Peter Figen
Offline
• • • • •
Upload & Sell: On
|
 p.3 #13 · p.3 #13 · Fuji GFX 100 II and new lenses - Pre-order / Shipping Thread | |
BrandonSi wrote:
Yeap.. hitting the right LUFS is something I struggle with. I am not an audio engineer, just learning via tutorials. I record the audio separately, cut into chunks in Audacity, normalize, and then do a bit of EQ in Resolve / Fairlight. Then I have separate track for music side chained to the voice-over track to compress down the music audio when the vocal track is active.
I had uploaded an original version that was coming back as -10 db on stats for nerds last night , so I just bumped up the entire audio bus a db or two and re-rendered / re-uploaded. If you see the other videos the tend to over around -4-6db, which seems to sound about right to me.
I am gun shy with the audio, I made a video once and it was like -20db, sounded fine locally but must've had my speaker volume jacked up and it was just the worst, no one could hear the video at all.
It was my understanding (perhaps incorrectly) that a negative db value on stats for nerds was a good thing, that I was -6db under 14 LUFS, and that you wanted to avoid YT reducing the volume of the audio. The goal obviously would be to hit 0, but if you were under 0, then YT wasn't touching your audio track. Is that wrong? (This is what I was basing this off of - https://productionadvice.co.uk/stats-for-nerds/ )
Audio is definitely not a strength of mine, just trying to learn as I go when it comes to most of the video / audio editing stuff! Thanks for the notes! 
...Show more →
When you see -6 db on stats for nerds, that means that they have knocked 6 db off of what you gave them and that your average was 6 db louder than their normalization target of -14, or at -8 db. And to make matters worse, we're all dealing in negative numbers where everything feels backwards.
The basics are that 0 db is as loud as you can be so in order to leave headroom for louder sounds or passages, the averages have to be less than zero or in the minus db range. YouTube's target standard is -14 for music and that leaves you with 14 db of headroom for a really loud passage - drums or guitars or cannonball blasts or whatever. The target levels for film and tv are -21 or -23 db which leaves a lot of room for sound effects in movies to get very impactful.
The goal generally is to have you Integrated LUFS, which is the average throughout the entire piece to be about -14 with peaks coming up to about minus 1db, just shy of clipping, and anytime your average LUFS is louder than -14 YT is going to turn is down to that level, but if you're already quieter than that, as you discovered, they do not turn it up.
This is also where using compression, even on speaking vocals can give it that broadcast quality by slightly clamping down on loud passages and raising quiet passages - not enough to be obnoxiously apparent but just enough to take any edge off. And then, if you have any spikes, which you probably don't in these types of audio tracks, but if you did that's where you put a limiter at the end of the audio chain to lop off that last db or two to keep them from clipping.
I hope that made some sort of sense. Now I need to borrow that lens from my neighbor.
|
So no worries at all.