Ohmic Audio Labs Knowledge Base

12.1 DSP Platform Selection

🔰 BEGINNER LEVEL: Do You Need a DSP?

When a DSP Is Worth Adding

A DSP (Digital Signal Processor) sits between your source unit and amplifiers. It intercepts the audio signal in the digital domain, applies all your processing — crossovers, equalization, time alignment, level control — and outputs a corrected signal to each amplifier.

You should seriously consider a DSP if:

You may not need a separate DSP if:

Illustration note: DSP signal chain block diagram from input through processing blocks to outputs

Entry-Level Options Overview

miniDSP 2×4 HD — $100

Two stereo inputs, four outputs. 24-bit/96kHz. USB programmable from a PC or Mac. Provides parametric EQ (10 bands per channel), Butterworth/Linkwitz-Riley crossover filters (1st through 8th order), time delay, and output level. Excellent value for a simple system.

Limitation: Only two inputs — won't work as-is with a 4-channel head unit without mixing. No high-level (speaker-level) inputs — requires head unit with RCA outputs.

Helix DSP Mini — $280

2 inputs, 6 outputs. Higher quality internal DAC than miniDSP. More filter options. Compact form factor. Good for 3-way front + subwoofer.

Helix DSP — $450

6 inputs, 8 outputs. Accepts both line-level and high-level inputs. Vehicle-specific integration for factory head units. Used in most professional SQ builds.

For the most common builds:

Two-way + sub: miniDSP 2×4 HD works well. Three-way + sub: Helix DSP Mini or miniDSP C-DSP 6×8. Four-way active + sub: Helix DSP or better.

🔧 INSTALLER LEVEL: Platform Comparison and Configuration

Selecting by System Architecture

Map your system before selecting a DSP:

Count inputs needed: - Stereo head unit with RCA: 1 stereo pair (2 inputs) - Factory head unit without RCA: use high-level input; count speaker pairs used (usually 2 pairs = 4 inputs) - Factory head unit with separate center/sub channel: may need 3 pairs

Count outputs needed: - Each individually driven speaker = 1 output channel - Tweeter (L, R) = 2 channels - Midrange (L, R) = 2 channels - Midbass woofer (L, R) = 2 channels - Subwoofer (mono) = 1 channel - Total 3-way + sub: 7 channels (requires 8-output DSP minimum)

Feature checklist:

Feature Why It Matters Min. Requirement
Parametric EQ per channel Room correction, driver response correction 6+ bands per channel
Crossover filter type Butterworth, Bessel, LR options LR12/24 minimum
Crossover filter order Steeper = better driver protection 24 dB/oct minimum
Time delay per channel Critical for imaging 0.01 ms resolution
High-level inputs Factory integration Required for factory HU
FIR filter support Phase-accurate crossovers Optional but premium
Software quality Stability, usability Mac/PC with active development

Configuring Input Signal Level

Test procedure:

  1. Play a 0 dBFS (full-scale) test tone from head unit
  2. Set head unit to reference volume (typically 75% of maximum)
  3. Open DSP software, observe input meters
  4. Input level should reach −1 to −3 dBFS on the DSP meter
  5. If below −10 dBFS: source level too low → increase head unit output or reduce DSP input attenuation
  6. If above −1 dBFS (clipping indicator): source level too high → reduce head unit output or increase DSP input attenuation

The goal: Use as much of the DSP's dynamic range as possible without clipping.

⚙️ ENGINEER LEVEL: DSP Architecture and Signal Path

Fixed-Point vs Floating-Point Processing

Fixed-point DSPs (older, budget platforms):

Process audio as integers with fixed decimal position. Common word lengths: 16-bit, 24-bit, 32-bit.

Advantage: Fast, low power, predictable. Disadvantage: Limited headroom for internal calculations. With 24-bit input, a +24 dB gain stage followed by −24 dB would require 48-bit intermediate precision — most fixed-point systems clip internally at 24 or 32 bits.

Double-precision accumulation: Better fixed-point DSPs use 48-bit or 64-bit accumulators for intermediate calculations even with 24-bit I/O. This prevents internal overflow during EQ operations.

Floating-point DSPs (modern platforms, SHARC and similar):

Process audio as 32-bit or 64-bit floating-point. Dynamic range effectively unlimited within the representable range. Internal overflow essentially impossible with reasonable gain structures.

Advantage: No overflow concern, full precision throughout processing chain. Disadvantage: Slightly more complex hardware, higher power consumption.

Practical implication: Modern DSPs (miniDSP HD, Helix, Alpine) all use floating-point or equivalent precision internally. The user should not worry about internal precision for these platforms — set gains reasonably and the math works correctly.