No announcement yet.

Record audio with synced decible log?

  • Time
  • Show
Clear All
new posts

  • Record audio with synced decible log?

    How would you go about setting up a long term cheap system to record audio every day over a month and also measure the audio levels and be cable to sync the level measurement with the recording? I have PC to dedicate and perhaps there is some free software and but I dont know how to measure the audio level in dba then sync it to the recording. Any thoughts?

    Last edited by rkruz-elSAT; 01-06-2018, 05:13 PM.

  • #2
    A Month of recording? Is this some kind of industrial application?

    I'd need allot more info before I could even begin to guess what might work, but based on what you posted I can tell you the problems you will have.

    First is the file format you would use. A One terabyte drive can fit about 65 days of 16/44.1 CD quality wave files. 45 days if its 24/44.1 files. If the audio quality isn't as important a 320kb mp3 will fit 303 days.

    The problem you'll have recording audio is file size. Win 7 and below has a maximum size of the authored file is 4 GB. Win 8 and 10 may be a little higher but its not going to allow a continuous audio file for 30 days. you wouldn't be able to open a file that size.

    Most audio programs record the audio to a temp file before its saved permanently. If the temp file exceeds the space limitations, or has a glitch like when the CPU get busy (very common in audio) or a read/write error can freeze the recording program and cause the entire file to be lost in a crash.

    They do have video programs that can record up to 70 hours but I'm not familiar with their ability to automate the recording of multiple files and they may require a manufacture specific DVR system for long recording.

    An audio program would need to break the audio up into smaller manageable files before 4G file limitations are met.

    Next you ask for the ability to A weight the audio before recording. An A weighted decibel meter uses a specific mic combined with filtering to remove bass so the meter hears sound more like our ears do based on the Fletcher Munson curve. for a computer to record an A weighted signal you first have to choose the reference mic you plan on using then EQ it to be A weighted. Using a software plugin to do that wouldn't be wise because running a plugin when recording is problematic. You'd either need to use a hardware EQ for that and properly calibrate it using a program like RAL or you could EQ the audio after its been recorded and simply render it with the proper EQ settings.

    The ability to log the changes is difficult too. You give no information on the application so I don't have a clue what you are looking for no less log.
    You can open an audio file in audio editor file and use its dB grid to see its loudness levels. The file is seen as a histogram in most audio programs so its already logged. From there you can even magnify its view to see very low audio levels and magnify its time line to view the entire span of time or zero in on an hour, minute second, or microsecond of audio time and see the audio peaks occurring by zooming in on it. many editors have a decibel time vs decibel grid as you view the file.

    Many editors will also scan the file and give you averages for RMS, Peak and Average audio levels of the entire recording. If you have certain thresholds beyond this you need to average, you'd probably need to get someone to write you program to do that.

    There are some audio tools that have thresholds that can be set manually. Some compressors and limiters have the ability to render the file and leave only what is being trimmed away (a differential rendering like a photographic negative) which might be used to detect levels exceeding the threshold you set.

    Again I'm only throwing out a few possibilities based of your vague description. More details may give you better options to use.
    First thing that came to my mind is it sounds like you need a tool that will only record results when a certain loudness threshold is exceeded. Like a voice activated recorder that allows you to set how loud the sound needs to be before the recording kicks on. Then have it set to turn off after the volume drops down below the threshold. This would collect the audio that's too loud and ignores the audio that's too quiet. Of course that could be reversed too. Have it ignore everything over a certain loudness and only record when things get quiet. Doesn't make allot of sense recording silence but with a clock you'd know when the events occur.

    There are allot of ways of skinning a cat, but I'd need to know allot more details if you want to use existing tools and do it cheaply. You may not need to skin the entire can and simply don't know that. As a tech we use certain specialized tools for measuring Long term events like companies AC voltages using a specialized monitor. It only saves the abnormal events however. You don't care how long the AC is running fine, only when its not running fine. If its there for a week and for 6 days and 23.999 hors its running fine, I'm only worried what happens in the .001 hour events. If they are spikes and I have a clock and that clock tells me the spike occurs day same time then you can track it down to some electrical timer kicking on.

    I suspect your need to read audio is similar.


    • #3
      thanks. the application is to record audio levels over long term to support leaf blower ban. So I need to log the audio level with time stamp along with the synchronized audio. Long term contiguous recording for raw capture of low fidelity, single channel then subsequent editing to create a file for presentation with time domain around the loud events showing audio levels relative to references like safe hearing levels, jet engine etc