APC400R Studio acoustic processor

The world's first high-order real-time FIR processor on FPGAs

The FIR filter is characterized by a linear phase that adapts to the unique requirements of audio - that is, changing the frequency domain without causing phase changes. However, due to the characteristics of the DSP serial processing, high order of the FIR filter cause more latency, these latency bring new phase problems. so the high-order FIR filter Due to DSP performance problems, can only stay in the theoretical stage. Even due to the complexity of processing algorithm coefficients, some manufacturers even use the PC's CPU to calculate the coefficients and FIR processing, this will result in greater system latency.Thus, we first introduced high-performance FPGA parallel processing, high-order FIR parallel processing,  to achieve zero latency performance, to fix phase error in your acoustic.

solve all the space phase problem in 20 seconds

The traditional solution to the problem of acoustic space is quite complex, and system engineers need to use different measurement and processing tools to solve different problems. For example, the placement of multiple subwoofer in the room, through multiple measurements, adjust the placement, Set different delay time, so that subwoofer can play in the same phase, but also take into account the subwoofer and the whole time of the time, so complex settings, often trade-off, can not get relatively correct results. At the same time the room and the speakers need to take into account not only the bass, in the bass zone occurred in the room wave, the midrange frequency division distortion, high frequency in the room early reflection, the system engineers physically and mentally exhausted, spend a lot of time After debugging, only a riddled sound system to the tuner, the tuner can only be based on a defective system in the tuning repeatedly solve the system problems, there is no time for music balance. This is why the false singing is prevalent in the studio of the major television stations.And we designed a unique algorithm only need a room or a large space in a number of points for microphone measurement, the time of each measurement point only three seconds, for the noise in space has a great tolerance, and then The algorithm will take ten seconds to discover the phase problems that exist in the system, and automatically send the processing parameters of the problem to the FPGA. This series of processes, without human intervention, only need to support the microphone, press the measurement key on it. The tuner will be able to get a standard acoustic space after 10 seconds. In this space, there is no bass bass caused by the cancellation of the subwoofer, nor is it low frequency resonance, no annoying feedback whistling. Because our intelligent algorithm has been well aware of the sound field, the tuner can even through our frequency response curve design to design the overall frequency of music frequency.It all takes only 10 seconds.

The universal algorithm can be applied to a variety of chip platforms

For the real-time implementation of multi-channel and multi-stage filter, we apply the method of multi-thread parallel processing in the field of high-speed processing by using the method of computing and processing separation, and the high-order filter processing is reduced to beyond the multi-track audio processing Delay, multi-stage filter, multi-channel parallel processing possible. After understanding the consumer application, the number of these applications channels do not need too much, but more sensitive to the cost price. Based on the parallel audio algorithm implemented by FPGA, we fabricate the DSP core of small parallel processing and integrate it into the ARM architecture to form an integrated audio application using ARM instruction and DSP core module parallel processing. Platform -HIMS (High speed instruction Multi-stream). The cost of audio processing does not increase, but the processing capacity greatly enhanced, can be applied to surround sound, acoustic space, high-end function processing, as to cope with the general DSP functions such as frequency divider, all kinds of equalizer, amplitude compression expansion, The At the same time as it is parasitic on the ARM platform, so that power consumption, BOM price and high computing power possible.