Tweet
Waveform Statistics
(Forensics Version Only)
There are times when it will become necessary to measure various amplitude related parameters associated with sections (or entire) audio files. The Waveform Statistics system provides that capability. You simply bring up the Waveform Statistics feature and then highlight the area of the file that you want measured. To calculate the data, click the “Build Statistics” button and a display like the one shown below will be created. These data can be copied to your clipboard for use with other software programs such as a spreadsheet or a word processor. The following data are analyzed:
Start Time of the Highlighted area
Stop Time of the Highlighted area
Selected Region (Time Span)
Sample Rate of the File being examined
Bit Depth (resolution)
SHA-1 Hash of File (Unique Cryptographic representation of a file)
RMS (Root Mean Squared)* Value of the highlighted area
Rectified Average Value of the highlighted area
Maximum Positive Peak Value within the highlighted area
Maximum Negative Peak Value within the highlighted area
Crest Factor (ratio of Maximum Peak divided by the RMS Value)
Maximum Low Frequency Value below 10 Hertz
Average Low Frequency Value below 10 Hertz
Number of Clipped Samples within the highlighted area
Number of times the signal passes through zero (zero crossings)
Average Frequency based on zero crossings (this is most meaningful on periodic signals.)
*Note: RMS is the square Root of the arithmetic Mean of the squared values of the selected set of samples.
The following specific waveform statistics can be used to assist in the estimation of the likelihood that two files (or file segments) were made of the same person(s) at the same venue using the same recording equipment. It is sort of a “litmus test” for audio file equivalence. The following waveform data can be useful in that quest:
Start Time (of highlighted area)
Stop Time (of highlighted area)
Selected Region
Selected Samples
Sample Rate of the File being examined
Bit Depth (resolution)
SHA-1 Hash of File
RMS (Root Mean Squared)* Value of the highlighted area
Rectified Average Value of the highlighted area
Maximum Positive Peak Value within the highlighted area
Maximum Negative Peak Value within the highlighted area
Crest Factor (ratio of Maximum Peak divided by the RMS Value)
Maximum Low Frequency Value below 10 Hertz
Average Low Frequency Value below 10 Hertz
Number of Clipped Samples within the highlighted area
Number of times the signal passes through zero (zero crossings)
Average Frequency based on zero crossings (this is most meaningful on periodic signals.)
Of special importance is the SHA-1 data, which is unique for any given audio file. It is a cryptographic hash consisting of a 40 digit hexadecimal number which is essentially a unique fingerprint of the entire audio file. This number is a good way to tell if one file is identical to another file. Thus, it can be useful in Forensics audio work to uniquely identify a file or to detect a file that has been tampered with. That hash number is always calculated over the entire file regardless of the selected area or the channels selected. If even just one audio sample of a file is modified, it will create a different SHA-1 hash number. For example, if you bring up the Demo Wave file called “Male Voice Test Sentence BW” and compare it to “Male Voice ID Test Sentence High Quality” they will sound the same and have the same length and other attributes. However, the first one has one sample edited compared to the other and thus they yield completely different SHA-1 values (proving that they are not identical).
Note: The Waveform Statistics function responds to the channel selected except for the SHA-1 hash (which is always calculated for the entire file, including both channels). When both channels are selected, it responds to the sum of the two (except for SHA-1).
Selected File Portion for Waveform Statistics Analysis
Selected File Portion Waveform Statistics Dialog Box
The Waveform Statistics shown in the “Selected File Portion” example above comes from a “Make Waves Signal” having the following characteristics. You should be able to reproduce these results with your software:
Frequency: 1000 Hz
Amplitude: -10 dB
Length: 30 Seconds
File Type: Stereo
Sample Rate: 44.1 kHz
Wave Shape: Sine Wave
Resolution: 16 bits
Note: Some amplitude related parameters (such as RMS) will be displayed in either Samples or dB or % of full scale depending on your display preferences setting (Edit Menu/Preferences/Display/Y Axis).
(Forensics Version Only)
There are times when it will become necessary to measure various amplitude related parameters associated with sections (or entire) audio files. The Waveform Statistics system provides that capability. You simply bring up the Waveform Statistics feature and then highlight the area of the file that you want measured. To calculate the data, click the “Build Statistics” button and a display like the one shown below will be created. These data can be copied to your clipboard for use with other software programs such as a spreadsheet or a word processor. The following data are analyzed:
Start Time of the Highlighted area
Stop Time of the Highlighted area
Selected Region (Time Span)
Sample Rate of the File being examined
Bit Depth (resolution)
SHA-1 Hash of File (Unique Cryptographic representation of a file)
RMS (Root Mean Squared)* Value of the highlighted area
Rectified Average Value of the highlighted area
Maximum Positive Peak Value within the highlighted area
Maximum Negative Peak Value within the highlighted area
Crest Factor (ratio of Maximum Peak divided by the RMS Value)
Maximum Low Frequency Value below 10 Hertz
Average Low Frequency Value below 10 Hertz
Number of Clipped Samples within the highlighted area
Number of times the signal passes through zero (zero crossings)
Average Frequency based on zero crossings (this is most meaningful on periodic signals.)
*Note: RMS is the square Root of the arithmetic Mean of the squared values of the selected set of samples.
The following specific waveform statistics can be used to assist in the estimation of the likelihood that two files (or file segments) were made of the same person(s) at the same venue using the same recording equipment. It is sort of a “litmus test” for audio file equivalence. The following waveform data can be useful in that quest:
Start Time (of highlighted area)
Stop Time (of highlighted area)
Selected Region
Selected Samples
Sample Rate of the File being examined
Bit Depth (resolution)
SHA-1 Hash of File
RMS (Root Mean Squared)* Value of the highlighted area
Rectified Average Value of the highlighted area
Maximum Positive Peak Value within the highlighted area
Maximum Negative Peak Value within the highlighted area
Crest Factor (ratio of Maximum Peak divided by the RMS Value)
Maximum Low Frequency Value below 10 Hertz
Average Low Frequency Value below 10 Hertz
Number of Clipped Samples within the highlighted area
Number of times the signal passes through zero (zero crossings)
Average Frequency based on zero crossings (this is most meaningful on periodic signals.)
Of special importance is the SHA-1 data, which is unique for any given audio file. It is a cryptographic hash consisting of a 40 digit hexadecimal number which is essentially a unique fingerprint of the entire audio file. This number is a good way to tell if one file is identical to another file. Thus, it can be useful in Forensics audio work to uniquely identify a file or to detect a file that has been tampered with. That hash number is always calculated over the entire file regardless of the selected area or the channels selected. If even just one audio sample of a file is modified, it will create a different SHA-1 hash number. For example, if you bring up the Demo Wave file called “Male Voice Test Sentence BW” and compare it to “Male Voice ID Test Sentence High Quality” they will sound the same and have the same length and other attributes. However, the first one has one sample edited compared to the other and thus they yield completely different SHA-1 values (proving that they are not identical).
Note: The Waveform Statistics function responds to the channel selected except for the SHA-1 hash (which is always calculated for the entire file, including both channels). When both channels are selected, it responds to the sum of the two (except for SHA-1).
Selected File Portion for Waveform Statistics Analysis
Selected File Portion Waveform Statistics Dialog Box
The Waveform Statistics shown in the “Selected File Portion” example above comes from a “Make Waves Signal” having the following characteristics. You should be able to reproduce these results with your software:
Frequency: 1000 Hz
Amplitude: -10 dB
Length: 30 Seconds
File Type: Stereo
Sample Rate: 44.1 kHz
Wave Shape: Sine Wave
Resolution: 16 bits
Note: Some amplitude related parameters (such as RMS) will be displayed in either Samples or dB or % of full scale depending on your display preferences setting (Edit Menu/Preferences/Display/Y Axis).