars are bowed outward to make the ITD easier to see.
The Room Signal Visualizer shows the room signal - the
sound that reaches the listener only after bouncing off surfaces. As
sound waves interact with walls, floor and ceiling, they may be absorbed
or reflected, with different surfaces reflecting different frequencies
to a greater or lesser extent. The reflected waves may bounce several
times before reaching the listener; the longer they travel, the more
energy is absorbed. The visualizer has two parts: an impulse response
envelope at the top, and a dynamic room signal at the bottom.
The Room Signal Visualizer. The upper part shows the impulse response envelope, while the lower part shows the dynamic room signal - the reverberation you are hearing at any given moment, which is the summation of all the reflected sounds arriving at the listener's position; later reflections have generally lower intensity and their spectral characteristics are changed by the absorption properties of the surfaces they encounter.
You can make timing adjustments by altering Reverb Time and Fade Time parameters within the visualizer. To adjust Reverb Time, double-click at the point on the IR where you want the tail trimmed. The trim indicator (a vertical bar with small arrow) appears at the trim location, and the Reverb Time parameter is adjusted (the trim indicator also appears when you use the Reverb Time knob in the control panel). You can adjust the trim location by shift-clicking and dragging the trim indicator. To adjust Fade Time, shift-click and horizontal drag elsewhere in the visualizer. Leftward shift-drag makes the fade time longer, rightward shift-drag makes it more abrupt.
The upper part of the room signal visualizer shows the impulse response envelope: a plot of the overall intensity of the IR over time. With room impulse responses the early reflections are typically more intense, and the sound gradually dies away. Imagine the sound of a balloon popping in a room. In a small living room with carpet and soft furnishings, the sound dies away very quickly. In a large cathedral you will hear it die away only gradually over several seconds.
The animated lower part of the visualizer shows the effect of the impulse response on the input signal in real time. At the far left we see the initial reflections arriving a few tens of milliseconds after the sound was emitted, having bounced off one or two surfaces. But at the same moment we are also hearing sounds that have been delayed longer (to the right of the plot), sounds that have died away a little and have different spectral characteristics as some frequencies are absorbed more than others. These sounds add together to create the reverberation we hear at a given moment - the process is called convolution.
As well as the intensity of the sound, the impulse response also affects the tone of the reflections. A room with hard reflective surfaces such as stone or wood panelling will have a much brighter tone than a room with heavy curtains that tend to absorb high frequencies. In some cases the high or low frequencies may die away more quickly than mid-range frequencies, so that later reflections can have a warmer or cooler tone than the initial reflections.
Spectral colours in the visualizer. The first plot shows the impulse response envelope, while the lower plot shows the dynamic room signal. The sound source was a frequency sweep, illustrating the way different input frequencies are represented by different colours. The second plot shows the effect of the room on a broadband sound (white noise). All the frequencies are present equally in the input, but the colours of the dynamic signal tracks those of the impulse response. The reverberation is being coloured by the absorptive properties of the room.
In THeVerb these spectral properties of sound are represented by colours which correspond to wavelengths of light. Every part of the audible sound spectrum is mapped to a part of the visual spectrum, and sounds that combine different frequencies are represented by a corresponding combination of visual wavelengths. Warm colours - reds, oranges and yellows - correspond to rich bass frequencies; cooler blues and violet to bright treble frequencies; greens and turquoises to midrange frequencies. It’s also possible to see other combinations such as purple (blending bass and treble), white (broadband sound with energy across the spectrum) and pink (broadband with greater energy in the lower frequencies). If you adjust the Low Cut and High Cut filters (see Room Tone, below) you can see their effect on the IR envelope and the dynamic room signal in the visualizer.
THeVerb’s spatial model simulates how a sound source’s position affects what the listener hears. This section explains how the XY pad maps to the room and how the three distance cues interact.
XY Pad Layout and Listener Location The XY pad represents the horizontal plane of the room. The listener is below the bottom-centre of the pad, so moving the puck upward moves the source further from the listener; moving it downward moves the source closer; moving it left or right moves the source in the corresponding direction from the listener's perspective.
The XY pad represents the horizontal plane of the selected room. The pad is square, but the sound stage - the area the source can move within the room can be rectangular. The size and aspect ratio of the sound stage varies by room, in some rooms it can be much smaller than the space (for example in a venue where the stage is much smaller than the audience area). In the current version of THeVerb you cannot visualize these aspects of the room’s geometry (although the overall dimensions are diplayed in the Room Browser), so it is best to rely on your ears to evaluate the effect of source placement in different rooms. We plan to include a more detailed room visualisation in a future version of THeVerb.
The listener sits below the bottom edge of the pad, facing the sound stage. The listener’s position is fixed - you cannot move it. The puck represents the sound source: moving it upward increases the distance from the listener; moving it left or right changes the source’s direction. At the bottom centre of the pad the source is at its closest point to the listener, directly ahead.
Because the listener is below the pad rather than at its edge, even a source placed at the very bottom of the pad is some distance away. This minimum distance is typically a couple of metres leading to some attenuation of the direct signal (which can be offset the using Gain Match feature).
The XY Pad coordinates are retained when you switch rooms, so you can audition the corresponding locations in different rooms. Clearly, very small rooms (e.g., Vocal Booth) do not allow much space for sources to be placed, and it can be difficult to hear spatial differences. Large rooms (e.g., Foundry) have much larger sound stages and the effect of moving the source is much more pronounced.
By double clicking you can move the puck directly to the nearest snap point, corresponding to the corners, near-centre, far-centre and centre stage.
THeVerb allows you to move the source seamlessly in real time. You should notice the direct signal provides a very detailed impression of movement, while the effects on the room signal are more subtle in most rooms. You can automate the X and Y parameters in your DAW, or - if it permits it - link them to a MIDI controller for real-time input.
If you move the source very quickly you may notice a subtle pitch shift which is analogous to the Doppler effect we experience in real life when a sound moves past us. This is normally subtle or inaudible when the source moves at realistic speeds.
The Overview describes the three distance toggles in the control panel. These all operate on the direct signal. Here is what to expect when you use them, and why you might choose to turn one off.
Distance Cues The different distant cues affecting the direct signal can be toggled on and off independently: a) distance attenuation b) distance delay c) binaural microdelays.
Distance attenuation () follows an inverse square law: double the
distance and the direct signal drops by about 6 dB. This is the most
immediately audible distance cue. With it on, moving the source further
away makes it noticeably quieter, which shifts the balance between
direct and room signals toward the room. This is realistic, but it means
the output level changes as you reposition the source. If you need a
consistent level while still using the other spatial features, turn
attenuation off and consider using one of the Gain Match modes instead
(see Gain Management, below).
Distance delay ()
delays the direct signal according to the speed of sound. At typical
room distances (a few metres) the delay is only a few milliseconds, and
on its own it is barely perceptible. However, the presence or absence of
this delay can affect the relative phase of signals when combining
THeVerb’s output with other tracks or with a parallel dry signal.
Turning it off preserves phase alignment at the expense of physical
realism. Unless phase interaction is causing problems, the default (on)
is a reasonable choice.
Binaural delays () model the interaural time difference (ITD): when
a source is off to one side, sound arrives at the nearer ear a fraction
of a millisecond earlier. This is one of the strongest directional cues
our hearing system uses, and on headphones the effect is vivid - the
source is clearly placed to one side. With binaural delays off, the
direct signal arrives at both ears simultaneously and is perceived as
straight ahead. You might turn them off if you are mixing primarily for
speakers (see below) or if you want a more conventional reverb sound
without strong lateralisation.
THeVerb’s spatial model is binaural: the direct signal processing is designed to place sound at each ear independently, producing the most vivid spatial impression on headphones. On speakers, acoustic crosstalk means each ear hears both channels, which dilutes the ITD cues. The room signal still conveys a sense of space on speakers, and the overall effect remains useful, but the precise left-right placement of the direct signal is less sharp.
If you are mixing primarily for speaker playback, you may find that turning off binaural delays and relying on the room signal for spatial character gives a more natural result. If you are mixing for headphones - or for a binaural or immersive format - keeping binaural delays on will give the strongest sense of source position.
Did you know: - between 2014 and 2024 sales of headphones grew by 59 percent (reaching 455 million worldwide). - 78 percent of streaming consumers now listen to music through headphones or earbuds. (Source: The Atlantic)
The Overview describes THeVerb’s tone and time controls. This section provides more detail on how they interact and how to use them in practice.
Room Tone Controls The Room Tone filters (Low Cut and High Cut) shape the frequency content of the room signal only.
The Room Tone filters (Low Cut and High Cut) shape the frequency content of the room signal only - they do not affect the direct signal. The filters are 12 dB/octave, giving a moderate slope that sounds natural rather than abrupt.
In practice, the most common use is managing the low end. Large rooms and rooms with hard surfaces can produce substantial low-frequency reverberation, which may muddy a mix. Raising the Low Cut filter cleans this up without affecting the direct signal’s bass content. Similarly, High Cut can tame overly bright reflections - useful if a room’s natural character is harsher than you want.
The effect of the filters is visible in the Room Signal Visualizer’s spectral colours: raising the Low Cut shifts the colour palette away from warm reds and oranges; lowering the High Cut removes the cooler blues and violets. This makes it easy to dial in the tone by eye as well as by ear.
Reverb Time and Fade Controls The Reverb Time control trims the effective length of the impulse response, shortening the reverberation. The Fade control adjusts the shape of the tail's decay near the trim point, determining how abruptly the reverb cuts off.
The Reverb Time control trims the effective length of the impulse response, shortening the reverberation. This is useful when you want the character of a large room but need a tighter tail for clarity in a mix. Very short trim values can produce a thickening or doubling effect rather than an audible reverb, which can be effective on vocals and close-miked instruments.
The Fade control adjusts the shape of the tail’s decay near the trim point, controlling how abruptly the reverb cuts off. A longer fade produces a smoother, more natural-sounding end to the reverb; a shorter fade gives a more defined cutoff.
When BPM units () are engaged, the Reverb Time control snaps to
musically meaningful durations derived from the host tempo, so that the
reverb tail aligns with the rhythmic structure of the track. This can be
particularly effective for rhythmic material where a reverb tail that
decays cleanly before the next beat or bar produces a tighter, more
controlled sound. BPM units require a tempo from the host DAW; if no
tempo is available, the control reverts to its standard proportional
behaviour.
BPM units retain their relationship to the host tempo when it changes, or when introducing a preset into a new project (assuming “processing” parameters are included in the preset), so the reverb tail will retain its musical timing.
When using absolute rather than BPM units, internally, and in your DAW, Reverb Time is represented as a proportion of the original IR length, meaning that when you switch to a new room the Reverb Time will retain its relative position. If the new IR is longer than the previous one, the absolute reverb time will increase; if shorter, it will decrease. This allows you to maintain a consistent relative reverb length across different rooms, and it means that whatever the value it receives, THeVerb will always apply a valid trim (i.e., it can’t be beyond the length of the IR or less than zero), but the absolute value will scale with the length of the IR when switching between rooms.
The Direct/Room Mix control blends the direct and room signals. At 0.0 you hear only the direct signal; at 1.0, only the room signal. It functions much like a wet/dry control, but with one important difference: the “dry” direct signal is not identical to the unprocessed input. When distance cues are active, the direct signal carries attenuation and delay processing, so even at a mix of 0.0 the output reflects the source’s position.
For subtle room ambience - the kind of treatment where you want a sense of space without obvious reverb - values in the range 0.05 to 0.15 are a good starting point. For more prominent reverb, values of 0.3 and above bring the room signal clearly into the foreground. At the extremes, a mix of 1.0 (room only) can be useful as a send effect, or for creative treatments where you want pure reverberation.
The interaction with Gain Match is worth noting: if you are using Gain Match to Input, the compensation tracks the overall output level, so changing the mix does not cause the level to drift. With Gain Match off, increasing the mix adds room signal energy and the output gets louder.
THeVerb’s spatial processing changes the output level in ways that a conventional reverb does not. Distance attenuation makes the direct signal quieter as the source moves away, while the room signal adds energy. The Gain Match modes address this, but they serve different purposes depending on what you are trying to achieve.
Gain Match Modes a) Gain Match off b) Match to zero distance c) Match to input level d) Match direct signal
Gain Match off ():
the most realistic option. Level changes naturally as the source moves,
just as it would in a real room. Best when realism is the priority and
you are comfortable managing levels manually - or when THeVerb is the
primary level control for the track.
Match to zero distance (): compensates for the minimum attenuation that
occurs even when the source is at its closest point to the listener (the
bottom centre of the XY pad). With this mode, the output at that nearest
position matches the input level, but the signal still gets quieter as
the source moves further away. This is a good compromise when you want
to preserve the sense of distance but don’t want the constant baseline
attenuation to throw off your mix balance.
Match to input level (): the output level tracks the input level
regardless of source position. This uses an RT60-matched envelope
follower - essentially a dynamics processor whose release time matches
the room’s reverb time - so that the compensation decays gracefully
during reverb tails rather than cutting them off abruptly when the input
goes quiet. The effect is a little like a compressor, and it does affect
the dynamics of the signal. Results depend on the character of the input
and the reverb time: with sustained content with only smooth changes in
intensity, the applied boost can be fairly transparent, but when the
input changes rapidly with sharp transients, you may notice the
artificial sound of the tail being shaped. Best when you need to drop
THeVerb into a mix and have the level stay consistent without manual
adjustment.
Match direct signal (): compensates the direct signal’s attenuation,
boosting the entire output (direct and room together) by the same
amount. The direct part of the output matches the dry input level, but
because the room signal is boosted proportionally, the overall output is
louder than the input. The net effect is that as you move the source
further away, the level stays consistent but the sound becomes more
reverberant - a useful behaviour when you want distance to affect
character rather than volume. Like Match to Input, this mode uses an
envelope follower to handle reverb tails smoothly, and this affects the
dynamics of the output, especially with transient-rich, percussive
signals.
A soft limiter prevents any gain match mode from producing excessive output in extreme cases (very large distance attenuation combined with a hot input signal).
User IR Import Import your own impulse response files via the Room Browser
You can import your own impulse responses (IRs) to provide the room signal. IRs can be obtained online, and in principle you can create your own (beyond the scope of this manual). User IRs must be wav files with durations between 0.05 and 10 seconds, and they should be stereo for best results. The quality of the reverb will depend on your chosen IR, and not every wav file will produce viable results.
Loading User IRs can fail for a number of reasons, including the file being too long, too short or in an unsupported format. If the file does not load you will see an error message and the plugin will continue to use the previously loaded room.
User IR Load Error If the selected User IR file cannot be loaded for any reason you will see an error message, and the previously selected room or IR will remain active.
A list of recently used IRs is maintained in the middle column. Use this to quickly reactivate an IR you have previously used. You may also find it useful to keep your IRs together in a fixed location on your disk so you can find them when needed. It is quite possible to use a subfolder of your IRLibrary for this purpose (but do not mix your IRs with the built-in IRs and bundles).
You can find some free IRs with Creative Commons licenses at operairlib.net and there are many other sources online. Such IRs do not inherently support the spatial features of THeVerb, but the full version of the plugin allows you to “Spatialize” User IRs, so that moving the source on the XY pad produces meaningful positional cues rather than a static stereo reverb. Spatialization synthesizes spatial cues based on statistics of thousands of IRs across different simulated spaces, but note that it cannot work in the same way as our own more precise spatial models (the necessary geometric information is not available). With spatialization, you should find that when the XYPad is in the bottom centre position the sound you hear is very similar to the original IR, the room signal varies subtly as you move the source location and the direct signal carries detailed spatial cues just as with the THeVerb rooms. Without spatialization you will be able to use THeVerb as a conventional convolutional reverb, but you cannot move the source.
When using Spatialize with User IRs, you can adjust the dimensions of the sound stage. These correspond to the space represented by the XY Pad (from 5-20m in width, and the same in depth) and can be freely adjusted and automated even after the IR is loaded. They do not affect the spatialization of the room signal (your IR), but are used to calculate the changes in the direct signal as your source moves around. As you would expect, larger sound stages produce more attenuation with distance, greater distance delay, and more pronounced inter-aural time differences (ITD). Although the physical changes can be larger, the psychoacoustic effect directional changes conveyed by ITD can be very striking in even small spaces.
User IR Import Controls Unlike our built-in room library, raw impulse response files do not include information about the spatial layout of the room where they were captured so sound stage dimensions must be provided to determine how the direct signal reacts to changes in the XY Pad's puck location when a User IR is loaded.
The spatialization process affects some frequencies more than others (depending on the IRs characteristics, the source material and its simulated position), so you may want to experiment with different IRs and sound stage sizes to find the best combination for your material. In general you should find that the process results in a more 3 dimensional effect as you move the source around.
When loading your own impulse responses, the output level can be unpredictable. THeVerb attempts to match an imported IR to the range used in library rooms, some User IRs produce substantially louder or quieter room signals than the included library rooms.
To avoid surprises: start with your DAW channel fader low and the Direct/Room Mix at a conservative value (0.10 or so). Audition the IR, then bring the fader and mix up to taste. If the room signal is persistently too loud or too quiet to blend comfortably with the mix control alone, consider adjusting the IR’s gain in an audio editor before importing it.
If User IR spatialization is active, resonant frequencies in the IR can interact with XY positions, so it is worth testing with representative audio across a range of source locations.
The Overview covers creating, loading and deleting presets. This section explains the three component groups and how to use them for more flexible workflows.
When you save a preset, you choose which of three independent groups to include:
Preset Options When saving a preset, you determine which set of parameters is included: one or more of Room, Position, Processing.
Because the groups are independent, you can create presets that store any combination. Some practical examples:
If you change any setting after loading a preset, an asterisk
(*) appears next to the preset name, indicating the current
state has been modified from the stored version. You can save over the
existing preset to update it, or save a new preset under a different
name.
Because of the way different presets can include or exclude certain parameter groups, you can combine them when setting up a track. For example, you can load a preset that defines the room, gain matching, reverb time adjustments and filter settings, followed by another that controls only the location of the source. By setting up a reusable collection of presets, you can quickly specify the acoustic environment and spatial layout of an ensemble.
When Sync is enabled across multiple instances, loading a preset or changing the room in one instance propagates the change to all synced instances. See Multi-Instance Workflows for more on how sync works.
THeVerb’s ability to run multiple instances in the same DAW project - each processing a different track but sharing the same acoustic space - is one of its distinguishing features. This section explains how to make the most of it.
When working with muliple instances, it is often useful to see them
in a single plugin window. To do this, ensure that the multi-instance
bar is active (). If you see the closed eye icon (
) instead,
click it to make other instances visible. The instance you are currently
acting on directly has a solid blue puck; the other instances have
hollow pucks of different colours. You can move any of the pucks.
Briefly clicking the key icon in the instance bar (
)
toggles a legend listing all the instances with their names and marker
colours. Hold down the mouse icon to show the instance names next to
each puck.
Multiple instances When the Instance Bar is active, you can see and manipulate the source positions of other instances of THeVerb in the same project.
The Sync button () puts an instance into a shared pool. When Sync is
off (
), the instance is independent and neither sends
nor receives changes.
What happens when a synced instance changes its room or loads a preset depends on the sync mode setting. In the default mode (“Room only”), room changes propagate to all synced instances but preset changes do not, so each instance shares the same acoustic space while retaining its own processing settings. The other modes allow preset changes to propagate as well, either together with room changes or independently. See Sync Mode in the Settings and Help section for a description of each mode.
A common use of sync is to place different instruments/tracks in the same room, and we might typically want those instruments to be in different locations. In this case use the default sync mode (“Room only”) to avoid moving other instances’ sources when changing presets. However many other uses are possible (for example applying common room tone or reverb time settings across instances) or moving groups of tracks to shared locations (e.g., as if mixed through the same monitor), in which case you may want to deliberately sync presets as well as, or instead of, rooms. If your provide your own impulse response instead of a library room (see User IRs, above) it will also be synced to other instances.
When a preset does propagate to synced instances (in modes that allow it), it applies whatever parameter groups it contains. If the preset includes a Room group, all receiving instances switch to that room; if it includes a Position group, each moves its source puck to the saved position; if it includes a Processing group, all update their processing controls accordingly. This is worth bearing in mind when designing presets for multi-instance use: a preset that includes only the Room group will change the room across all synced instances without disturbing their individual positions or processing settings. One exception is the “- No Preset -” option, when selected it does not propagate to other instances, even if the sync mode allows preset changes. This may be useful if you want to avoid unintended changes to other instances.
If you want preset changes as well as room changes to propagate, you can switch to a different sync mode. For a one-off setup you can also enable sync temporarily to establish a shared room, then disable it on instances you want to adjust individually.
The basic workflow for placing several instruments in the same room:
The result is that each track sounds as if it were performed at a specific location within the same space. On headphones, the binaural processing gives a vivid sense of each source’s position; on speakers, the room signal provides a coherent shared ambience.
Name your instances (using your DAW’s track naming) so they are easy to identify in the Instance Bar. If your DAW supports it, track colours will be used to distinguish instances in the plugin interface.
THeVerb’s spatial model assumes a single source location. When the input is set to Mono, the signal is summed to a single channel before spatial processing, which produces the most realistic result - the sound genuinely comes from one point in the room. This is the natural choice for individual instruments, vocals, and any source that has a single physical location.
Stereo mode processes the left and right channels independently through THeVerb’s spatial pipeline. This can produce interesting and creative results with stereo sources such as drum overheads, stereo synths or stereo-recorded instruments, but because two independent channels are being placed at the same XY position, the spatial impression is different from what you would hear from a real single source and the stereo information can conflict with the spatial cues generated by the plugin. In a multi-instance setup where realism is the goal, Mono is usually the right choice for each track. Stereo can produce some interesting and often subtle creative effects.
If you want both stereo and realism, consider routing left and right channels to separate mono tracks and positioning them independently with different instances of THeVerb.
THeVerb’s rooms are stored on your disk in the “IR Library”. The library is installed alongside the plugin and placed in a folder within your Windows user account:
%appdata%\HartleyDSP\THeVerb\IRLibrary.
The library is divided into bundles - curated collections of rooms with related characteristics. Several bundles are included with the plugin. A small subset of rooms (the “Trial Bundle”) is available in Trial Mode, before the full product has been activated. In addition, a few rooms are built into the plugin binary itself, so that basic operation continues even when the IR Library on disk is missing, incomplete or damaged (if a built-in room is being used in place of a library room, a notification will tell you so).
You can browse the individual bundles (described below) or click the magnifying glass icon to search by name or by built-in tag. Click on a tag to list rooms with the relevant character. Multiple search criteria (names, tags) can be combined as filters to reduce the number of listed items to those which fit all of them.
Room Browser Search. Left: By clicking the magnifying glass icon you can search the available rooms by name (text box, top left) or tags (first column of browser). Right: When multiple criteria are selected the list narrows down to include only those rooms fulfilling all criteria.
The plugin comes with three curated bundles of rooms each including 15 rooms:
The Trial Bundle is a small selection of rooms included with the plugin, available in Trial Mode before activation. There are two rooms from each of our three curated bundles:
From Studios & Venues: St Johns Wood 2 (the default room which is loaded with a fresh instance) and LA Coffee From Rooms & Booths: Pine Room and Vocal Booth From The Works: Foundry and Pit
The room browser marks each room with a status dot, and unavailable rooms display an explanatory message in the details column. There are two distinct problems the plugin can detect:
In both cases, if the room belongs to one of the included bundles, re-running the installer is likely to fix the issue - the installer replaces the included library files. Your license and settings are stored separately from the library and are not normally affected by reinstalling. Just in case, it is wise to make a note of your registered email address and license key before reinstalling (both can be found in the activation panel, or in your account at hartleydsp.com). The same advice applies when an entire included bundle is unavailable.
In our experience, the most common cause of missing or damaged rooms is manual changes to the library folder - files moved, renamed or edited outside the plugin. See Changing the Library Location before reorganizing anything by hand.
Rooms shown with a lock icon() are present on disk but not
available under the current license:
Free and premium expansion bundles may be made available at hartleydsp.com. If you obtain an expansion, follow the instructions included with it to register and install the bundle, then refresh your license (as described above) so that the new rooms become available in the plugin.
Advanced users can relocate the IR Library - for example to a larger drive. The “Change Library Path…” button at the bottom of the room browser changes the location where the plugin searches for the library. It does not move, copy or modify any files.
To relocate the library safely:
The change applies to all instances of THeVerb on the machine. If the
new location proves incorrect (for example, the browser shows “No
bundles found”), use the same button to point back at a working copy -
the default location is
%appdata%\HartleyDSP\THeVerb\IRLibrary. Any unintended
change to the structure or content of the library files and folders can
cause rooms or bundles to become unavailable; re-running the installer
restores the included bundles to the default location.
The “Change Library Path…” button never moves your files. Pointing it at an empty or incomplete folder will make rooms unavailable until you point it back at a complete library.
If you open a DAW project that references a room which is currently unavailable - missing, damaged, or locked under the current license - the plugin loads a default room instead and posts a warning in the notification area, so your session keeps working.
The project’s original room selection is generally preserved: the plugin retains the original room reference, and saving the project keeps it, so the correct room returns once the library is repaired or your license refreshed and the project is reopened. The preserved reference is replaced only when you actively select a different room in that instance - at that point the plugin treats it as a deliberate new choice.
If a project reports an unavailable room and you want to be certain of keeping the original selection, the cautious approach is to close the project without saving, repair the library (or refresh your license), and then reopen it. For extra safety, you can confirm the repaired installation is working in a fresh project before reopening important saved work.
User IRs (your own WAV files loaded via the User IRs section of the room browser) are referenced by their file path. If the WAV file has been moved or deleted, the project will load a default room and report the missing file; restoring the WAV to its original location and reopening the project restores the selection. The original file path is not retained across repeated saves, so restore the file before saving the project again.
The header panel provides access to the Help menu () and
Settings menu (
). This section describes both menus in detail.
Click the help icon () in the header to open the Help menu.
Online Help: opens the THeVerb support page at hartleydsp.com in your web browser. This page provides getting-started guidance, answers to common questions, and links to further resources.
Manual: opens this user manual. The plugin first looks for the PDF installed alongside the plugin (which always corresponds to the version you installed); if it cannot be found or opened, it falls back to the online version at hartleydsp.com (which is updated to the latest version). The manual is often the best port of call for detailed information - if you experience a technical issue that is not covered, you can raise a support request by using the support form (via Online Help - you will need to log in) or email to support@hartleydsp.com. Beta users should use the Beta Report form.
Gather Diagnostics: opens a diagnostics window showing detailed information about your system, plugin configuration, audio settings and license status. You can copy or save this report and include it when contacting support – it helps us understand your setup quickly. The report does not include personal information unless you explicitly enable the checkbox at the top of the diagnostics window.
Trial Mode: opens an in-plugin help dialog explaining how Trial Mode works and how to activate. This entry appears only when the plugin is running in Trial Mode (for example, after a fresh installation before activation).
Rooms Help: opens an in-plugin help dialog covering common room availability issues – missing or damaged rooms, locked bundles, and how to resolve them. See also The IR Library for more detailed information.
Click the gear icon () in the header to open the Settings
menu. At the top of the menu you will find the activation or license
management entry (see Activating Your
License). The remaining entries are described below.
Sync
Mode: When Sync is enabled (),
changes you make in one instance can propagate to other synced
instances. The Sync Mode setting controls exactly what propagates
Room only: when you change the room in one synced instance, all other synced instances switch to the same room. Preset changes do not propagate. This is the default state.
Preset only: when you load a preset in one synced instance, all synced instances load the same preset. If the preset includes a room, the room changes accordingly on every synced instance. If the preset includes source location - all synced instances move their source puck to the same location. If the preset includes processing settings, all synced instances update their processing controls accordingly.
Room or preset independently: room changes and preset loads both propagate, but independently of each other. When you change the room, only the room is synced. When you load a preset, the preset is synced and applies whatever it contains (including its room, if it is included in the preset).
Room and preset: both room changes and preset loads propagate to all synced instances. A room change also re-syncs the current preset, and a preset load also re-syncs the current room.
The sync mode applies globally – all synced instances share the same mode. The setting is saved between sessions. For more on multi-instance workflows and when to use sync, see Multi-Instance Workflows in the In Depth section.
If you want to set up several instances in the same room but then adjust presets independently per track, try “Room only” mode. Set the room once (it propagates to all synced instances), then work with presets per track without affecting the others.
To avoid unintended preset changes you can select “- No Preset -” from the preset menu. When this is selected - the setting is not propagated from the sending instance even in “Room and preset” or “Preset only” modes.
Auto-dismiss Room Browser on load: When this setting is active the Room Browser closes as soon as a new room is successfully loaded, meaning the browser does not block the underlying plugin interface. When it is not active (the default), the Room Browser remains open so that you can continue to audition rooms while browsing. You can actively close the Room Browser at any time by pressing the “Esc” key, by clicking on the Room Selector (the button that shows the name of the current room which you use to open the browser) or anywhere else outside the browser.
Enable drag other instances: When active (the default) and where the instance bar is set to “show”, you can move other instances’ source locations by dragging them on the XY pad.
Show instance name on drag: This option
(active by default) displays the instances name next to its puck when
you are dragging it on the XY pad. Note that at any time you can display
all instance names by holding down the mouse icon on the key icon (),
and briefly clicking the icon toggles a legend that shows all instances
marker colours and names.
Show/Hide Visualizers and Clean Mode. THeVerb provides options to reduce visual clutter when you need to focus on the sound.
Show visualizers: this checkbox toggles the Visualizer panel on and off. When hidden, the visualizer row collapses and the plugin window becomes shorter. Visualizer processing is also suspended, which may reduce CPU usage slightly. The setting is saved between sessions and shared across instances.
Hiding the visualizers can be useful when screen space is limited (for example on a laptop, or when running many plugin instances), or when you have already dialled in your settings and no longer need the visual feedback.
Hiding the visualizers does not affect the audio output in any way – it only affects the display and the associated processing overhead.
Clean Mode: this checkbox toggles an alternative, simplified interface with fewer icons and less colour, for times when you want to focus on the audio without too much visual distraction. The setting is saved between sessions and shared across instances. Enabling Clean Mode means that you lose the spectral colours which can be useful in understanding the frequency content of the reverb. You also lose some less critical indicators such as the envelope follower indicators, but you may find the result more soothing.
By hiding the visualizer and enabling Clean Mode you can have a very minimal interface, while retaining all the controls you need and leaving the audio output unaffected.
The Advanced submenu contains settings that most users will not need to change. They are provided for troubleshooting, performance tuning, or specific workflow requirements.
THeVerb’s convolution engine can operate in different modes that trade latency (the delay between input and output) against CPU usage. Performance is affected by your system hardware, your DAW/soundcard’s configured sample rate and buffer size, and the particular room/IR you are using. In most circumstances the default mode is the best choice.
Each mode in the menu shows its effective latency in milliseconds and its relative CPU cost, computed from the current host buffer size, sample rate and IR length. The values update whenever these conditions change.
Efficient (default): a good all-round choice that works well for most host configurations. You should not change the default unless you are experiencing performance issues or need to reduce latency for a specific workflow.
The remaining modes are alternatives that may suit specific situations:
Balanced and Detailed: these use progressively smaller internal partition sizes which can smooth out the processor load at the cost of higher overall CPU usage.
Lowest latency: If you select a small buffer size in your DAW (<256 samples), this setting can provide reduced latency in principle.
The change takes effect immediately. The convolver mode is saved between sessions.
The latency and CPU figures shown in the menu are estimates. Actual CPU usage depends on many factors including your system hardware, other plugins in the signal chain, and the DAW’s own processing overhead. If you experience audio dropouts or glitches, try switching to a mode with lower relative CPU cost.
In most cases the default mode provides the best balance and you do not need to change it. The alternatives exist for situations where you are working at unusual buffer sizes, encountering performance issues, or need the absolute minimum latency.
Low latency will not typically be important in a mixing scenario as, during playback, the host compensates for plugin latency by aligning all tracks in the signal chain (plugin delay compensation, sometimes called PDC). However, if you are using THeVerb in a live monitoring context, or in a DAW that does not compensate for plugin latency, lower latency might seem attractive.
With conventional reverb plugins latency is less problematic as the wet signal is already somewhat diffuse and delayed with respect to the dry signal. In many scenarios the small amount of delay introduced by a plugin is insignificant in the context of predelays and reflections introduced by the effect; low latency is not critical. THeVerb is atypical in that it also processes the direct path - altering the timing of the “dry” signal - making misalignment more noticeable, especially when the mix control is set to a low value. However in a live monitoring scenario (e.g., live performance, recording vocals) you are unlikely to need the direct signal, and you may prefer to set the mix control to 1.0 or toggle all distance processing off. In these circumstances lower latency may not be as important as preserving CPU headroom.
THeVerb can use hardware-accelerated graphics (OpenGL) or software rendering for its interface.
Auto (recommended): the plugin tests whether hardware acceleration is available and uses it if so. This is the best choice for most systems. Some older systems and those with built-in graphics chips may not support OpenGL or may only support older versions of OpenGL that do not work with THeVerb - this is detected automatically and the plugin falls back to software rendering.
Force software: graphics are drawn using the CPU. This should work on any system, but may be slower and contributes to the CPU load. Use this if you experience graphical glitches, display corruption or crashes that you suspect are related to your graphics driver.
Renderer changes take effect when the plugin window is next opened. After changing this setting, close and reopen the plugin window (or remove and re-add the plugin in your DAW) to see the effect.
For best compatibility, use standard uncompressed WAV files. THeVerb currently supports:
.wav)In practice, common IR files such as 24-bit PCM WAV and 32-bit float WAV are supported. Compressed WAV codecs (for example ADPCM, mu-law, a-law, or MP3-in-WAV) are not supported in this import path. Due to their very short duration IRs used to capture guitar cabinets may be compatible with THeVerb. A minimum sample rate of 44.1kHz is recommended. The IR is resampled at the session sample rate selected in your DAW, so unless you are using a very high sample rate for your session there is no advantage to using an IR with a higher sample rate.