The most recent one I created is writing a function for extracting the initials from a full name string:
Here is a second one for anyone interested:
Audio on this particular video is particularly very soft.
on which video do you mean?
The spiral problem video. Great solution btw
ahh ok, yeah I messed on which audio input was being record during the whiteboard session, and I’m still trying to find the best way to EQ the audio. The next one should be closer to perfect
I sometimes run sound/mix for events/musicals/plays/concerts. Some tips…
Roll-off the low end, there is nothing there but rumble, hum and noise. Use a high-pass filter.
This one tip alone will make a big difference in vocals. (will cut out background ambient noise, AC/fan noise, wind noise, rumbling noise, footsteps, etc)
Make sure there is enough gain but not clipping.
Use a condenser microphone instead of a dynamic microphone (note: will need 48V phantom power)
If the “ss” aren’t intelligible, a little boost around 6khz to bring some nice sibilance, but not too much. Just give it some “air”.
Depending on voice, a 1khz slight bump will improve intelligibility. But not too much as ears are most sensitive in this frequency too.
If you’re talking softly, or vary your volume levels too much, compression will help as it makes the overall levels more constant… it boosts the quiet portions, and compresses back down the loud portions of your audio.
Wow, thank you so much for that info. I plan to run future videos through NI solid Eq in Ableton, and I think I found a decent configuration, but I’ll tweak it some based on your suggestions. I do have a basic large diaphragm condenser mic, but I’ve just been running a bass boost in imovie (which sounds bad now that I’ve used real EQ). I assume these settings change based on my natural vocal range right? Sometimes it’s hard for me to tell what sounds best after spending too much time turning knobs and dials. Also I’m trying to figure out what volume my audio should be to match other YouTube videos. I read Europe uses -9db or something
The most important thing is not exceeding 0dBFS, anything that exceeds it is an overload and you’ll hear ticks/clicks in your audio, i.e. distorted sound, lost information. After compression, one can make the audio average around -6dB. Note: There’s a lot of energy in the low end frequencies and it’s easy to peak 0dB and overclip, and yet the overall audio still sound weak. That’s why it’s good idea putting a high-pass and getting rid of these low end frequencies. Then what you see on your meters will be mostly energy from the vocal range frequencies without the spurious low end.
I’m uploading to Dropbox an mp4 video of your spiral problem with the audio adjusted. I’ll message you the link when it’s done. Let me know what you think. You can use this to replace your YT video without losing your view stats. (The quality though of your voice by the whiteboard and in front of your computer are so different and I didn’t spend a lot of time matching it. I just made it more easier to listen to.)
- yeah, don’t use bass boost.
Awesome, thanks! Could I get a screenshot of whatever eq settings you used please?
Did you get a chance to see/hear the updated version I sent via Dropbox ?
Settings I used:
I didn’t do anything wild with EQ’ing, just rolling off the low end, putting a bit more body in your voice (because the mic seemed far away and the volume very low) and a touch of sibilance to improve intelligibility.
Sounds good, heard loud and clear!