Graphite Box,Graphite Gear Box,Molded Graphite Box,Parts For Graphite Gearbox Henan Carbons New Material Technology Co., Ltd. , https://www.hncarbons.com
The sound of the voice changer can also be identified
The "speaker" changes the physical properties of the speech and does not change the acoustic characteristics of all the senses. At present, the main advantage of artificially passing the phonetic identification method is that it can distinguish the "advanced voiceprint features" that are difficult to recognize by computers, such as dialect accents, idioms, proverbs, verbal defects, and prosodic features. How is this "advanced"? When we distinguish one's voice from acquaintances, we first pass these characteristics; when imitating the show imitation, we also pay great attention to such features. But computers are difficult to recognize, called "advanced features." Yes, the human ear is the most sophisticated voiceprint identification instrument. Non-professionals are not familiar with the "low-level voiceprint features", such as: formant, fundamental frequency, but the computer is most "familiar", and even the sound intensity, sound length, VOT, etc. are computer-acceptable.
Whether it is manual identification or automatic identification, it is compared by recording the case - "checking materials" and recording the suspects - "samples". Since the voice changer is a "sample" that has been changed by a certain setting, the same setting can be used to change the "sample". As mentioned earlier, the voice changer is a generalized channel. There used to be a case in which the appraiser was a "voice changer" and thought that the speech had been distorted. It was necessary to restore the previous speech by signal before it was authenticated. This kind of thinking, let alone channel reduction can not be restored in the current technology, mainly did not consider, we identify the recorded recording, which is not the result of signal processing through the channel? The sampling rate is higher and the precision is finer. It is also the result of discrete changes. Can it be truly equivalent to a continuous signal such as human voice? The same goes through the channel, and is subject to signal processing, but the difference in size is different in the sense of hearing.