Any mastering tips?

Jun 2, 2005
2,865
2
38
Just done with a project.. mixed a band, and i am mastering it as well.. however, i got a few questions, some might be weird, but i got to know:

- what are the general rules?

- how many seconds between songs... how many seconds silence in front of the first song on a CD?

- I want to send out unmastered songs as well.. just in case they want someone else to master it.. however, at what volume level should i send them out? at 0 DB on the master slider, it clips... should i still send it out at that standard volume level? or should i turn the fader down untill it stops clipping and let whoever is mastering sort out how to get more volume? (im thinking the first one)

- What kind of CD to use to send the masters out?

Is it a good idea to export all the songs as single and mastered WAV files out of nuendo, and then to import them into a new nuendo project file, sort out the tracklist/song order and burn it to CD? or is there loss of quality in between that process?

Cheers in advance, im on a schedule, so any help will do! :kickass:
 
I would think sending it out without it clipping... When I bounce the 2-bus when I'm going to master, I rarely have the fader at 0db. If you mix with compression on the 2-bus its probably easier to have the fader closer to 0 at least. I guess everyone's techniques for mastering are different, thats just something I've been doing for a while now. I guess it's all based around how the song(s) are mixed ya know?
 
i think some general things mastering engineers want are split stereo wavs. also i think they want about 6 dB of headroom on the bounce (i'm pulling this number completely out of my ass, but i could swear i've read it somewhere). I think general consensus is to not turn your mix down by lowering the master fader (although i do it out of lazyness).

someone with some concrete knowledge will elucidate the situation.
 
Keith called it on leaving some headroom for mastering...if your mixes are printed too hot, there will be no room left to do EQ'ing/compressing/whatever without clipping

the way i work when i "master" projects is to render each track at whatever bit/sample rate it was recorded at(i usually go for 24/48, sometimes 24/96), then load each track into a separate project file

then arrange the tracks in the order necessary, take care of track fade in/outs, and determine how much silence will be between each song(i would just go be "feel" on this one).

next, listen to the whole thing through, taking notes as you go...write down what sticks out that needs to be corrected, what's lacking, etc.

then start your 2-buss processing - i usually use a variety of EQ's, compressors, limiters, and sometimes stereo imaging and harmonic enhancers/tape emulators. a lot of people seem to have a "go-to" chain for their 2-buss work, but i like to experiment with different plugs, and stick them in different orders in the chain.

edit: also make sure to dither down to 16/44.1 - but not until VERY LAST!
 
same thing, as far as i know

i'm just trying to get across that one should leave the source material at the highest possible bit and sample rates during all processing, then chop it down very last
 
same thing, as far as i know

i'm just trying to get across that one should leave the source material at the highest possible bit and sample rates during all processing, then chop it down very last

it's not the same thing.

When you take a higher bit rate and just save or render it as a lower bit rate without dithering (quantizing and requantizing), you are truncating or rounding the bits (depending on how your DAW does it). This becomes cyclical which can result in digital artifacts, expressed as additional frequency content. Our ears hear this as harmonic distortion.

Dithering "smooths" the transition from one bit depth to a lower bit depth by randomizing the cyclical calculations that result in the digital artifacts. By randomizing, the "average" value of the randomization mirrors the actual value of the original bit rate audio. And since the cyclical nature is interrupted, our ears do not perceive the additional audio as harmonic distortion - just noise, which is more pleasing to the ear than the distortion we would otherwise hear.

I hope I didn't explain that too wierd. It's still something I'm grappling with as I learn more about mastering...Which seems to have become what I'm "known for" around my area, lol...:kickass:
 
When you take a higher bit rate and just save or render it as a lower bit rate without dithering (quantizing and requantizing), you are truncating or rounding the bits (depending on how your DAW does it). This becomes cyclical which can result in digital artifacts, expressed as additional frequency content. Our ears hear this as harmonic distortion.

Dithering "smooths" the transition from one bit depth to a lower bit depth by randomizing the cyclical calculations that result in the digital artifacts. By randomizing, the "average" value of the randomization mirrors the actual value of the original bit rate audio. And since the cyclical nature is interrupted, our ears do not perceive the additional audio as harmonic distortion - just noise, which is more pleasing to the ear than the distortion we would otherwise hear.

thanks for the clarification on that

now my question is: if there's only a single rendering from 24 to 16 bit, is it going to make an audible difference? i'm asking this, of course, because you mentioned the cyclical nature of these processes...but i don't see how cyclical it can be if the process is only being applied once.
 
So what's your average squash in dbs? I find after 4-6db of average reduction (the ideal number in the L3 manual), the song starts to sound harsh and "squashy":p the hats always get too loud, and I start losing the low end in my kick, so I back off, and I'm rarely getting an even level. A good trick for "final volume" I have learned in my "ghetto mastering" endeavors is to just look at the meters of my playback channel for my favorite new (loudest) cd, and match my work to those levels. At this point, I hate my mastering work, though.:lol: I use the L3, by the way. What are some other "go to" plugs with waves?
 
thanks for the clarification on that

now my question is: if there's only a single rendering from 24 to 16 bit, is it going to make an audible difference? i'm asking this, of course, because you mentioned the cyclical nature of these processes...but i don't see how cyclical it can be if the process is only being applied once.

I believe the cyclical nature refers to the cycling of the algorithm for the length of the song and the values vary with frequencies/amplitudes.

Re-rendering would just amplify the distortion becuase you would be rounding the bits on already rounded bits, thereby creating more distortion.

This is how I take it.
 
So what's your average squash in dbs? I find after 4-6db of average reduction (the ideal number in the L3 manual), the song starts to sound harsh and "squashy":p the hats always get too loud, and I start losing the low end in my kick, so I back off, and I'm rarely getting an even level. A good trick for "final volume" I have learned in my "ghetto mastering" endeavors is to just look at the meters of my playback channel for my favorite new (loudest) cd, and match my work to those levels. At this point, I hate my mastering work, though.:lol: I use the L3, by the way. What are some other "go to" plugs with waves?

Since I *think* i've learned more about mastering techniques I rarely, if ever, go above (below? haha) a 2dB reduction. My goal is to only use a limiter for ~1dB reduction.
 
A good trick for "final volume" I have learned in my "ghetto mastering" endeavors is to just look at the meters of my playback channel for my favorite new (loudest) cd, and match my work to those levels.

i would totally NOT do this!

mix/master as much with your ears, and not your eyes. i know we all want our masters to reach the volume of other commercial releases...but the truth is, most of the shit being released now is just way too pancaked.

also, try do everything in your power to avoid smashing flat with the L3...there's a lot of things you can do to raise average volume other than lopping off peaks.
 
Double check the individual tracks during the fadeouts – when the tracks don’t all end at the same time.
I recorded my buds CD over here (acoustic g + vox) and during mixing, the tracks sounded like they faded to infinity, but it was evident during mastering at the pro studio (better monitors) that some individual tracks sounded slightly “cut off” because they weren’t quite down to infinity.
Close is not good enough.

Granted it was barely audible and you had to listen closely to notice it, but it was still there and could have easily been avoided - I just wasn't paying close enough attention to the very ends of the songs.

Was definitely one of those “how could I have missed that” moments.