catsittingstill

A Whole Lot of second-stage mixing (for me) is making things louder, often by making parts of them softer.

This probably sounds counterintuitive. But in, for example, a track of me strumming away on the octave mando, some of the strums will be louder than others. This is partly me getting excited, or me striving for expression, and partly the sad fact that I am just not very good, and sometimes I hit the strings harder than I meant to.

And the peak for the track is the loudest point of the track, whether or not that is a strum that is 4 dB louder than necessary. If I go in and clip out that loud strum and reduce its volume to make it only a little louder than the other strums, that will often let me raise the volume of the entire strumming track.

By the same token, if I reduce the volume of the loudest sung syllables, I can raise the volume on the entire vocal track. Thus I can make the voice and the instrument both louder, without blowing out the listeners eardrums--because I can make it more consistently loud.

And while people like songs to have softer parts and louder parts, in general if you want to, for example, listen to music in the car you need the softer parts to be only a *little* softer, or the car noise covers them up and you can't hear them at all. So more consistently loud is good.

The interesting thing is just how much louder my sung high notes are than my sung low notes. I have low notes I'm boosting by 4 dB to get them audible over the instrument, and high notes I'm cutting by 2.5 dB. I end up changing the levels dozens of times in a single verse. Fortunately once I have them entered, the computer handles it automatically for every replay and rendering, or I couldn't possibly do this.

So I skipped one song today because my Ear said it was fine, and worked on three more, and am still on track to finish Friday. But I really need to get a chance to listen in the car again.

I finished the preliminary mix of Yellow Truck Song today. Which means 1) barring something really unforeseen, all 20 songs will be on the album! 2)I finished a day early! Giving me an extra day for mastering, though we are going to Skype-game with Jake tomorrow, so I may not be accomplishing a whole lot with it.

I also fixed a couple of egregious errors in some of my mixes: Ludington's Ride, I discovered when I went to remove the pre-instrumental that just wasn't working, had gone up without EQ or reverb, so I fixed that, and Hydrogen Burns had gone up as just the lone harmony part--no melody, no instrument, so I fixed that, and I had a couple of things that were peaking much higher than expected in the finished version so I tracked that down and fixed it.

Next comes listening to everything in the car and also incorporating Peter's suggestions about fixes. I need to get the sound reasonably consistent, which may involve moving parts in the auditory field (changing the pan) and also making soft parts of the vocals louder so they aren't drowned out by other vocals or by the instrument. The volume adjustments for a single song can take hours--and go up as the square of the number of parts. I'm going to have to spend some time messing with the esses; the standard de-esser is just not working as well as I'd like, and the esses only seem to really become unreasonable after the reverb is on, so I wonder if the reverb is set wrong somehow. If I can find a fix for that it would be easy to make the adjustment for everything.

I got up at 3 am and recorded this morning. I forgot to set the alarm but fortunately I woke up at 3:08, so I hopped out of bed and got to it.

I recorded harmony vocals for Scale (substituted yesterday evening when I discovered that I couldn't reliably sing one transition in the Oblivion Bridge harmony), Hydrogen Burns, and fixed a wrong note in the Wyverary harmony. I tried to record harmony vocals for Neil Armstrong but that part goes too low and my voice was too tired and I could hear that it just wasn't working.

I also recorded octave mando bass lines for Come Flying With Me, Driver's Lament, and All Things Pass.

I tried to go back to bed afterwards but didn't sleep much. About 8:30 I gave up, got up and sat in a hot bath to force my temperature back up (I got really cold while recording.) I don't normally accomplish much on recording days, but I wrote an octave mando bass line for Alice Day that I quite like (and that is reasonably easy to play,) checked out Villain's Cat and discovered no easy way to fit another harmony on that, and rewrote the harmony for Song Of The Makers as a high harmony rather than a low harmony because I think that will be more interesting and also easier to learn. Then I patched in the bit that needed to be fixed in Wyverary, and mixed the octave mando bass line for All Things Pass. Having four different lines in that song is making it quite crowded and I'm not entirely sure that is going to work. I also practiced the transition in the Oblivion Bridge harmony that was giving me trouble and I think in another day or so I will have it down.

Left to record: Song Of The Makers vocal harmony, Alice Day octave mando bass line, Dr Faber's Medicine Show mimmoth harmony, Neil Armstrong chorus harmony, Oblivion Bridge harmony. I hope I can fit it all in one morning but that's a lot of singing, so maybe not.

And another twelve days or so of mixing. Eep. But it is almost 8 pm and then I can go to bed.

I had a very productive day yesterday; I finished cleaning up the mando line for Today Is The Day and then patched and checked the pitch on the whole of the vocals for Fairy Willow (which were not up to my usual standards for some reason but I think I've got a good version of it now.)

But I'm behind on some other things.

So, I'm guessing mixing (recording-type mixing) is probably different things for different people. Probably lots of people are much better musicians than I am and they play their rhythm guitar part and maybe they have to try a couple of times, but they get one with no mistakes in it. No wrong chords, but also no buzzes and no places where one of the strums in their strum pattern didn't happen, or whatever.

That's not how it works for me. It's generally not worth trying more than three times, though if I really crash and burn I delete that take and do another. (And if I'm *really* crashing and burning I give up and put that song off to the next recording session after more practice.) I do two (if things are going well) or three (if I feel insecure) vocal takes also. Or sometimes more than that.

( Read more... )
So perhaps it is no wonder that mixing a short easy song can take four hours and mixing a longer more challenging one twelve. But this is why I am concerned that "October" may not be time enough. Especially since I also want to write and record harmonies, which takes time, and will produce more tracks that have to be mixed.

Whuf. Probably if I were a better musician, this would go quicker.

I went down to Atlanta for the house filk at Dawn's place. I had a nice visit with Alice and Beth and Marie, and the filk itself was a lot of fun. Matt Leger skyped in from Africa, where he and his Songbird are stationed. That was the longest distance skype attendence I have seen, but it worked pretty well; the connection didn't fail once and there were only a few sound glitches on our end. It occurs to me that for filk circles it would be handy to have a camera under the control of the person on the other end; we kept having to turn the laptop so Matt could see the person performing. But it was good to see him and to hear a bit about what his life is up to, so far away.

I took Monday off, feeling kind of flattened, and having a new Temeraire book to read, but got back to work yesterday. I mostly finished mixing Villain's Cat, and I checked its sound against that of Today Is The Day (similar arrangement) and they sound consistent. I also tweaked the bass line for All Things Pass (I'm strumming the mando as the instrumental on that, and it occurred to me that I could use the octave mando to play a bass line, though I think it will be a bit higher than most bass lines, but we'll see.) And I wrote a bass line for Driver's Lament.

I have much much *much* mixing to do, and I need to write harmony parts and record them. So that is keeping me kind of busy.

Oh, and I have a new song; I need to solidify the tune a bit and put it up.

So the 3 am recording on the 17th went better than I thought. Two of the songs seem quite useable, the third--I think I can make it work, though I have so far spent eight hours mixing it and I am not done yet, which is a bit worrying.

I did another 3 am stint today. It seemed to go quite well. I was particularly pleased with the instrumental for Dr. Faber's Medicine Show, which is quite fast (the song is in swing time and to be different I wrote a jig to go under it--lots and lots of fast little eighth notes) and which I basically nailed on the first try.

I do most of my songs by laying down a scratch track in which I sing and play at the same time, then singing "for real" to the scratch track and playing "for real" to the scratch track and matching them up. That wasn't working well for this song; I was playing right at the edge of my abilities when I made the scratch track and it just wasn't regular enough for the "for real" tracks. So I spent several days practicing to the metronome, laid down my mando track to the metronome, and then sang to mando-plus-metronome. That worked pretty well.

I also did Come Flying With Me and That Kind Of Mouse. They seemed to go okay also. We will find out the truth of the matter when I mix. I mean to spend today and tomorrow mixing (and practicing Oblivion Bridge, Neil Armstrong, and Song Of The Makers) and record again Monday am. I'm hoping to have all my basic tracks down before I go to the Atlanta filk circle in a week. But we'll see.

I put together a verse of the vocals and a verse of the mando from Today Is The Day, recorded in the blanket fort, and compared them to some tracks recorded at Carson Newman back when I had access to the padded room there.

I really think the blanket fort recordings are better. I hear a distinct chest hump (reinforcement of lower frequencies) in the CN recordings that I don't hear in the blanket fort recordings. That is a load off my mind, because I made the CN recordings sound okay, so the blanket fort recordings will probably be fine. I'm not in the best voice at 3 am, but I will make it work somehow.

I recorded 2 and possibly 3 more songs early this morning. That made for a rather challenging day, as this was also my morning to volunteer at the Rural Clinic from 9 am to noon. I'm still not doing very well at getting back to sleep after recording. I came home and got most of my practicing done and also ordered new octave mando strings.

Along those lines, I listened to recordings of the mandolin and octave mandolin from Saturday morning versus this morning, to see if the timbre of the strings was changing. They say you should put fresh strings on when you record--but I think they were invisioning recording once a week or once a month rather than every day. (It takes me forty minutes to change the strings on one instrument--and it costs 6-8$)

I think I hear a subtle change in the mandolin sound. I don't hear any change in the octave mandolin sound. It may be that my ear is just not all that good. On the other hand it may be that the mandolin strings are lights and I think I heard that lights "go stale" faster than medium or heavy strings.

I'm out of the light strings (I had about six packages but used them all up for concerts and the like, I guess.) I have ordered more and they should be here Wednesday; I have enough octave mando songs that I don't need to record mando between now and Thursday morning. I have also ordered octave mando strings because I only have one more set. They won't be here until Friday at the earliest, but it's not as urgent. Now that I am recording CDs these are business expenses (I sure wouldn't be changing strings so often if I weren't recording.) Must remember to put these into the spreadsheet.

If all goes well, Saturday should see me with the basics (instrumental track and melody vocals) of all 21 songs recorded. Then I need to leap into writing out harmonies for some of the songs and getting those recorded. Also mixing. The detail work alone (picking which version of a track is best, patching any mistakes, reducing the volume of the breaths, checking the pitch of everything) for a single track can take a full day so I don't have many days to waste.

Must go to bed soon. Getting up at 3 am...

It looks like this:



Inside the blanket fort is my microphone. The theory is that the blanket fort reduces the amount of sound echoing back from the walls and lets me get a cleaner recording. This is a very low-tech way to do this but it is basically free (because I have a lot of blankets) so it is worth a try.

To deal with the noise from outside I got up at 3 am to do the actual recording. Almost nobody mows the lawn at 3 am, the birds aren't singing yet, very few people are driving up the hill by the house.

Early thoughts on the results: 1) I was considerably more comfortable recording in my familiar practice room than in the studio. I think I did a much better job playing as a result. I recorded 3 songs in an hour and a half. 2) I went back to bed about 5:45 am but didn't sleep very well and feel pretty dragged out today. My throat also feels scratchy. I hope I'll be able to make this fly. 3) Kip usually uses my laptop to play videogames on the weekends. He turns it over without complaining whenever I ask (and he doesn't play at 3 am, so there's no problem with recording) but I hate to cut into his fun time, so I'm going to wait until Monday to go through the tracks and see what I think of the sound quality. 4) I have verified the next 3 tracks will work to record to. (Sometimes they don't keep a constant enough beat and I can't play along to them very well. A lot of my strummed pieces are having this problem. I have taken to practicing with a metronome and will record new scratch tracks for them soon.)

Given the difference in comfort (and price!) of recording at home versus in the studio, if the blanket fort isn't good enough I'm giving serious thought to getting the rock wool and putting it up on the walls.

 I have been laying down scratch tracks for the new album at the rate of two a day for nearly a week.  I only did one yesterday, but I wrote a new song also, so I'm good with that, I think.