(01-Jun-2017, 05:29 AM)rryyyaann Wrote: ... I've read that it's not wind noise its just the mic struggles with that high of volume and clips...
The thing about clipping is that it causes distortion not noise. I can confirm that an unaltered action cam is likely to have clipping when sitting between 4 noisy props but the wind noise is like clipping plus noise.
Because both are likely happening you probably want to troubleshoot each issue independently. Wind noise could be resolved or reduced via the Bruce method that Thibault shared. Clipping would require reducing the input audio before the cause of the clipping (I don't know if the clipping happens in analog at the mic, or in digital during processing/encoding of the file).
To tune your ear to hear each issue it helps to understand both. One way to get a feel for which is which of course is to record some HD video on a windy day. Do a punchout way up into the sky then dive back down camera first toward the ground (recover before crashing of course :p) You'll be able to hear the growl of the props. They have pitch and sound like a distorted high pitch growl. That's your actual audio clipping.
Then as you fall... all the static noise is what's left from wind noise.
Some more info for anyone who is curious about clipping vs noise:
Here is an example of clipping
The gray portion of the wave is deleted because it's amplitude was greater than the maximum recordable
As you can see even on clipped audio you maintain pitch to a certain extent. As multiple pitches modulate eachother you might loose some of the sounds but ultimately you're going to get some sort of tones even out of very clipped audio...
To think of clipping mathematically consider that each digital point on a wave gives instructions to the computer on how much or little pressure should be applied to the driver (speaker). So for example the crest of a triangle wave in an example audio file with amplitude range from -10 to +10... If the crest values were 0,3,6,9,12,15,12,9,6,3,0... once clipped the values would be 0,3,6,9,10,10,10,9,6,3,0 (notice the 10s in the middle). What this does it turns the triangle wave into more of a trapezoid wave lol... This distortion usually sounds like a buzzing sound.
With wind noise because the pitch is random it covers up other noises. When there is wind noise plus clipping then you pretty much cover up the existing audio all together. It would be like typing several layers of random letters over existing letters ultimately you wouldn't be able to tell which letters really say the words on the page. (I could go into noise floor and why you can hear both but the non-noise audio sounds quieter behind the noise but I'll leave that out in this case but it might be worth more research)
Example of noise:
To imagine the noise being clipped just imagine removing the information from the wave from the top 40% of the image and the bottom 40% leaving just the middle 20%... you'd still have random pitch but the peaks above and below -10%/+10% would all be cut flat (or near flat depending on what source caused the clipping. This is an area where digital vs analog plays a big role. And it's why many say you can't exactly recreate the sound of a tube amp)
anyway... I probably overshared but hopefully it helps with troubleshooting audio engineering issues