Sunday, 5 October 2014
13th Week - June Hymn (378th Song)
In this week I've been recording my brother. He told me to record him singing a song of The Decemberists called 'June Hymn'. I took the chance to become his producer.
The first thing to bear in mind was that he wanted to try to sing it very low. He is not very sure of his normal voice and was quite sure than doing it lower it would sound better, more intimate, more whispering, more soft. If we were it to be so intimate I would like to get all the low end in his voice so I recommended him to get close to the mic so we got some proximity effect (when you get very close to cardioid pattern mics, which mine is, they capture more low frequencies, it's a classic technique that film teaser broadcasters use to get very deep and present voice). But first my brother had to record the guitar. We used my main acoustic guitar, it's a hard guitar to play and sounds a little tense but still we can get good sounds from it which it's been very handly in my recordings. He wanted frets to sound as he moved along to them. So I placed one main mic (Rode NT1) in a spot that I thought it was cool behind the hollow of the guitar and another one (Shure SM57) directly to the frets. After recording it I wasn't able to hear his fingers sliding over the strings but it's because he barely move hands, everything's played in the same spot, close to the nut. I got a very warm sound from the first one and a nice brighter sound from the other one.
We recorded his voice. Lately I like to record with the mic as far from the interface as possible and with the back of the mic pointing to it. The back part of a cardioid-pattern mic is the place where the sound gets worst to it, so pointing to the source of louder noise in the room, which is the interface, I avoid it to be recorded, not sure if totally but it's totally hidden by the voice, even more when he's singing from so close and the gain of the input signal is so low. So he did his performance. As my role of producer I had to tell him some tricks to get better his voice. Unfortunately the lowest notes were very difficult to him and they are not very accurate, but the rest is fine. He needed some takes as his voice was shaking in the first ones. But after 4 takes and few overdubs we did it. We decided that he should do his own backing vocals as he knew the song very well, I might have messed it up. As we recorded them I noticed that they were in tone but sonically main voice and backing vocals were clashing, but I thought a little bit of EQ would be fine. Later I'd discover that it wasn't that easy. After all this I just made a viola section with a virtual instrument to give it more expression and some fun to the song. As I suck playing keyboards my viola notes had to be fixed manually: some pitches, some lengths and positions and lots of velocities.
My approach to the mixing was to make it sound like Leonard Cohen or so, where the voice is very characteristic and the rest of the instruments just are in the background. Lonely man telling with sorrow his story. The first thing I worked on was in the voice and tried to give it personality. I got rid of some of the ugly frequencies that the proximity effect brought to the recording which were way too much. I tried to remove some unwanted frequencies first rather than boosting some others, although at the end I added gain slightly on 1k and 5k. I used the default Logic's EQ and compressor. I'm trying to use just one so I master it before I just jump from one to another as I'm used to. And I got a solid and neat voice. I also sent a delay to the voice because I wanted the voice sounding really in the front and reverb use to mess with that, so delay should be the answer. I used a spreader for guitar, so I could put it in the middle without bothering the way of the voice, and EQ (with lots of high end) and a bit compression (I liked the sound of the guitar the dynamics were a little bit too wide which I fixed). Then the backing vocals... This was by far the most difficult part of the production. No matter the volume they always covered the lead vocal. I experimented with many EQs to fix that and was so difficult, I was being too gentle. I started to tweak the knobs more radically and things started to work, the voices didn't sound very well but they fit in the mix if I did the right adjustments. Still don't know if that's what a professional would have done but it worked for me, maybe not the best sound but I could hear the words and I wasn't interfering with the other instruments. Last part of the mixing was automating the volume of the voice as some parts were dug into the track, mainly the low pitched notes and few others, but this part was easy to do.
Masterwise I just used L2 and tried if it would sound better or not. I think it will, now sounded less natural but instruments coexisting better. Finally I showed to my mom and the first thing that she said was: It reminds me Leonard Cohen. Such a compliment.