The points below are intended to explain a little about how Peak works with respect to noise. Peak’s output does contain some noise as it has a true analogue signal path. Peak’s noise is comparable to other synths in its class and its signal output compared to this noise is still very powerful.
- Peak’s headphone output is loud.
It was designed to deliver enough output level to be able to be used in a gigging environment. Low impedance headphones will be louder still. Hiss will be audible if you turn the master volume all the way up but at this level the patch volume will extremely loud.
- Output level select.
Firmware version v1.1 (262.255) introduces an attenuation that can be selected. This will affect both main outputs and headphones.
- Press Settings
- Then press the Page/Select buttons to view page 8/9 ‘VOLUME RANGE’
- Set this to -6dB, -3dB (default) or 0dB
Setting this to -6dB will reduce the level of the main outputs and headphone out
We have measured Peak’s THD+N using a single voice at 987.77 Hz. We also carried out the same test on the DSI Prophet 12 for comparison.
- Peak THD+N = 0.32%
- DSI Pro 12 THD+N = 0.62%
Our test procedure for measuring noise is:
- 1 voice playing
- 1 Osc set to sine tone playing a B4 (987.77hz)
- All gain stages up to max or just before clipping
- Take the measurement of peak output
- Use an AES compliant filter to remove the fundamental sine
- Signal left is THD+N which can be turned into a percentage using the following calculation http://www.sengpielaudio.com/calculator-thd.htm
4. Signal to Noise Ratio
We don't publish SnR figures for Peak because with all the different wave types and gain stages that Peak has it is difficult to define what the true maximum output is.
5. The analogue filters will introduce the most noise into the signal path. Turning down the oscillators and pressing keys will open the VCAs allowing the filter noise to be heard. There are 8 filter circuits so 8 keys will allow all the noise from these filters to be heard.
6. In firmware v1.0 we found a bug with the oscillator that meant that some voices would generate additional digital noise. This was found and fixed in v1.1