Why FAR digital?

Introduction

The FAR Active monitor™ loudspeakers, created and manufactured by ATD2, are using the most advanced audio digital technologies

The signal processing is entirely achieved in the digital domain by 2 DSP (Digital Signal Processor). These monitors were designed to accept the IEC60958 digital audio streams (AES/EBU – SPDIF).

By this resolutely digital perspective, did we have to treat the analogue input like a poor relation? Certainly not! The objective was to propose a digital monitor able to compete with the best analogue systems while offering functionalities and ergonomic impossible to be obtained with a classic analogue technology.

Block diagram of the monitor

The XM range proposes compact multi amplified 3 way monitors using the same medium-tweeter CAC™ (Central Acoustic Coherence) configuration which is an invention of ATD˛. The AV range includes 2 or 3 way multi amplified models. Both series share the same electronics and by consequence have the same functionalities.


The monitor is configured by a control panel located at the back of the monitor (an alphanumerical LCD plus a keypad) or by the FAR LINK ™ network by using a remote control or the PC SCS software developed by ATD2.

Function blocks

The following scheme describes the chain of the digital processing applied to the signal by the monitor.


Digital Domain <-> Analogue Domain

Analogue to Digital conversion

The balanced analogue signal (input by XLR Combo connector) transits through a programmable gain amplifier which allows adapting the analogue full scale sensitivity of the monitor to all the analogue levels commonly encountered. For example, the “broadcast” standard imposes a level limited at +6dBu with a reserve of +9dBu. The sensitivity of the analog entry will then be +15dBu at full scale. For weaker signals, we will choose a sensitivity of +9dBu or even +3dBu. To adapt to the mixing tables capable to output very high levels, we will select +24 dBu.

The analog signal is then filtered by a low pass filter (Bessel type) to attenuate the very high frequencies (>1MHz) before going in the digitizer (ADC).

The quality of the analog chain and the AD conversion is such that we obtain at the digital output of the converter a total harmonic distortion + noise (THD+N) ratio lower than 0.0007% (<-103 dB) for a signal level of -0.5dBFs (at all frequencies).

At this stage, it is interesting to compare these performances with a purely analog circuit. A TL072, an operational amplifier very often used in electronics of active monitors and very carefully brought into operation, will give as output from a balance to unbalance converter circuit a total harmonic distortion + noise ratio of about 0.003% (-90 dB) for a level of signal of -0.5dBFs, that means performing 13dB less than the analog input of FAR monitors.

Digital to Analog conversion

After the digital processing in the DSP, the signal is converted back to the analogue domain. The converters are of a very high quality and performances are matching.

The total harmonic distortion + noise (THD+N) ratio is lower than 0.0008% (<-102 dB) for an amplitude signal of -1dBFs (at all frequencies).

The output filter of the digital to analogue (DAC) converter is of a second order Bessel type and has a cut off frequency equal to 80 kHz.

The analog filtering in input as well as in output has been designed in order not to induce a variation of group delay in the audio band and till 40 kHz. A flat group delay is the first key to achieve great signal dynamics, in particular with a sampling rate frequency of 96 kHz.

This is worth noting that all analog signals are systematically balanced to guarantee a very high immunity against external electromagnetic perturbations. For example, a mobile phone won’t generate any buzz in any FAR Active monitor™ loudspeaker when starting ringing.

One of the reasons why the conversion is of such a high quality is because of all the special care brought to the clock circuit which has a very low phase noise (jitter <  20ps) to guarantee the best possible conversion. The 2 converters (ADC and DAC) are synchronized on the same clock. This is demonstrated at the listening by an outstanding stability of the sound stage and by the most realistic reproduction of the reverberations.

The total harmonic distortion + noise ratio of the entire chain (ADC+DSP+DAC) is lower than 0.001%.

Digital Signal Processing

The signals entering in the DSP have a 24 bits resolution and a sample rate frequency of 96 kHz (supplied by the Asynchronous Sample Rate Converter - ASRC).

The fixed point DSP has a 48 bits resolution (to be compared with the 24 bits resolution of the input signal). The theoretical dynamic range of the processing is about 288 dB…


The quantization errors that inevitably happen due to the discrete representation (quantified on 48 bits) of the data to be processed, are relegated into the first 8 least significant bits of the 48 bits word (precision/noise bits).

The multiplication / accumulation operations (MAC) are made thanks to a 76 bits accumulator. This means that all operations done in the DSP are made without any loss of dynamic neither of resolution.

In practice, the level of distortion induced by the digital processing within the DSP is inferior to -192dB.

When the 2 DSP have finished their calculations, they output the samples towards the Digital to Analog Converters (DAC).

These output samples are necessarily coded on 24 bits. Signal normalization is then a prerequisite condition. The transition from 48 bits to 24 bits inevitably leads to a truncation error at the level of the least significant bit (LSB) of the output sample. This is why a pseudo-random noise (dither) is applied on this bit in order to spread on the whole audio spectrum the effect of the truncation distortion and to make it completely inaudible.

The resulting dynamic range of the signal which goes through the DSP is about 138 dB (23 effective bits).

By comparing the values of the total harmonic distortion + noise ratio of the converters and the DSP, one can notice that the DSP itself induces no significant and even less audible worsening of the signal. The quality of the Analog to Digital conversion is then entirely preserved by the implementation designed by ATD2.

Of course, the final quality highly depends on the way the signal is processed in the DSP. For example, a decrease of the signal of 96dB followed by an amplification of 96 dB will not result in a null operation. The sample will loose at least 8 bits of resolution.

The experience of ATD2 in the digital signal processing applied to professional loudspeakers guarantees the preservation of the resolution and of the dynamic of the signal during all the calculation.

The different functional blocks implemented in the DSP are :

  • Delay and relative level of the monitor for the optimum setting of a 5.1 system
  • Equalization of the monitor (flat frequency response anechoic chamber)
  • 10 bands full parametric equalizer
  • Preset curves (each of them having 8 fully configurable filters)
  • 3-way cross-over with specific treatment for the woofer in closed volume.
  • Time alignment of the different divers (woofer, medium and tweeter) to guarantee an almost perfect impulse response.

Let’s compare now a similar implementation made with classic analogue components (operational amplifiers, resistors and capacitors). Beside the fact that the purely analogue electronic is far less flexible than its digital equivalent, the precision of the filters are directly linked to the tolerance of the passive components (often of about ±5%, at best 1%).

With an analogue design, each component will add its sound signature to the signal which goes through. For example, a capacitor is everything but transparent. To less extent, it goes the same for the operational amplifiers. Above that, the noise of each active component adds up, which damages everything along signal chain.

The FAR monitors are then as neutral as it is possible to be.

The large number of active cells required to obtain a result not even close to the functionalities of the FAR monitors would impose consequent analogue electronics that would be very difficult to fit in a monitor.

Preset curves

The SCS control software allows, via the FAR LINK ™, to define and to save in the monitor up to 7 different curves , each featuring 8 filters from the 1st or the 2nd order.

These curves can be recalled using the control panel of the monitor or via the remote control. The main use of these curves is to simulate the acoustic result at the final user premises.

The following example shows a frequency response simulating a TV set. Depending on the content, it is possible to adapt the acoustical result of the monitor to get an accurate idea of the result at the end user. The monitors are programmed at the factory with 4 curves. It is however possible to modify them in order to create new ones.

The « FLAT» curve can not be modified. It is used as reference for studio mixing.


10 band full parametric equalizer

The FAR Active monitor™ loudspeakers feature a 10 band parametric equalizer.

The configuration can be done by means of cursors and text entries or via a graphical and ergonomic interface that allow making a «visual» setting. It is possible to configure 7 sets of 10 filters and to store them into the monitor.

These equalization settings can also be recalled using the control panel of the monitor or via the remote control.

The « FLAT » equalization setting cannot be modified. When the “FLAT” curve and the “FLAT” equalization setting are on, the monitor by itself has a flat frequency response.


Room Acoustics Correction

The SCS control software proposes a tool that helps the optimization of the acoustic of the listening or working room. It is possible to load in the SCS software the frequency response of the loudspeaker in its working room and to “flatten” it manually by using the 10 band parametric equalizer.


Even if this method has the advantage of leaving the full control of operations to the technician or sound engineer, it requires a certain experience to obtain an optimal result.

ATD2 has developed an assistant for the optimization which offers (but not imposes) a solution to the user.

You just need to give a model in which the frequency response of the pair monitor / room must be found. The model must not necessarily be flat, even if it is in most of the cases the searched aim. Any arbitrary form is impossible.


The optimizer will search the parameters of filters which lead to an optimal solution. The user can than judge the effect of each filter by using the “OFF” button which is associated to it. The user can then modify the parameters of each filter if the proposed solution does not please him.


Once the satisfactory result is obtained, the user can then store it into the monitor with an appropriate name.

Loudspeaker Control System.

Beside the major functions such as the preset curves and the 10 band parametric equalizer, the FAR Active monitor™ loudspeakers propose an environment of integrated control.

Via the FAR LINK ™ network, it is possible to control and configure up to 7 loudspeakers and 2 subwoofers.

The tones controls (tilts & shelves) as well as the choice of source, the analogue full scale sensitivity, the choice of digital channel, the delay applied to the loudspeaker and the relative level are directly accessible.


The delay and gain of each driver are accessible in the tab « Xover Settings » in order to compensate an inclination of the loudspeaker which is not perfect with the listening position. It is also possible with these settings to decrease the effects of the first reflection on the mixing desk. However, it is there a powerful tool reserved to an experienced user that can very well control the acoustic of the room.

Conclusion

This document is made as an explanation and relates the major functions of our monitors.

For more technical explanations or more specific points, do not hesitate to contact Mr. Xavier Lambrecht (xavier.lambrecht@atd2.com).