These are chat archives for highfidelity/hifi

8th
Apr 2017
Tiago Andrade
@TagoWill
Apr 08 2017 19:11
@sethalves maybe. I will take a look.
@themelissabrown thank you very much for the help. If you find anything just let me know.
Philip Rosedale
@PhilipRosedale
Apr 08 2017 20:38
@amvmoody the largest possible value for an audio sample is 32,767, which could be positive or negative. I suggested using 15000 as the max scale for the meter because the chance of everyone driving their audio to full scale is minuscule. You will want to experiment with what a good number is for the top.
Midnight
@MidnightRift
Apr 08 2017 20:39
@PhilipRosedale I thought that might be the case with the 16000. have you checked out the comments on that worklist task. I dont think the meter is going to work as expected atm. more details are in the comments.
Philip Rosedale
@PhilipRosedale
Apr 08 2017 20:40
From using the first version of your work (which seems great!) it looks like you will also want to average over a longer period so that the number is more stable and readable.
Midnight
@MidnightRift
Apr 08 2017 20:41
after 1 second any sound in the same frequency range will go down to zero in audioLoudness
Philip Rosedale
@PhilipRosedale
Apr 08 2017 23:20
@amvmoody looking at the worklist job...
@amvmoody Hmmm.... do you have "Audio Noise Reduction" turned on in the developer->Audio menu (I think that is where it is).
I'm trying your latest right now.
Midnight
@MidnightRift
Apr 08 2017 23:29
that setting seams to help a little. Instead of going down to zero it goes down to about 10-20% of what it was after the second is up.
Philip Rosedale
@PhilipRosedale
Apr 08 2017 23:52
Let me look at the C++ code that generates it... I think you may be right, I might be doing something that isn't exactly what we want... hang on.