KTV Sound Design With Headphone Playback

With the development of Internet technology and digital music technology,music has become one of the main applications of the Internet,and online KTV has also become popular.Sound effects processing plays a vital role on the effect of KTV,but the sound processing function from all kinds of KTV software's will be limited by the hardware and software resources of mobile devices.To make the online KTV occupy less hardware and software resources of computer or cell phone,the microphone with KTV sound effect processing has emerged,however,most of sound effect processing only possess the echo function due to the limit from power dissipation and cost.On the other hand,as the entertainment for individuals,the online KTV is replay by using headphones,so there will be the phenomenon of "Inside the head location".At present,there is no the product with sound effect processing to improve this phenomenon.In view of the problems above,this paper designs a KTV sound algorithm adapted to the earphone replay and implemented it on DSP.The details as following:Firstly,the design of reverberation sound effect.By analyzing and comparing the KTV software with the commonly used reverberation algorithm model,and the early reflection,later reverberation and holistic reverberation algorithm model was designed by comb and all-pass,and the late reverberation model was also designed for the stereo output and the comb delay length.The two channel correlation coefficient was reduced from 1 to 0.006,and the amplitude-frequency characteristic became more flat.The designed reverberation model was compared with the algorithm and software KTV.As a result,the echo density of the design reverberation model and the algorithm is the highest,with the complexity of Freeverb 53.1% up to 179% echo density,and the echo density of the four KTV software is higher than the national K song and the singing bar,while the reverberation time is second only to the good singing.Secondly,improve the reverberation algorithm sound coloration.The accuracy of the three main parameters(the central frequency f0,the gain G and the quality factor Q)of the design method parametric equalizer was analyzed,as a result,the accuracy of the center frequency and the gain were consist.But compared with the design of the analog domain,the Q value of the digital domain design method is more stable under the same condition.A ten segment equalizer is designed by digital indirect method,which improved the amplitude frequency maximum peak point of the reverberation processing,as well as the overall flatness of the amplitude frequency characteristics.Thirdly,the design of the virtual surround sound effect.Decorrelation processing was carried on accompaniment,and correlation coefficient and calculation were compared for the decorrelation method.The new upmix was designed by using the selective delay decorrelation method,and the conversion of the four channel to the five channel was achieved.According to the simplified HRTF model,the downmix design of the five channel to the double channel was also completed.The above design algorithm was implemented on the SHARC ADSP-21489 development board of ADI,and the calculation of the algorithm was 91.03%.Compared with the four commonly used KTV software,the results showed that the design model improved the shortcomings of all the KTV software,the amplitude frequency flatness was more flat than the four KTV software,the reverberation time was secondly only to good sing,and the echo density is higher than the national K song and the singing bar.According to the principle of membership degree of fuzzy theory,we designed the sound effect comparison between the listening experiment and the national K song,i.e.,the proportion of each index in each grade is calculated by the people's number in each grade,and each evaluation index and the overall sound effect were evaluated at five grades of membership degree.The results showed that the membership degree of the National K song is 17%,while the design model has a good membership degree of 92% in the evaluation index of sound image externalization.In the whole sound effect,the design membership is 65.6% and the membership degree of the National K song was 57.8%.