Doug Young

1) Opened the files as a stereo file in Sound Forge and listened. The track seems pretty well recorded (nicely played, too!), and my first impression is that there’s little to do. The room noise doesn’t seem too evident during the piece, except on the fade out, where I can clearly hear the room noise. Overall, my only complaint is that the tracks seem a bit thin, I’d like to hear a bit more body.

2) Normalized the files to -3db, using Sound Forge’s normalize function. The level of the original recording is a bit low, especially with only 16 bits. The level of the originals is -13DB, which leaves a lot of headroom on the table. If you could boost this level without boosting the noise, the noise would be much less of an issue.

3) Muted the beginning noise and trimmed the silence to a reasonable length.

4) There are a few squeaks that stand out. Not serious, and I’d tend to just leave these alone if it was my recording, but since we’re playing mastering engineer, I thought I’d see what I could do. I selected a squeak, and used the spectrum analyzer to look at the sound. The squeak seemed to be most prevalent at 1.8Khz (and above). I used a de-esser setting on a multi-band compressor to compress at 1.5K and above, selected just the squeak, and compressed. Repeat for 3 or 4 other squeaks. It’s not real clear to me that this is an improvement. The squeaks are softer with the edge taken off, but they sound different now. If they sound too artificial, it’ll call attention to them more than just leaving them alone. Well, we’ll see if anyone notices :-)

5) Added a bit of compression, using SoundForge’s WaveHammer. I used only a touch of compression, so little that the meters never even show any action. But you can hear the effect, the track just gets pushed a bit more out-front.

6) I loaded the file into an EQ comparison tool, HarBal and compared to two other tracks that I like the tone of: Ed Gerhard’s Moon River,and Laurence Juber’s Cat Hill. I keep these two tracks around as references. The tool confirms what my ears are telling me. The track has a lot more high end than my references, especially between 1K and 8K. I cut about 3-4 DB in that range.

7) The noise on the fade out is a bit of a problem. After a bit of experimenting, I hit on a solution that seems to work reasonably well. I loaded the file into Adobe Audition, which has a “Dynamic EQ” plugin. Basically, you can change the EQ as the tune progresses. I selected the tail, and set the EQ to start with no EQ, and to progressively cut the highs as the tail fades out. I both increased the amount cut, and lowered the cut-off frequency as the noted faded out. This seemed to eliminate most of the higher frequency noise, but there was still a rumble. So I did the same thing again, but this time used a high pass setting, and cut more and more bass as the noted faded. Not perfect, but the result is a fade that cuts both highs and lows as the note fades out. Hopefully after a bit of reverb, this will sound fine. At least less distracting than the noise.

8) I started experimenting with Acoustic Mirror for adding reverb, and on a whim, tried out some of the mic models. Even after the EQing, the tracks still seem a tad thin to me. I discovered that I had a ribbon mic impulse that produces a very fat sound. Blending that with the original produced a nice sound with both the highs of the original and a fatter low end. I decided to save two versions of the file, one with the ribbon modeling applied, one without.

9) I used Acoustic Mirror to apply the Lexicon Concert Hall model. The reverb was mixed in at -24DB, where I could only just start to hear it. I applied the same reverb to both of my tracks, with the only change being a slight boost to the high end (reverb only) for the ribbon model.

10) I burned both versions to a CD and listened in the car, just to hear them a different way. Seemed a bit muddy there. I decided to cut about 1.5 DB at 120 Hz to reduce that just a bit, even tho the tracks sound fine on my studio monitors. Finally, I checked my volume levels, and again normalized both tracks to -3DB, save to APMEdit_DougYoung1 (non-ribbon), and APMEdit_DougYoung2 (the ribbon model version)

I just updated my submissions, adding a 3rd, which uses the Audition spectrum view to all but eliminate those squeaks. This works really well! Thanks for the tip, Fran. This was a great find, no more squeaks! I started with the Ribbon version for this, and added a little more high end and cut a few DB at 200, based on comments that it was still a bit muddy.

Link to page of additional info from Doug.

http://www.dougyoungguitar.com/APMEdit.htm