Lars Lentz Audio™

Why? Why would I want to do this?

Well, through trial and error and sheer luck, I found how much I like the sound of the Sony DPS R7 reverb and set out to match this with a software plugin version. There is no direct 1:1 conversion, though, so here’s what I did to get there.

The Sony DPS-R7 Sound

Biosphere (Geir Jenssen)

It started when I began analyzing Geir Jenssen’s music, which he made as Biosphere. In particular, I like the albums Cirque, Dropsonde, Substrata, and Shenzhou.

I know he used a Sony DPS R7 reverb based on interviews with him (see sources at the bottom of the post), and I checked out what it sounds like (see the video below).

Instantly after hearing this, I thought, “this is Biosphere.” So it is the reverb that makes the sound for me.

I sought to get the same reverb sound from a plugin, but none exactly produces the R7 sound. However, BReverb2 from Overloud, while not specifically modeled on it, fits the time period nicely.

Here’s the final result (video) of my implementation of the DPS R7 within BREVERB 2, so you can see where I’m going with all of this:

You can get all these BREVERB 2 presets by following the link in the sources section at the bottom of this post.

Given how Overloud describes BREVERB 2 (and how those eras’ studio racks looked), the “most famous/popular units” they’re alluding to would broadly include:

• Lexicon 224 and 480L — flagship studio reverbs of the era, with lush halls, plates, and rooms; these are explicitly referenced in Overloud’s promo material.
• Lexicon PCM series (PCM70/80/90) — workhorse rack units with versatile algorithms and distinctive modulation, widely used in late-80s/90s production.
• Sony DPS-R7 — a high-quality Japanese 1U digital reverb from the early 90s with smooth algorithms, full-bandwidth reverbs, and a mix of subtle and characterful spaces.
• Yamaha SPX series (SPX90, SPX990, etc.) — multi-effects units with recognizable room, hall, and gated reverbs that show up in many 80s/early-90s mixes.
• Roland SRV series (SRV-2000, SRV-330) — classic Roland digital reverb racks that provided dense halls, plates, and rooms with a characteristic early-digital coloration.
• EMT digital units (e.g. EMT 250/251, later 266/267) — high-end reverbs from a company known for plates, giving studios algorithmic options alongside their hardware plates.

The Sony DPS-R7 is what I am after, so what exactly makes it sound like it does, and what are the corresponding parameters in BREVERB 2 that I would need to change to get there?

• Diffusion high (75 to 95%). The R7’s multi-ER architecture builds density fast; BREVERB reproduces this with Diffusion, not Size.
• Ceiling at ~18 kHz, not lower. Keep Damping/High Cut at 18 kHz. Rolling off earlier makes it sound like a plate emulation rather than the full-bandwidth Sony, so don’t do that.
• A touch of Motion/Depth (15 to 30%). The R7’s converters and delay-line modulation add subtle movement; a little Motion prevents the sterile “frozen” tail of purely static algorithms.
• Presence bump ~4 kHz (2 to 3 dB). Recreates the R7 “Presence” high-mid tone control that gives its reverbs their recognizable forwardness.
• Low shelf 1 to 2 dB @ ~120 Hz gives the low-end warmth that the R7 has.

Of these, the damping set at 18 kHz seems to be the most important. The R7’s headline trait is near-full bandwidth, as it passes lows down to 10 Hz and tops out cleanly at 18 kHz.

So, following these rules and my ear, I set BREVERB 2 up for this, then experimented with different algorithms to create various presets. With BREVERB 2’s dry input at 0 dB and wet (reverb) at about –6 dB, I could get reverbs that sound surprisingly like the R7 as used by Biosphere. Varying amounts of wetness are needed, but I can judge this by ear.

What follows is the longer version of that work: the full parameter-by-parameter translation I ended up with, and then the album-specific presets I built from it.

A note on method, and one rule I set myself

This isn’t a “here’s an IR, load it and you’re done” post. BREVERB 2 is an algorithmic reverb, not a convolution engine, so there’s no way to just capture the R7 and drop it in. What I did instead was go through the R7’s documented specs and BREVERB 2’s actual parameter list side by side and build a real parameter-for-parameter translation. Every setting below is a real, named control that exists in BREVERB 2. I checked each one against the manual before writing it down.

And I made myself one rule going in: no separate EQ plugin in front of the reverb. If BREVERB 2 has a native way to do it, like Low Cut, Damping, a shelf on its own EQ, etc., then I use that instead of reaching for another plugin. It’s a good exercise in knowing your tools properly, and it keeps the signal chain simple.

One thing to flag about the numbers: the dry 0 dB / wet –6 dB balance above is what I landed on running BREVERB 2 as an insert, judging it by ear. Several of the presets further down are written for a 100% wet aux send instead, which is how the R7 was actually used as outboard gear hanging off a mixer. Same reverb, different routing. Use whichever fits your session and ignore my Master numbers when your ears say otherwise.

Why the R7 Sounds the Way It Does

The DPS-R7 is an 18-bit, 40 kHz stereo reverberator from 1991 with a genuinely wide frequency response of 10 Hz to 18 kHz, only –1 dB down at the edges (Sound On Sound, Dec 1991; Retrosonic). That full-bandwidth character is a big part of the “sheen” people talk about. It’s not dark, and it’s not brittle; it just passes almost everything. It ships five algorithm types: Hall, Room, Plate, Gate, and Early Reflection, and inside something like the Plate program there are around 97 editable parameters: multiple pre-delays, cross pre-delays, three early reflections plus a cross-ER, a “Presence” tone control, “Rotate High” for HF damping, Spread, and Size (Sound On Sound). That’s a lot of depth for a rack unit from 1991. Check out the manual for it at the bottom of this post.

BREVERB 2 is built on different bones (it’s a Lexicon-lineage algorithmic design), but it turns out the two units think about reverb in remarkably similar terms. BREVERB 2’s Hall, Plate, and Room algorithms each give you Time, Size, Diffusion, a Shape/Spread pair, a Motion/Depth modulation section, a frequency section with high/low decay multipliers and crossovers plus Damping, and a two-band fully parametric EQ (Overloud BREVERB 2 Manual). Once I laid the two parameter lists side by side, the mapping mostly fell out on its own.

Mapping the DPS-R7 to BREVERB 2

The key structural translation between the DPS R7 and BREVERB 2:

DPS-R7 concept	BREVERB 2 equivalent	Notes
Reverberation Time (0.3–99 s)	Time (Main) + Size interaction	In BREVERB, Time is scaled by Size; set both to reach a target RT60.
Pre-delay 1/2 (1–22527 words)	Predelay (Pre/Mod page)	R7 word-based; convert to ms.
Cross Pre-delay 1/2	Regen L/R + stereo Predelay	Approximate; BREVERB has no true cross-feed pre-delay.
Size	Size (Main)	Direct conceptual match.
Spread / Shape of envelope	Shape + Spread (Hall/Plate)	Controls attack/build of the tail.
Diffusion (echo density buildup)	Diffusion (Main)	Direct match — the “density” control.
“Rotate High” (HF damping)	Damping (Hz) + High decay multiplier	Splits into a crossover + multiplier in BREVERB.
“Presence” (high-mid tone)	EQ band 2 (bell, high-mid)	Use parametric EQ, not damping.
Full-bandwidth 10 Hz–18 kHz response	Low Cut ≈ 10–20 Hz, High Cut/Damping ≈ 18 kHz	See EQ shelf section of the BREVERB manual.
Gate (post-effect)	Gate section (Threshold/Hold/Release/Shape)	BREVERB gate is built-in and more flexible.

---

Getting the Frequency Response Right — Without a Separate EQ

This is the part I spent the most time on, because it’s also the part that makes the R7 sound like the R7 and not just “a reverb.”

Pre-delay conversion first, since it trips people up. The R7 sets the pre-delay in digital words at its 40 kHz sample clock, so one word corresponds to about 0.025 ms. The formula is just:

Pre-delay (ms) = words ÷ 40

So 400 words is 10 ms (tight room/drums), 1000 words is 25 ms (a standard hall separation), and 1200 words gets you to 30 ms for a large hall. I just punch these straight into BREVERB 2’s Predelay field, and I toggle the note-sync symbol off so it stays in milliseconds instead of locking to tempo.

Now the actual tone-shaping, done entirely with BREVERB 2’s own controls:

• I set Low Cut to around 10-20 Hz wherever the algorithm exposes it, which keeps the tail’s low end as extended as the R7’s real 10 Hz floor. Hall gives you a genuine Low Cut on top of its Low decay multiplier and crossover; Plate only gives you the multiplier and crossover, no separate cut, so I lean on those instead.
• Damping is the single most important knob in this whole exercise. I set it to around 16-18 kHz so the tail rolls off exactly where the R7 does. That’s what gives you “airy, not fizzy.”
• I keep the High decay multiplier around 0.9-1.0× on Hall and Plate, because the R7 keeps its highs alive just a touch longer than a natural room would, and that’s the bit that reads as “digital sheen” rather than “acoustic space.”

For the R7’s tone controls (its low end and its “Presence” high-mid lift), I don’t add another EQ plugin. BREVERB 2’s own two-band EQ can turn into shelves at the extremes of its Q control: Q all the way down turns a band into a Second-Order Low Shelf, and Q all the way up turns it into a 12 dB/octave High-Pass Filter (Overloud manual). So:

• EQ1, Q at minimum (Low Shelf), 120 Hz, +1 to +2 dB. This is the R7’s extended low-end.
• EQ2, a normal bell around 3.5–5 kHz, +1.5 to +3 dB. This is the “Presence” control, which is really just a high-mid lift.
• I don’t add anything above 18 kHz. The ceiling comes from Damping, not from a boost.

If I had to boil the whole R7 flavor down to one sentence: warm-ish lows, a gentle presence bump around 4 kHz, and a hard-ish ceiling around 18 kHz. Never dark, never brittle.

Parameter Mapping, Algorithm by Algorithm

Once the frequency-response piece was sorted, the rest was just walking through each R7 mode and finding its BREVERB 2 counterpart.

Hall (R7 “HLR”): I run Time at 1.6-2.6 s for a classic vocal hall, Size at 60-80%, and Diffusion pushed up to 70-85%, because that high diffusion is what gives you the dense “90s wash” instead of a sparse, modern-sounding tail. Shape and Spread I set around 40-60% and 50-70% to shape how the envelope builds. R7’s cross pre-delay doesn’t have a true equivalent, so I approximate it with Regen L/R at 5–15% for subtle stereo recirculation. “Rotate High,” the R7’s HF damping control, splits cleanly into BREVERB’s Damping (16–18 kHz) and High multiplier (0.9×). And for low support I run the Low multiplier around 1.0–1.2×, Low Freq around 200 Hz, with Low Cut at 20 Hz.

Plate (R7 “PLR”): This one is shorter and denser. Time 1.2-2.0 s, Size 45-65%, Diffusion up around 80-95% because plates are inherently very dense. Envelope maps to Shape at 30-50%, pre-delay 10-20 ms. HF damping again splits into Damping (16–18 kHz) and High multiplier (0.85-0.95×). The chorus-y movement you hear in R7 plates comes from Motion around 30-45% with Regen at 10-20%.

Room (R7 “RMR”): This is the one place BREVERB 2 does something genuinely different from Hall or Plate: instead of a multiplier-plus-crossover, Room gives you three plain, independent Hz values for High, Damping, and Low, with no multiplier involved. I treat all three as separate settings, never as one merged “damping/high” figure, because they really are three distinct controls in the plugin. For a tight, small Room I run Time 0.3-0.9 s, Size 20-45%, Diffusion 60-80%, and use the Decay (%) control to balance early reflections against the tail (lower percentage = more ER-forward). HF trim lands around 16-18 kHz, HF damping around 14-18 kHz, LF trim around 40 Hz, and a touch of Motion/Depth (10-20% / 10%) for movement.

Gate (R7 “GTR”): This is the R7’s Gate, which is a post-reverb envelope gate. In BREVERB 2, I build the density first with Room or Plate, let the tail run long, and then clamp it hard with BREVERB’s own Gate section (Threshold/Attack/Hold/Release/Shape). If you want the rising, reversed swell instead of a true gate, swap in the Inverse algorithm instead.

Three Starting Presets

Here’s how I’d actually dial these in on a session, algorithm by algorithm.

Tight Room for Drums. I pick Room. Time goes to 0.45 s, Size to 28%, Diffusion to 75%. High diffusion here is what keeps the room from sounding like marbles rattling on a plate, and it’s what lets it thicken a snare or kit without smearing the transient (Craig Anderton on reverb basics). Decay I set to 55%, Predelay to 8 ms, Motion/Depth to 12%/8%. On the Freq page I keep it bright and close: Low at 45 Hz, High at 17 kHz, Damping at 15 kHz. Then EQ1 as a Low Shelf at 110 Hz, +1 dB, and EQ2 as a bell at 5 kHz, +2 dB, Q around 1.0. Gate stays off, or I’ll set a fast 120 ms Hold if I want it snappier.

Standard Hall for Vocals. Hall algorithm, Time at 2.1 s, Size 72%, Diffusion 78%, Shape/Spread at 50%/62%. Predelay is 25 ms, which is about 1000 words on the original R7, and that is enough to keep consonants clear without smearing the front of the word (Beat Kitchen’s mix primer on reverb). I add Regen L/R at 10%/12% for a little cross-feed, and Motion/Depth at 25%/18% for that gentle 90s shimmer. Freq page: High multiplier/Freq at 0.9×/6 kHz, Damping at 17.5 kHz to mirror the R7’s ceiling; Low multiplier/Freq/Cut at 1.15×/200 Hz/20 Hz. EQ1 is a Low Shelf at 120 Hz, +1.5 dB; EQ2 is the “Presence” bell at 4 kHz, +2.5 dB, Q ~1.0.

Gated Reverb Patch. I use Plate here because its Shape and High-multiplier controls give the pre-gate tail more density and brightness than Room would (swap to Room if you’d rather have a plainer character, then just drop the Shape row, since Room doesn’t have one). Time: 1.8 s, which is long and “raw,” because the gate does the actual shaping, with Size 60%, Diffusion 90% for maximum wall-of-sound density, and Shape 35%. Predelay 12 ms, Motion/Regen at 35%/15%. On Freq: High multiplier/Freq 0.9×/6 kHz, Damping 16 kHz, Low multiplier/Freq 1.0×/60 Hz. EQ1 a Low Shelf at 120 Hz, +1 dB; EQ2 a bell at 4.5 kHz, +2 dB for a little bite. I run the Master fader at -6 to -8 dB dry / 0 dB wet, since this one usually sits as a wet-dominant blend rather than a pure aux send. Then the Gate section does the real work: Threshold set so the tail passes on hits and closes around -24 to -18 dB, Attack 0.5-2 ms for an instant open, Hold 80-180 ms (this is really your “gate length” knob), Release 20-60 ms, and Shape set to Sigma with the slope around 70% for a cutoff that’s abrupt without being clicky.

One thing worth remembering about BREVERB 2’s Master section: it isn’t a single wet/dry knob; it’s two independent dB faders: Dry Level and Wet Level. As a rough guide, 100% wet on a send is Dry at -∞ dB / Wet at 0 dB; a wet-dominant insert blend like the gated patch above is roughly Dry -6 to -8 dB / Wet 0 dB; and a mostly-dry halo is more like Dry 0 dB / Wet -10 to -14 dB. Treat those as starting points and trust your ears and your meters over the exact number.

Where the Sound Actually Comes From

Worth restating, now that all the numbers are on the table: of the five rules I listed at the top, the 18 kHz ceiling really is the one that matters most. Roll the top off earlier and it stops sounding like a full-bandwidth Sony hall and starts sounding like a plate emulation (Vintage Digital). High Diffusion and a little Motion do the rest of the heavy lifting; Size matters much less than people expect.

---

The Biosphere Presets

This is the half I actually built the mapping for. Everything above gets you a faithful R7 — bright, wide, full-bandwidth, the way the unit shipped. But that is emphatically not how Geir Jenssen used it. What makes the Biosphere sound is that he took a full-bandwidth reverb and deliberately threw most of that bandwidth away.

What we know and what I’m inferring

The documented part is short. Biosphere’s September 1997 kit list, the Substrata period, names the Sony DPS-R7 as the only reverb in the rack, next to an Akai S-3000 sampler, a Yamaha FB01, a Roland A-33, a Mackie 1202 mixer, and a PowerMac 7600 running MOTU Performer (Modwiggler / Future Music). So essentially every reverberant space on that record went through one box. The chain was minimal too — sources into the Mackie, into the R7, to two-track. “Nothing fancy,” per that same interview reference.

Everything below that is reverse-engineered from listening. There are no published patch numbers, and Jenssen never documented his edit values. So treat these as best-reconstruction settings rather than an archival patch dump. I’m confident in the character; I can’t be confident in the exact numbers, and nobody honestly can.

The signal chain

I set up one 100%-wet aux send and route sources to it, which is how the R7 was used in the first place as outboard gear hanging off the Mackie. Then, and this is the important bit, I let Hall’s own Low Cut and Damping do all the filtering. Hall’s Freq-page Low Cut genuinely stops everything below its setting from ever being reverberated, and Damping rolls the top off the tail. That one-two, done entirely inside BREVERB 2, is the single biggest factor in the distant, weightless quality. Optionally, I’ll put light compression on the return to lift the tail and thicken it, though in the R7 era that would have come from the Mackie’s headroom, not a plugin.

Vast Hall — the Substrata signature

This is the default Biosphere space: long, glacial, filtered, extremely wet.

I use Hall. Time goes to 4.5-7 s (I’ll automate it longer for pads), Size to 85-95% for cathedral-or-valley scale, Diffusion to 80-90% so it’s a smooth wash with no discrete echoes. Shape/Spread at 55%/75% gives a slow bloom rather than an instant onset. Predelay 30-45 ms keeps the source readable inside that huge tail, Regen L/R at 8%/10% adds width, and Motion/Depth at 25%/20% gives the gentle drift that reads as cold air moving.

Then the Freq page, which is where this stops sounding like a normal hall: High multiplier/Freq at 0.6× / 4 kHz so the highs die fast, Damping at 6-8 kHz — nowhere near the R7’s 18 kHz ceiling, and this is the key to the “far away” tone — and Low multiplier/Freq/Cut at 1.2× / 200 Hz / 220 Hz. That Low Cut at 220 Hz is doing the work an external high-pass on the send would normally do, and it’s why you don’t need one. EQ1 as a Low Shelf at 150 Hz, +1.5 dB for slight warmth; EQ2 as a bell at 1.2 kHz, -2 dB, Q ~0.8 to scoop out the “digital honk” and push it further back. Master sits at -∞ dB dry / 0 dB wet (fully wet send, no dry path).

Feel target: a source dropped into an enormous, cold, softly-lit space. Long, dark, diffuse, slightly moving. That aggressive high-frequency roll at 6-8 kHz instead of 18 kHz is precisely what separates Biosphere’s use of the R7 from a bright, showy “90s Sony hall.”

Variations for specific textures

Deep Space Drone for pads and sustained tones, Substrata and Shenzhou style. Hall, Time out at 8–12 s, Size 95%, Diffusion 90%, Shape/Spread 50%/90% so it reads as a drone bed rather than a discrete space. Predelay short at 20 ms — let it merge with the drone. Regen L/R 10%/12%, Motion/Depth 35%/25% for more drift. Freq: High mult/Freq 0.55× / 3.5 kHz, Damping 5 kHz (very dark), Low mult/Freq 1.3× / 200 Hz, Low Cut 40 Hz. EQ1 Low Shelf 150 Hz, +2 dB; EQ2 bell 1 kHz, -2.5 dB, Q ~0.8. Master -∞ dB / 0 dB. I automate Time and send level in and out for the slow-tide dynamics you hear on things like Poa Alpina and Kobresia.

Cold Room for percussive clicks, found sound, sampler stabs. Room algorithm, Time 1.2-1.8 s, Size 45%, Diffusion 75%, Decay 60%, Predelay 10 ms, Motion/Depth 10%/8%. Remember Room’s three separate frequency values: High 8 kHz, Damping 6 kHz, Low 60 Hz. EQ1 Low Shelf 250 Hz, +1 dB; EQ2 bell 2 kHz, -2 dB, Q ~1.0. Master 0 dB dry / -10 to -12 dB wet. This gives short, dark, diffuse ambiance to rhythmic elements without washing them into the big hall.

Metallic Distance — dark, distant, glassy timbres. Plate, Time 3-4 s, Size 60%, Diffusion 90%, Shape 55%, Predelay 25 ms, Motion/Regen 40%/15% for that subtle chorus-y R7 plate movement. Freq: High mult/Freq 0.7× / 6 kHz, Damping 8 kHz, Low mult/Freq 1.1× / 280 Hz. EQ1 Low Shelf 220 Hz, -1 dB; EQ2 bell 5 kHz, +1.5 dB, Q ~1.0. Master -6 to -8 dB / 0 dB. Denser and slightly metallic, but kept dark enough that it reads as distant ice rather than a bright vocal plate.

Ice Cavern — the liquid-metallic tones of Cirque

Cirque is a different animal from Substrata. Reviewers describe a “liquid-metallic sheen” over the music (Touch Records) and “geometric ice formations muffled by snowdrifts,” with rhythms whose reverb envelops you “as though you’re hearing it inside an ice cavern” (Cue Records, Electronic Music Critic). Where Substrata is a pure dark hall, Cirque is metallic, dub-treated, and glassy — it shimmers instead of just receding, and the top end is blurred by dub-style filtering rather than killed outright.

The trick is a Plate for the metallic ring, tuned cold. Time 2.5- 3.5 s, which is long enough to envelop, short enough to stay glassy rather than muddy. Size 55-65% (a cavern, not a cathedral), Diffusion pushed to 92% for that seamless liquid wash, Shape 40% so transients still ring. Predelay 18-25 ms keeps the dub-delayed source distinct from the sheen. Motion at 45% is critical here (that’s the “liquid” movement), with Regen at 20% to thicken the tail.

The Freq page is where this diverges from every Substrata patch: High mult/Freq at 0.85× / 8 kHz, so the highs live longer than the dark hall presets and stay glassy. Damping at 10-11 kHz blurs the very top (the dub haze) while keeping the metal glint. Low mult/Freq at 1.0× / 250 Hz keeps the lows tight so the sub-bass sits under the reverb, not in it. EQ1 as a Low Shelf at 200 Hz, -1 dB thins the low-mids for a colder read; EQ2 as a bell at 6.5 kHz, +2.5 dB, Q ~1.2 is the glassy glint. Master -8 to -10 dB dry / 0 dB wet. Cirque‘s rhythms stay more present than the fully-wet Substrata patches.

Plate has no internal Low Cut, but you still don’t need an outboard one. If a source has real sub-bass rumble, switch EQ1’s Q to maximum at the same frequency and it becomes a genuine 12 dB/octave high-pass filter (EQ1’s gain stops doing anything in that mode, so the –1 dB setting just becomes irrelevant). Hard cut, still entirely inside the plugin.

Feel target: a cold, dub-treated plate that rings like light on ice. Add a short tempo-synced delay before the reverb to properly nail the dub-echo-into-ice-cavern effect on rhythmic material.

Sub-Zero Bass — the subterranean low end of Cirque

Cirque‘s bass gets described as “blurred sub-zero bass” (Touch Records) and bass that “booms like hidden substrata” (Cue Records) — deep, dub-treated, felt more than heard.

Read this one as a routing recipe, not just a reverb patch. Bass reverb that keeps a mix clean is really about a tiny, dark, filtered space plus dub-style low-end movement, not a big wet tail. The rule that matters most: do not send raw low bass into a long reverb. It turns to mud, every time. The R7’s full 10 Hz bandwidth let Jenssen put space around bass without it collapsing; in BREVERB 2 you get there with a short, heavily-filtered room and a mostly-dry blend.

The clever part, and the bit that keeps my no-extra-plugins rule intact: I high-pass the reverb feed inside BREVERB 2 rather than before it. Set EQ1’s Freq to ~130 Hz and its Q to maximum, which turns EQ1 into a 12 dB/octave high-pass filter. Only a filtered harmonic ghost of the bass reaches the wet output — the fundamental never gets reverberated. Technically BREVERB 2’s EQ sits after its reverb engine internally, but the practical result is the same as ducking the send beforehand, and the sub itself still comes through the dry path untouched. For the dub movement I use Room’s own Motion/Depth instead of an external LFO; it modulates the algorithm’s internal delay lines and gets you the same slow breathing quality the old trick got from a 0.1–0.3 Hz auto-filter sweep. Keep it mono and mostly dry so the low end stays centered.

Settings: Room, Time 0.8-1.3 s, Size 35-45%, Diffusion 85% so the low tail is smooth rather than fluttering, Decay 45% (slightly ER-forward so it thickens rather than washes). Predelay 8-12 ms keeps the dry sub punchy. Motion/Depth 20%/15%. Freq: Low 35-40 Hz to let the deep weight through, High 5 kHz, Damping 4 kHz — very dark, essentially no high content in the bass reverb at all. EQ1 as described, the high-pass at 130 Hz. EQ2 a bell at 300 Hz, -3 dB, Q ~1.0 to scoop low-mid mud. Master 0 dB dry / -10 to -14 dB wet.

If it starts booming or smearing the groove, shorten the Time and lower the send — don’t reach for more EQ. One optional authenticity add: a saturation or tape stage before the reverb send. Cirque‘s bass booms partly from analog weight, and a bit of drive generates the harmonics that let a filtered send imply a fundamental it isn’t actually carrying.

Dub-Blurred Rhythm — the obscured beats of Cirque

Reviewers keep circling this: Nook & Cranny “lazily lopes along with a dubbed-out rhythm,” Le Grand Dôme is “obscured by its dub treatments,” and even the brisk drum-n-bass of Algae & Fungi is “strangely obscured by their distant tone” (Electronic Music Critic). The album overall runs on “muffled beats,” with drums kept firmly in the background (Wikipedia).

This is genuinely different from putting a room on a drum. It’s delay into reverb — real dub technique — plus heavy filtering, so hits smear into a receding cloud instead of decaying cleanly. The dub part comes first: a tempo-synced eighth or dotted-eighth delay with 40-60% feedback, and then I send the delay’s output, not the dry hit, into BREVERB 2. That’s what makes the rhythm feel like it’s happening somewhere else. Then I automate or duck the send so the reverb swells in the gaps between hits and recedes on the beat.

For the reverb itself: Room, Time 1.5- 2.2 s (long enough to blur into a cloud, short enough not to bury the next hit), Size 50- 60%, Diffusion 88% so individual repeats melt together, Decay 35% — tail-forward, de-emphasizing the transient. Predelay near zero at 0-5 ms so it merges immediately with the repeats. Motion/Depth 30%/20% for the underwater wobble. Freq: High 5 kHz, Damping 6-7 kHz, Low 80 Hz so kicks and toms still read as present. EQ1 Low Shelf 150 Hz, +1 dB; EQ2 bell 2.5 kHz, -2.5 dB, Q ~1.0 to scoop the attack-y upper-mids and push hits back. Master -6 to -8 dB dry / 0 dB wet. The dry transient stays a click; the smear does the distance work.

This is the only preset here where the delay stage matters as much as the reverb. Skip it and use a room alone and you’ll get “reverb on a drum,” not dub-blurred rhythm.

Two Shenzhou Presets

A caveat before these. The R7 is only confirmed in Biosphere’s rack as of that 1997 kit list. Shenzhou came out in 2002, and no gear list or interview names the reverb on it. Given how long Jenssen keeps gear — he’s mentioned still opening 1993 session files and still using 1970s synths (Headphone Commute) — it’s plausible the R7 was still in service, but that’s my inference, not a documented fact. These are built from the documented sound of the record.

Shenzhou is built almost entirely from looped, pitch-shifted fragments of Debussy’s La Mer and Jeux — hyper-minimalistic, largely just a core sample loop plus reverb, deep rumbling bass, and occasional faint percussion (Wikipedia, Touch Records, Brainwashed). When I replicated the title track a few years back, what surprised me was how ordinary the reverb turned out to be — a fairly plain orchestra hall. The character was coming from a tilt EQ centered around 230 Hz and a low-pass filter swept toward its dark end on the sampler, not from anything exotic in the reverb.

Debussy Loop Hall. The good news is that the tilt EQ and the filter sweep are both reproducible inside BREVERB 2 with no extra plugin. The tilt becomes EQ1 as a Low Shelf (min Q, +3 dB at 230 Hz) paired with EQ2 as a bell (-4 dB at 4 kHz, ordinary Q — a cut that broad doesn’t need anything extreme), and the low-pass sweep is already covered by tuning the Freq section dark.

Hall algorithm, Time 2.0-2.8 s — enough space to feel orchestral without going endless like the Substrata presets. Size 65-75%, Diffusion 75% (moderate; Shenzhou is less washed than the other records, more like a worn recording), Shape/Spread 45%/55% for a gentle unhurried bloom. Predelay 20-30 ms preserves the vinyl-sample transients. Regen L/R 12%/15%, Motion/Depth 10%/10% — minimal, because this album runs on stillness rather than shimmer. Freq: High mult/Freq 0.55× / 3.5 kHz, Damping 4-5 kHz (a darker ceiling than even the Substrata hall), Low mult/Freq/Cut 1.15× / 220 Hz / 30 Hz. Master -3 to -4 dB dry / 0 dB wet — less drenched than the pure ambient patches, because this is a sample loop with reverb on it, not a wash. Add light vinyl crackle for authenticity; the darkening itself is already handled.

Deep Current. Underneath those loops, reviewers describe “low end currents,” “ominously rumbling bass notes,” and “faint echoing drumming” (Brainwashed, EvilSponge). It’s a companion to the Cirque sub-bass patch, but darker, slower, and less rhythmic — a drone rather than a groove.

Room, Time 1.5- 2 s, Size 40%, Diffusion 80% for a smooth, non-fluttering rumble, Decay 30% (tail-forward). Predelay 15 ms for slight separation on the faint percussion. Motion/Depth 15%/12% — barely perceptible drift. Freq: Low 30 Hz to let the true sub current through, High 3 kHz, Damping 2.5 kHz, which is extremely dark; almost no top end reaches this layer. EQ1 Low Shelf 70 Hz, +2.5 dB for the rumbling weight; EQ2 bell 500 Hz, -2 dB, Q ~1.0 so it doesn’t clash with the loop’s own low-mid body. Master 0 dB dry / -12 to -14 dB wet — felt as pressure, not as a spatial event.

The Five Rules That Make It Sound Like Biosphere

If you remember nothing else from Part 2, remember these.

1. Filter hard — but inside BREVERB 2. Hall’s own Low Cut to 200-250 Hz (or EQ1/EQ2 at maximum Q as a high-pass on Plate and Room), and let Damping handle the top roll around 5-7 kHz. That weightless, band-limited tail is easily 50% of the sound, and you can build all of it without a plugin in front of the reverb.
2. Kill the highs in the tail too. Damping at 5-8 kHz, not the R7’s native 18 kHz. This reverb is dark and distant on purpose. It is deliberately not using the unit’s full brightness.
3. Go very wet, and keep the source either dry-forward or very low. These are mostly wet ambiances on a send.
4. Long, high diffusion, slow Shape. 4.5 s and up, 80-90% diffusion, slow bloom — no discrete echoes, just an evolving field.
5. A little Motion/Depth. 20-35% adds the cold, drifting air that keeps long tails alive instead of static.

Sources

My BREVERB 2 presets: LL Presets

R7 factory-preset impulse responses (for authentic capture rather than modeling): Sony DPS-R7 IRs

Sony DPS-R7 review and full parameter walkthrough: Sound On Sound, Dec 1991 (muzines archive), Tape Op review

DPS-R7 specifications (18-bit, 40 kHz, 10 Hz–18 kHz): Retrosonic Pro Audio, Vintage Digital

Overloud BREVERB 2 parameter reference: Overloud BREVERB 2 User Manual (PDF)

Reverb parameter theory (diffusion/density/damping/pre-delay): Craig Anderton, Beat Kitchen

Biosphere’s 09/97 studio kit list, confirming the DPS-R7 as his only reverb: Modwiggler forum / Future Music

Ambient/dark-hall reverb technique (filtering, high-cut, wet sends): Valhalla DSP, Music Production Wiki

Cirque sonic character (liquid-metallic sheen, ice-cavern reverb, sub-zero bass, dubbed-out rhythms): Touch Records, Cue Records, Electronic Music Critic, Wikipedia

Shenzhou production and sound character: Wikipedia, Touch Records, Brainwashed, EvilSponge, Electronic Music Critic

My own ear-matched Shenzhou breakdown (tilt EQ, filter, hall reverb): Lars Lentz Audio

Jenssen’s long-term gear retention habits: Headphone Commute studio interview

The Sony DPS R7 manual: