I do mostly acoustic music in my studio. I record a lot of serious players, and some newer artists. I get this because my studio is not open to the public.

More and more, my clients want me to master their stuff, and the more I learn, the more I know I'm not qualified. However, to the extent that I do master, I'd like to get better at it. I get to listen to much of my stuff on the radio, so I can tell what works there or not, although each station is different, it seems, in how they eq and limit for broadcast.

I'd like anyone willing to tell me what I'm doing right or wrong in my basic process.

1) In recording, I often compress * very gently * with my pendulum compressor, to disk. This is mostly to get a little better sn, but also to get some of the beautiful sound of the Pendulum onto disk before (forever) exiting the analog world. This also allows me to take a bit more risk in the gain structure. I try not to compress percussive sounds at all. It seems the best determination here is the envelope of the incoming sound... but there's also spectrum to consider - the attack of some instruments is defining, and loud. Thus, compressing kills the attack transient and character. It turns out that, in the end, those attack transients stand out, and so become the first sound to be eaten by a limiter. Thus, percussion seems to get bigger in the mix after mastering, but not more percussive. For this reason, I leave fast attack instruments very slightly lower in the mix, knowing that limiting with bring them up. This allows me to preserve a little more dynamic range. A slower attack time (tuned to the length of the dominant transient) also seems to be a key "knob".

2) In mixing, I'm ultra sensitive to the failures of cheap speakers - esp subs, that will be the ultimate listening environment. The drivers in these "ring" near their resonant frequencies and harmonics thereof. Also, 8 foot ceilings and 2 foot speaker stands prevail - so nodes in the 90-150 range correspond to these distances. For this reason, I create notches (-2db) often in material that seems prone to bad resonances in this area. Its just gut feel, really, I guess. Also, I don't use low rolloffs if I can help it during mix, notches work better to preserve the low lows. The attack transient on an a440 plucked string goes down into the 30's. Why take it out without a reason. I often find that in the end, a string attack has as much power as a kick drum to move the music along.

On high-end , I've found that much apparant high-endis noise. I LOVE the dehisser in Samplitude, and noise modelers in software. Sometimes these can perfectly tame a fizzy top.

3) Then in mastering, I try several attack release settings, usually fast attack (8-20) and slower release (12-50). I will listen for frequencies that jump out as I increase the ratio. Often, I will go back to the mix, and fix some source of these so that I can get a little more loudness but still a clean sound after limiting. Then I will do rolloff at a good spot (usually 12db/oct at around 50, depending on how clean the program material is down low. Then I go back and re normalize (because the rolloff gave me back some headroom). I don't use MBC's much because they sound a bit odd, but for certain problems - esp if I can't go remix - I suppose they would be handy.

4) Then I listen on several sets of speakers esp auto, and have the client take it home and make them listen to it in every venue they can. Good ones come back with good feedback. I respond with a little eq, as required, remaster, and then repeat until done.

Should I team up with a mastering house so I can be interactive with the engineers, and perhaps remix for them from time to time if a mix has issues they would prefer to see fixed in the mix? Throwing my mix across the wall to a mastering house seems so ... not musical. Can anyone recco someone good and not too expensive for my clients?

Is my head about in the right place maybe? Am I forgetting any basic steps. I figure the radio stations do a lot of limiting so I don't want to over-do it just to be "radio ready". I prefer not to have clients in the room at all for mastering unless they are thoughtful, in which case they can be very helpful in explaining how their genre "sounds".