SineMedia APC - to Break Digital Audio Limits

Release time:2018-07-09 13:21
Author:Wu Jian

SineMedia was founded in Late 90’s when China's audio industry turned from analog to digitalphase. In first ten years of inception, SineMedia helped out musicians and engineers understand and capitalize the new digital audio technology to create more creative music and impacted sound scene. In this supportive process of thousands of musicians and audio engineers, we came to understand music creators’ concepts, and picked up the idea of working for audio media communication from the concept of art back to sound technology. Working together with the technical development team that loves music, we extracted the approximated sound closer to the essence of music out of the noise or distortion and developed a better digital audio platform-SineCore.

SineCore—the Next Generation Realtime Audio Processing Platform

SineCore technology aims at creating up-to-date digital audio that rejects the compromise between perfect digital audio algorithms and latency and reproducing the nature of the sound with three-dimensional standards of sampling precision,sampling frequency, and phase accuracy. We endeavor to lead a more scientific audio era and furnish professional musicians, engineers, digital product manufacturers and end-users with a brand new solution which will inspire the industrial revolution and offer the audience an opportunity to experience the artists' expression in any space and any time.

SineCore audio algorithm processing IP:

1. IIR filter for parameter(include lowpass hi-pass, low-shelf high-shelf, peak,band pass) and graphic EQ;

2. IIR for speaker crossover-includeButter-Worth, Bissell Linkwitz–Riley;

3. FIR filter convolution for physical oracoustic modulation;

4. FIR filter room correction, phasealignment;

5. Single or multi-band compressor;

6. 32 bit Single precision floating pointaudio mixing engine;

7. multi audio metering, including VU, PPM;

8. AI mixing include audio featureextraction and Automix;

9. HRTF surround;

10. Noise cancellation;

11. Reverb include hall, room, plate,echo;

12. Dynamic: compressor, limiter,expander, gate.

SineCore hardware IP:

1. ADDA receiver IP for various commercial markets;

2. AES3-AES / EBU Digital Audio InterfaceIP;

3. AES10-MADI Digital Audio Interface IP;

4. AES67- Network Audio Interface IP;

5. DSD-Multi-Rate Direct Stream Audio IP;

6. BNC Word Clock IP;

7. Open Sound Control Protocol IP;

8. ARM High Speed Transfer IP;

9. STM32 SOC IP.

SineMedia  APC -the World's 1st High-order Real-time FIR Acoustic Phase Calibrator in FPGA

The FIR filter is characterized by alinear phase that adapts to the unique audio requirement - to change the frequency domain without causing phase changes. However, due to the characteristics of the DSP serial, the FIR filter with over ten thousands taps, because of its long delay time, cannot meet the real-time audio performance requirements, and  the extra added delay brings new phase problems. Due to DSP inborn performance issues, the high-order FIR filter, has to stay in the theoretical stage. Even with the complexity of processing algorithm coefficients, some producers even use the CPU of the PC to calculate thecoefficients and process the FIR, which result in greater system latency.

Therefore, for the first time, we haveachieved low-latency performance and ensured that the filter runs without newphase delays by introducing high-performance FPGA parallel processing ofhigh-order FIRs.

Audio Debugging Caused by All Spatialphases Can Be Resolved in 10 Seconds

Traditional solutions to acoustic space problems are quite complicated, and system engineers need to use different measurement and processing tools to solve various problems, such as placing multiple subwoofer in the room, adjusting placement through multiple measurements, and setting different delay time, so that the subwoofer can play in the same phase. What is more, they have to take into account the time relationship between subwoofer and full frequency. 

Following these complex settings, mistakes can be made so easily that the correct results are always hard to obtain. At the same time, the room and speakers need to be considered not only the bass, room standing waves in the bass and midrange areas, the distortion of the crossover phase of the midrange, and the early reflections of high frequencies in the room, leaving system engineers exhausted and spending plenty of time indebugging, afterwards, passing over the riddled sound system to the tuner. The tuner can only struggle with the system problem based on the defective system in tuning repeatedly, and there is no time for music balance. This is why fake chorus is common in the studios of major television stations.

In contrast, the unique algorithm we designed only need nothing but a microphone measurement at one or multiplepoints in a room or in a large space, with each measurement point taking only three seconds, and a great tolerance to noise in the space. And then, thealgorithm will take ten seconds to figure out the phase problem in the system and automatically send the processing parameters of the problem to the FPGA. In this series of processes, no human intervention is required, and just press the measurement key in the microphone, all is done! 

The tuner gets a standard acoustical space in 10 seconds, where there is enough bass due to the cancellation of multiple subwoofers, no low frequency resonance and no annoying feedback howls. Since our smart algorithms are fully aware of the sound field, the sound engineer is able to design the overall style of music frequency response with our frequency response curve design.

All above, it only takes 10 seconds.

Measurements in Real Studio

Comparison of phase diagram before (green) and after (red) processing: from 50Hz, the phase delay is reduced to less than 36 degrees, above 100Hz phase in 20 degrees, so that the speaker in any space becomes a minimum phase system.

Comparison of the frequency response before (green) and after (red) processing in real room: (93 Hz, 270 Hz) caused by the resonant frequency of the cabinet, crossover (400Hz; 2.8 KHz), and there flection in the room (93 Hz, 270 Hz).

Scan the qrcode to reading this article on your phone