KR20240012684A - Kimjun hdr(high dynamic range)sound algorithm for wireless headphones & earphones) - Google Patents

Abstract

KIMJUN HDR algorithm is installed in wireless-based devices for wireless earphones and wireless headphones.
Most wireless devices adopt the Bluetooth transmission method, and in the case of wireless earphones and wireless headphones, like smartphones or tablets, in order to transmit audio signals wirelessly, the encoded signal is sent in the form of a packet through the decoder of the wireless earphones or wireless headphones. It goes through the process of being recompressed (encoding) and then decompressed (decoding).
The reality is that the sound quality deterioration that occurs at this time does not meet the standards of the original sound, even though related codec technology has developed greatly.
This means that in terms of hardware and codec specifications set as international standards, there is only an international standard for the decoder, that is, a receiver, and there is no international standard for encoding, but only recommendations, so the sound quality changes depending on the encoding method of each company. .
Therefore, audio signals transmitted through smartphones and tablets equipped with encoders that convert audio signals have quality differences for each company, and this is due to the data input to wireless earphones and wireless headphones equipped with decoders that receive the signals. Because it only receives transmission, there is great difficulty in realizing the same quality as actual Hi-Fi wired.
Accordingly, the KIMJUN HDR Sound Algorithm is installed in the chipset after decoding and is different from dithering, which minimizes digital noise for damaged sound sources with only software. It actively masks digital noise according to the frequency characteristics and waveform of the sound source being played and repairs the damaged frequency region. Correct it first.
*Dithering: A function that masks the digital noise generated during the process of converting a signal to digital so that it is not heard. A function that minimizes digital noise by applying White Noise of 1 Quanta (It is an old technology, so it is not natural, and many frequency bands are damaged. Sound quality deteriorates.)

Description

HDR(HIGH DYNAMIC RANGE) algorithm for KIMJUN wireless headphones & earphones {KIMJUN HDR(HIGH DYNAMIC RANGE)SOUND ALGORITHM FOR WIRELESS HEADPHONES & EARPHONES}

The present invention is different from dithering, which is mounted on a chipset and minimizes digital noise for damaged sound sources using only software. It actively masks digital noise according to the frequency characteristics and waveform of the sound source being played and preferentially corrects the damaged frequency region. It's about method.

Bluetooth's wireless system uses the ISM (Industrial Scientific and Medical) frequency band of 2400~2483.5GHz. Among these, to prevent interference from other systems using frequencies above and below, a total of 79 channels are used, from 2402 to 2480 GHz, excluding the range of 2 GHz after 2400 GHz and up to 3.5 GHz before 2483.5 GHz. ISM is a frequency band allocated for industrial, scientific, and medical purposes. It is often used in personal wireless devices that emit low-power radio waves because there is no need to obtain permission to use radio waves. Amateur radio, wireless LAN, and Bluetooth use this ISM band.

Because it uses the same frequency band as several systems, there is a risk of radio interference between systems. To prevent this, Bluetooth uses a frequency hopping method.

Frequency hopping is a method of moving quickly through a large number of channels according to a specific pattern and transmitting packets (data) little by little. Bluetooth hops 1,600 times per second across 79 assigned channels. This hopping pattern must be synchronized between Bluetooth devices to operate wirelessly. Communication takes place.

Bluetooth connects devices in a master and slave configuration, and if the slave device is not synchronized to the frequency hopping generated by the master device, communication between the two devices does not occur. This allows for a stable connection while avoiding radio interference from other systems. This is a situation that is very unfavorable to the standards for appreciating music or sound sources. In the process of encoding and transmitting data, deterioration occurs and it is decoded and transmitted.

Research and products in the field of 3D stereoscopic sound from various audio companies around the world have been developed over a long period of time and are being applied to many areas.

However, in terms of implementation method, each company implements stereoscopic sound with its own algorithm. As the left and right images, which are the most basic elements of a stereo image, become wider, the sound image in the center becomes increasingly blurred, and the position and tone of each clear sound source changes.

Therefore, KIMJUN is clearly differentiated from other companies' algorithms in terms of clear sound image localization, wide sense of space, and constant tone when converting to 3D after inputting a stereo signal.

The technical problem to be solved by the present invention is to provide a technology for converting a 2CH sound source into a 3D 2CH sound source.

The technical problem that the present invention seeks to solve is to provide a technology that clearly implements sound image localization, which is an existing problem when converting to a 3D stereoscopic sound source.

The technical problem that the present invention seeks to solve is to provide a technology that prevents changes in tone, which is an existing problem, when converting to a 3D stereoscopic sound source.

The technical problem that the present invention aims to solve is to provide a technology that accurately implements the pure, original sound transmission of Hi-Fi.

In order to solve the above problem, in one embodiment of the present invention, the input two-channel signal is upmixed into a 5.1-channel multi-channel, and the upmixed multi-channel is placed at the designated coordinate value of the HRTF. Afterwards, two signals that passed the HRTF are output.

Adjust the mid to adjust the distance of the phantom image created between L and R.

Signals beyond the L and R stereo images are controlled through side signals.

Eliminates interference during speaker output.

According to the present invention, after the decoding of A2DP, which receives only unidirectional transmission for listening to the lower sound source, there is an advantage that the software algorithm is mounted on the MCU before being amplified through the Audio Amplifier and then transmitted to Digital => Analog => Audio Amplifier => Speaker.

According to the present invention, when a 3D stereoscopic sound effect is implemented binaurally through earphones or headphones, it has the advantage of excellent resolution and separation.

This has the advantage of forming 3D sound using only stereo input rather than multi-channel input through HRTF technology, which is different from existing virtual surround.

After upmixing the Left and Right channels of stereo audio signals such as streaming or MP3 and AAC into the Left, Center, Right, Left Rear, and Right Rear channels using the formula below, each channel is down mixed through the coordinate values of the HRTF to create a stereo signal. Audio Amplifier transmission has the advantage.

Crosstalk Cancellation is a technology that allows signals from the left to be transmitted only to the left ear and removes signals transmitted to the right. Conversely, it removes signals transmitted to the left so that signals from the right are transmitted only to the right ear.

The path for the signal from the left channel to the right ear and the signal from the right channel to the left ear are erased so that the left channel signal goes only to the left ear and the right channel signal goes only to the right ear. To achieve this, it is necessary to secure data on the transmission path. need.

The impulse response (IRs, impulse response) recorded in the anechoic chamber is considered as a basic h(x) transfer function and played back in the left and right channels, respectively.

At this time, the position of the listener must be fixed using Crosstalk Cancellation. Crosstalk cancellation causes the sense of localization to fluctuate even if it deviates slightly from the designated position, so the positions of the speaker and listener must be fixed from the recording stage.

Using a dummy head 2-channel microphone such as NEUMANN KU100 at a fixed listener's position, impulse responses reproduced in the left and right channels for the left and right ears are recorded and secured as data.

(Impulse response_LL, Impulse response_LR, Impulse response_RR, Impulse response_RL 4 files are created as data for both ears for 2 playbacks.)

Crosstalk cancellation sound source is created through post-processing. The formula is as follows:

KIMJUN1_L = (I_L*J_LL+I_R*J_RL)+M

KIMJUN2_R= (I_R*J_RR+I_L*J_LR)+M

KIMJUN3_L = (I_L*J_LL+I_R*J_RL)+S

KIMJUN4_R= (I_R*J_RR+I_L*J_LR)+S

Here, the final generated left and right channel sound sources have Crosstalk Cancellation applied. (i.e. output value)

I is the left and right channel sound source of the sound source to which crosstalk is to be applied. (i.e. Input value)

J is transfer function data obtained through recording in advance as described above. (Total 4)

M uses a formula to find the mid image. M=(L+R)/2

S uses a formula to find the side image. S=(L-R)/2

* here is convolution, not multiplication. (convolution)

The caution here is that in order for the denominator not to be 0 when I_L and I_R are derived through the two equations, the value of J_LL*J_RR - J_LR*J_RL must not be 0. When the value becomes 0, it means that the signals coming into both ears are completely the same when playing the left channel. (Conversely, when playing the right channel, the signal coming into both ears is the same).

Afterwards, the obtained OUTPUT = KIMJUN1-4 is transmitted to KIMJUN1, KIMJUN3=L, KIMJUN2, KIMJUN4=R.

At this time, the ratio of KIMJUN1 and KIMJUN3 according to the distance from the listener and room conditions applies the 3:1 rule.

Claims (1)

Dynamic range, which is damaged when transmitting a sound source wirelessly, is implemented as a Bluetooth speaker by passing it through the KIMJUN HDR algorithm.
At this time, when implemented as a speaker, crosstalk cancellation is applied, and in the case of smartphones, TVs, sound bars, etc., automatic conversion through the earphone jack and automatic conversion through the Bluetooth pairing signal are possible.
After decoding during Bluetooth playback, it passes through the KIMJUN HDR algorithm and is implemented in wireless headphones and wireless earphones.
In the case of smartphones, TVs, sound bars, etc., automatic conversion is possible through the earphone jack and automatic conversion through the Bluetooth pairing signal.