I am primarily a musician who has always been interested in the audio production side of things. My band recently recorded base tracks for what is to be our second album. A typical track on this project has drums, bass, piano, an acoustic guitar track, an electric rhythm track, a lead track (mostly guitar, keys on some songs/sections) and vocals (which haven't been recorded yet).
This covers the basis of what I "know" about making recordings sound good:
We recorded everything to levels that filled out the wave field (or whatever you would call it) without peaking, and I adjusted each track to levels as close to peaking without peaking as I could. I compressed each track with a threshold of about -10db to begin with, then upwards of -20db, because of the issues I'm having with the results. Even when every track is compressed to -20db, with the ratio set at 8.00, the master still peaks unless I pull down the master volume, or throw a compression or limiter on it. i feel like it would sound better if the master wasn't peaking at all, with the track set at 0.0db and no effects or EQ. What am I missing here. The individual tracks aren't peaking, but when all played together the master file peaks. Is this OK? Can i just put a Limiter on the master file and call it a day or will it sound too gainy? The only other option i can see is turning every track down by the same amount until there are no more peaks on the master. Is that the way to go? It doesn't seem quite right. Is there some sweet-spot where if each individual track is not peaking past it, the master track won't peak when all of the tracks are played together. I've been looking through this forum for a few weeks now and I really appreciate all the knowledge you guys have about this kind of thing, so any help would be greatly appreciated.
Thanks!
Dave
Comments
This covers the basis of what I "know" about making recordings s
This covers the basis of what I "know" about making recordings sound good:
We recorded everything to levels that filled out the wave field (or whatever you would call it) without peaking, and I adjusted each track to levels as close to peaking without peaking as I could.
Forget what you "know" then. Bad, bad form there.
If you're even approaching clipping on an individual track at the tracking phase, you're FAR too hot and the damage is already done. Add some compression to that and make it twice as bad.
(Dead Link Removed)
*OF COURSE* the 2-bus is going to clip - YOU HAVE NO HEADROOM ANYWHERE!!! You used it all up at the very first possible chance you had!
Where does this "record as hot as you can" mentality come from?!? It completely flies in the face of common sense!!! [/THINKING TO SELF]
Seriously though - Back up - WAY up - The things you "know about making recordings sound good" are KILLING your recordings from the start.
Long story short - Pretend -15dBFS is your clipping point when you're tracking. MAYBE -12dBFS for percussives. You're worrying about the wrong things at the most wrongiest times...
(Dead Link Removed) That was a cool article back there, explain
(Dead Link Removed)
That was a cool article back there, explaining the dBFS and 0dBVU nicely, not to mention the rest.
Here's another question though. I've recently started summing through an analog desk. I basically take stereo feeds from the DAW, with drums, guitars, vocals etc, maybe a couple of mono tracks as well, tweak the mix a little, and record back into the DAW. My question is what level should I be aiming for when recording back? Does the -18dBFS or whatever spec of the converter still apply? Or should you just be aiming to leave enough headroom for the mastering process?
ALWAYS leave headroom (every single step of every single point i
ALWAYS leave headroom (every single step of every single point in the process).
That said - 0dBVU is probably going to be giving you somewhere around -18dB(FS)RMS anyway... You might consider going a whisker lower (my tracking/mixing converter set is calibrated to -20 and I 'pretend' they're at -24).
I liked the article. Certainly describes my learning curve - ma
I liked the article. Certainly describes my learning curve - making the mistake of not leaving enough headroom at some intermediate stage after I had corrected a lot of other beginner mistakes. One point I think you might make is the difference between RMS levels and peak levels. -18dB(FS)RMS will have higher peaks. How much higher depends on the source. I usually aim for peaks at about -12dBFS - alowing for higher peaks for a "spikey" signal like drums, lower peaks for a compressed signal like electric guitar. I never thought of it this way,but I am probably aiming for a consistent RMS reading.
Massive Mastering, post: 297781 wrote: ALWAYS leave headroom (ev
Massive Mastering, post: 297781 wrote: ALWAYS leave headroom (every single step of every single point in the process).
That said - 0dBVU is probably going to be giving you somewhere around -18dB(FS)RMS anyway... You might consider going a whisker lower (my tracking/mixing converter set is calibrated to -20 and I 'pretend' they're at -24).
OK, but do you ever see tracks coming in for mastering that peak at -18dB? If you've recorded following the aforementioned guideline with regards to tracking levels, and mixed down ITB, surely you pre-master will be happily peaking a lot higher than that? Should the signal from the 'summing desk' be coming back in at -18dB, as if I were tracking an instrument... and then what, normalize to something higher?
I'm sure the answer will be 'no... '
I'm kinda new to the analog summing world; been getting great results from the desk though!
1) I never said anything about peaking at -18dBFS... I said -18
1) I never said anything about peaking at -18dBFS... I said -18dB(FS)RMS. Big difference there. A mix that's sitting at -18dBRMS can easily be clipping the digital output.
1a) It's very common (especially among the better sounding mixes) for mixes to come in around -25dBRMS. That generally allows for a nice crest (maybe 18-21dB) and a bit of digital headroom (maybe 3-6dB).
That all said -
2) It's really not that uncommon for mixes to come in peaking at -18dBFS (maybe in the range of -36dBRMS or so, typically).
YES: That's certainly lower than necessary in nearly any given situation.
NO DOUBT: That's in no way "too much" headroom - especially for classical recordings (I regularly work on perfectly fine recordings peaking much lower).
LONG STORY SHORT: If you're TRACKING using converters calibrated to -18dBFS (pretty typical), individual tracks may peak as low as -15ish (dBFS) for more 'dense' sources (synth pads, distorted guitars, etc.), to maybe -12dBFS for 'soft attack' sources such as vocals, acoustic guitars and what not, to as high as -6ish (dBFS) for high-transients like drums. Undoubtedly, you're still going to have to attenuate those sources to have a mix that will even have an 18dB crest. A whole bunch of tracks at those levels are still going to clip the digits.
At the summing desk, it's still the same deal as anything else - Line level is always line level. If your converters are calibrated to -18dBFS (or "somewhere around there") and you hit it with a mix that's hovering around 0dBVU, you're going to be around -18dBRMS. As long as the crest is less than 18dB, you won't have a problem with clipping the input at the converter. Chances are pretty good that it won't be, which will require a little attenuation on the faders.
Personally, I don't find it mysterious at all that where there are "standards" for mix level, those standards are -24dBRMS - regardless of converter calibration. It's a nice, "safe bet" level that has a crest factor higher than nearly any typical 'rock/pop' mix, while allowing a whisker of headroom as a bonus (although I'm pretty sure I repeated myself from an earlier point there...).
P.S. Big fan of analog summing...
I must say, I have found this thread extremely helpful. I've si
I must say, I have found this thread extremely helpful. I've since opted to "cool off" my input levels when recording. Right now the only recording equipment I have is a cassette 4-track running at normal speed, and the advice I got here has resulted in a less-crappy-sounding tape. Thanks!!!
I wouldn't call it "too low" at all. Necessary? Probably not i
I wouldn't call it "too low" at all. Necessary? Probably not in most cases. But certainly not "too" low. I get projects in all the time with levels that don't top -12dBFS (they're usually the better sounding projects for that matter).
-30, -40dBFS, probably too low. Although those come in too on rare occasions. In 24-bit, it's still not a big deal.
Granted - We're (well, we were originally) talking about tracking levels - where -12dBFS could be too hot in many cases.
Keep the peaks! They're cool. [I made this exact mistake at firs
Keep the peaks! They're cool.
[I made this exact mistake at first]
Ignore the peaks, and listen to a version with all the tracks -20 at 8:1 ratio; then listen to the same song but with no compression at all.
What sounds better?
Then worry about pushing the final waveform volume up, which is a whole new ball game.
*unless* by peaks you mean clipping/red lights. In that case, pull all of the tracks down in volume a little.