EAW UX8800 User Manual download pdf (Page 6)

Eastern Acoustic Works One Main Street Whitinsville, MA 01588 tel 800 992 5013 / 508 234 6158 fax 508 234 8251 www.eaw.com

EAW products are continually improved. All specifications are therefore subject to change without notice. Part Number: Preliminary UX8800 April 2007

UX8800 Preliminary Specifications Digital Signal Processor

REAR PANEL - INPUTS / OUTPUTS

NOTES

TABULAR DATA

1. Measurement/Data Processing Systems: Primary - FChart: proprietary EAW software; Audio Precision. ??

2. Measurements: Dual channel FFT; length: 32 768 samples; sample rate: 48 kHz; logarithmic sine wave sweep. ??

3. Measurement System Qualification (includes all uncertainties): Level: accuracy +/-0.05 dB 20 Hz to 20 kHz, precision +/-0.1 dB 20 Hz to 20 kHz, resolution 0.01 dB;

Frequency: accuracy +/-1 %, precision +/-0.1 Hz, resolution the larger of 1.5 Hz or 1/48 octave; Time: accuracy +/-10.4 µs, precision +/-0.5 µs, resolution 10.4 µs. ??

4. Volts/Amperes: Measured rms value of the signal or as noted.

5. Performance: Input, DSP (Digital Signal Processing), outputs, and ac mains characteristics.

6. Functions: Operating controls, function parameters, and indicators.

GRAPHIC DATA

1. Graphs are plotted using raw data.

2. Frequency Response: Variation in output level with frequency for a constant input signal.

3. Phase Linearity: The difference in phase between the input signal and output, with signal processing latency removed.

UX8800 TECHNOLOGY

UX8800 Design Challenges: The goal for UX8800 digital signal processor was to provide the standard set of digital

processing tools in a 4 input by 8 output processor. In addition, the processor’s hardware had to be able to implement

Gunness Focusing™ in order to provide preset, factory-optimized processing for specific EAW loudspeakers. This would

include those intended for standalone use and those normally used in arrays. The preset processing allows unused input

and output channels to have standard processing for other uses.

In addition to its functions as a digital processor, the UX8800 needed to be network capable using standard Ethernet protocol

and hardware. However, using the EAWPilot software, Ethernet communication would need to be plug-and play to avoid

troublesome IP addressing issues. In order to facilitate control of and connect audio to multiple additional products, EAW

designed a new networking topology called U-Net. U-Net’s 32 channels allow a combination of control signals and audio

signals to be networked between products over standard CAT-5 cabling. Designed as a mesh network, no specific wiring

configuration is required as is for Ethernet, CobraNet, and other popular network topologies used for audio signals and

equipment control.

Gunness Focusing™ : Using innovative analysis tools and methods developed by EAW, specific, long-standing,

loudspeaker problems were isolated and analyzed as to their solutions using DSP. However, the desired complex filter

responses required accuracies grossly lacking in conventional DSP filters based on the Bilinear Z Transform (BZT). These

filters sacrifice response accuracies in the upper audible octaves to avoid mathematical difficulties involving the Nyquist

frequency. Using standard algorithms can result in filter response magnitude errors of over 15 dB, with equally flawed

phase performance. Using FIR (Finite Impulse Response) filters would have resulted in latencies in excess of acceptability

for real time use. Gunness Focusing uses EAW-developed filter algorithms that avoid these issues while providing the

exact, complex, filter responses required to correct the loudspeaker problems.

Gunness Focusing is implemented in the UX8800 for many EAW loudspeaker models. Of particular note are EAW array

loudspeakers. Gunness Focusing not only optimizes the performance of the individual loudspeakers, but it also optimizes

the integration of adjacent loudspeaker outputs and the off-axis performance of entire arrays.

Gunness Focusing Example: Figure 1 shows the spectrogram of an ideal, point-source loudspeaker. Figures 2 and 3

show two different spectrograms of a 2-way loudspeaker optimized with conventional digital signal processing (DSP) and

conventional measurements. In Figure 2 the time domain performance is emphasized. In Figure 3 the frequency domain

performance is emphasized. In both cases there is significant energy to the right of the main energy spectrum compared to

the ideal loudspeaker shown in Figure 1. These are all caused by inherent, mechanical properties of both the cone LF

driver, the HF compression driver, and the HF horn itself. Although the frequency response (not shown) is nearly an ideal,

flat line, these anomalies obviously exist in spite of the conventional processing. Because the usual measurements and

corrective filtering lumped the undistorted signal and the anomalies together, the flat response is actually a combination of

the energies from both. Anomalies like these are generally described as coloration and are responsible for why two, similar,

flat-response loudspeakers can sound quite different.

In contrast, the result of applying Gunness Focusing to this same loudspeaker is shown in Figure 4. The anomalies in both

time and frequency are largely gone, making the spectrogram in Figure 4 look quite similar to that of the ideal loudspeaker

in Figure 1. While the frequency response is also nearly an ideal, flat line, it is almost entirely a result of reproducing the

energy from flat input signal.

Summary: EAW's engineering efforts resulted in the UX8800 digital signal processor which provides complete user

control as well as factory-optimized, Gunness Focusing settings for EAW standalone and arrayed loudspeakers. User-

friendly, advanced, processing functions and plug-and-play audio/control networking facilitate its use for first-time DSP

users, seasoned professional operators, system designers, and audio aficionados all over everywhere. In keeping with the

performance of EAW loudspeakers, the UX8800’s sonic performance is superb.

SPECTROGRAMS: EAW's proprietary spectrograms show the spectrum or frequency content of sound (vertical axis) and its variation in time (horizontal axis), the colors

representing intensity. The width of the data reflects the size of the sliding time window applied to the data, which increases in size with lower frequency. The “data” in the

upper right is simply a limitation of the spectrograph's mathematics and has no relevance.

Figures 1 to 4 (top to bottom)