It is a good read and all the tips are good.
Point 4 There are different schools of thought though on some points like mixing into your mastering chain.
When you do that, in a way you limit what a good mastering engineer could do. From that point on say. You are printing in a whole lot of mastering effects. And you may have altered overall mix transients making the mix less snappy and more mushy.
There is no reason why you cannot get a perfect pristine beautifully balanced mix with all the important areas controlled in a nice way dynamically and all without an ounce of mastering processing. Then a really excellent mastering engineer has the best thing to work with. And they will more than likely take it well beyond any mix buss processing you probably did. That pre mastered mix needs to sit rms wise at a certain ref level such as -20. Then transients should occupy most of the dynamic range with the loudest peaks close to - or -2 dB FS. Transients sound the best at this point.
One could say going this way will still also produce a spectacular sounding finished product as well. Having un-processed mixes also lets you go back and re-master much more easily and the outcome will be better too.
No harm in checking your mixes with mastering chains on though. Just make sure you always print a totally un processed mix as well as a semi mastered one.
(Even if it means altering the mix for both) For the highest quality I work at the K System K-20 level so the rms level of a pre mastered mix is at that level. Peaks vary and maximum peaks should end up close to -3 dB etc for most of the music. If I were printing a semi mastered processing chain into a mix I would also ensure it only ends up still at K-20. Then a mastering engineer can at least take it higher in loudness if need be.
Always compare (A/B switching) a total un mastered mix to any processing mix at the exact same rms level and then you will really hear what the mastering processing is doing. Especially to transients.
Digital is giving us the best possible ability to capture and reproduce transients. And it much better than analog in this regard. It is our duty to preserve them and make sure they are still there right at the end of the mix. Any processing be it track, buss or main mix should not be effecting transients too much. Compare the sound of transients before and after any processing anywhere. Listen for them. They will tell you in many ways what is happening to the sound.
Point 9 With regard to automating before any compression the same thing can be achieved by opening up a full vocal track in an editor such as Adobe Audition and simply run a VU meter over it. I create a version where I tweak phrases etc to improve how the VU moves and makes the vocal track much more even. Use your ears here to catch any un-natural edits. I just save this with ED1 in the title. ED1 tells me one level of level editing has been done. Drag that out to replace the original vocal track and you now have a near perfect vocal track level wise, from start to finish. Any compression from now is going to be light conditioning which always sounds best. No automation also required now. Less CPU resources being used and no strange automation behaviour. Final vocal automation tweaks could be added in here too when the whole mix is in for sure though. But they will usually only need to be very minor and either just lift or turn down a vocal level by a dB or two.
post edited by Jeff Evans - 2016/07/24 16:16:02