The HEDD Lineariser® or
'A Brief History of Time'
Subwoofers are in use since many decades, their integration into the reproduction chain allows to add fundamental frequencies where small speakers reach their limits. However, the use of separate bass units often shows two typical unwanted side effects: the woofer section appears to be separated or not really integrated into the overall sound image, and the low frequencies themselves tend to sound somewhat boomy or smeared. Before we started to build subwoofers we took a closer look at the reasons for these widely experienced problems. The answer we found is: time.
Already in single loudspeakers one can observe that lower frequencies need more time to travel through the speaker than higher frequencies. In Satellite-Sub systems additional time shifts occur at the crossover points and the involved high-and lowpass filters. All in all some timely confusion is added in the reproduction chain by loudspeakers.
As we deal with sound a principal physical remark has to be made: sound waves are completely characterized by time and frequency. Between these two there is a fundamental physical relationship called the Harmonic Analysis. It states that each change in the frequency response of a loudspeaker inevitably causes a corresponding change in the Phase Response.
But what is a “Phase Response”?
A phase describes the timely distance between two events, nothing more. So if a tone starts at exactly 12 pm and another one at 12:01 pm the phase delay is 1 minute. As a consequence of the Harmonic Analysis phase delays in loudspeakers necessarily occur due to the nonlinearities in their frequency response. These delays range from <1 ms (Millisecond) to approximately 50 ms, depending on the transmitted bandwidth and some DSP internal settings.
Within a subwoofer all frequencies between 20 Hz and typically 80 Hz should be reproduced at the same point in time of course, but as mentioned due to internal filters and driver properties delays of different durations are added at different frequencies.
Lower frequencies travel longer through speakers than higher ones
In other words: The phase response of a subwoofer falls from higher traveling times - like 50 ms at ~20 Hz – down to 10 ms at ~80 Hz, which has become the standard crossover frequency between subs and satellites. With respect to frequency these different delays vary in a characteristic way for each subwoofer, they are represented in a curve called phase response. In analog times there was no way to avoid these phase errors.
With today's powerful DSP it is possible to apply FIR (for Finite Impulse Response) filters that are able to correct these deviations, but they only can achieve it on cost of an overall delay for the signal passing through. To do the job the filters need to know what's wrong in the time and frequency domain of the speaker. Technically speaking they need the measured Impulse Response (IR) of the subwoofer under consideration.
The IR's of BASS 08 and BASS 12 have been measured in the big anechoic chamber of the Technical University Berlin and implemented in the DSP boards to deliver that information. It contains all their frequency and phase nonlinearities, which now can be corrected in real time. The HEDD Lineariser®, an adapted FIR (Finite Impulse Response) filter is the mathematical tool of choice that reinstalls the original timely relationships of the incoming musical signal. The audibly improved transient response and a better focused stereo localization are evident benefits.
Having done this the phase response in the subwoofers now is linear, but as mentioned this has to be "paid" by an overall delay of ~50 ms for a down to 20 Hz (- 3 dB) frequency response. The DSP has to wait these 50 ms before the sound can be put through to the power amplifier and out of the box. As a consequence all frequencies between 20 and 80 Hz are delayed by 50 ms, it is called a group delay of 50 ms. It means the sub starts the low frequencies 50 ms later compared to the sound of the satellites. The HEDD MK2 speakers have a Lineariser® too, in their case - because they their response is set to >80Hz - their group delay is 10 ms only.
Eliminate inhomogeneous and boomy sound in Satellite-Sub systems
To solve the timing problems completely, we implemented two more DSP channels that add a 40 ms group delay to the analog “Outputs to Satellites” XLR sockets at the backplates of the BASS 08 and BASS 12 subwoofers.
What this means is, that the Satellites now "wait" these 40 ms until the subwoofer emits its low frequency signals. Then - if the satellite Lineariser is ON - add their own group delay of 10 ms - and well, here we are: 50 ms later, all 3 loudspeakers start reproduction at the very same millisecond, and they do it for each and every frequency in the complete audio band.
One more thing: In practice it often is difficult to set up satellites and subwoofers at the same listening distance. As only with equal distances the above considerations really work we have added a distance control called “SAT. OFFSET [m]” at the back panels of the subs. It allows to compensate the potential difference in traveling time between satellites and subwoofers if necessary, it actually would mean a new additional phase error. You may adapt to a ± 2m runtime difference between satellites and subwoofers in a .5 m resolution, that should cover the problem in most of the cases.
All components of HEDDs Linear Phase Satellite-Sub systems start the sound reproduction at the same point in time, and do so with the timely correct relationships within the audible band as to be found in the incoming music signal.
With the combination of HEDDs Linear Phase satellites, BASS Linear Phase subwoofers and an appropriate compensation for the longer group delay of the subwoofers, we actually have wiped out the standard problems that have been inherent in the use of subwoofers right from the beginning - until now.
We think we have found and eliminated the main reasons for inhomogeneous and boomy sound in Satellite-Sub systems. We at HEDD would be happy if you take the time for an audition, the proof as we do know is in the listening.
