Koenig Bruce E, Lacey Douglas S
BEK TEK LLC, 12115 Sangsters Court, Clifton, VA, 20124-1947.
J Forensic Sci. 2014 Jul;59(4):960-6. doi: 10.1111/1556-4029.12452. Epub 2014 Feb 6.
In this research project, nine small digital audio recorders were tested using five sets of 30-min recordings at all available recording modes, with consistent audio material, identical source and microphone locations, and identical acoustic environments. The averaged direct current (DC) offset values and standard deviations were measured for 30-sec and 1-, 2-, 3-, 6-, 10-, 15-, and 30-min segments. The research found an inverse association between segment lengths and the standard deviation values and that lengths beyond 30 min may not meaningfully reduce the standard deviation values. This research supports previous studies indicating that measured averaged DC offsets should only be used for exclusionary purposes in authenticity analyses and exhibit consistent values when the general acoustic environment and microphone/recorder configurations were held constant. Measured average DC offset values from exemplar recorders may not be directly comparable to those of submitted digital audio recordings without exactly duplicating the acoustic environment and microphone/recorder configurations.
在本研究项目中，使用9台小型数字录音机，在所有可用录音模式下，采用五组30分钟的录音进行测试，录音素材一致，声源和麦克风位置相同，声学环境相同。测量了30秒以及1、2、3、6、10、15和30分钟片段的平均直流（DC）偏移值和标准差。研究发现片段长度与标准差之间呈负相关，且超过30分钟的长度可能不会显著降低标准差。本研究支持先前的研究，表明测量的平均直流偏移仅应在真实性分析中用于排除目的，并且当一般声学环境以及麦克风/录音机配置保持恒定时，其值具有一致性。如果没有完全复制声学环境以及麦克风/录音机配置，来自示例录音机的测量平均直流偏移值可能无法直接与提交的数字音频记录的测量值进行比较。
