OK, how to podcast, if you’re not going to use Anchor or similar offers. Here’s what you need:
1. a microphone or a mobile phone
2. a webspace where you can upload files
3. Audacity for sound editing (that’s what this page is about)
a webspace where you can upload files
Use Audacity and your microphone to record your first episode, or use your mobile to make an “voice memo”. I’m sure your phone has such an app. Share the recording with yourself via mail, save it on your computer and open the recording with Audacity.
No matter how you started, you now have your recording in Audacity.
Select the silence at the beginning and the end using the mouse and cut it.
Double click the track and apply the *Limiter* effect:
Next, apply the *Truncate Silence* effect:
Now comes editing! Listen to your track and cut any garbage you don’t like. Unless you’re a great speaker, chances are that you’ll be cutting a third of your recording! This easily takes an hour or two. Cut out all those uhms, and-uhms, and whatever other filler sounds you make and can’t get rid off.
When you’re done, generate a few seconds of silence at the very beginning.
Import your “jingle”, some noise, music, or whatever (your own, if you can, some music licensed to you if you cannot) as a second track. Proceed as follows:
1. select the entire second track and use the Amplify effect to make it less loud than your podcast; I use -9db
2. select the first second of this track and use the Fade In effect so listeners have a moment to adjust
3. select a few seconds of the second track where it overlaps your podcast and use the Fade Out effect until (fiddle until you’re happy with the transition)
Export as MP3; I use variable bit rate, quality 7, standard speed, force export to mono. On the second form I set the metadata: I have the metadata in a template so all I need to do is load the meta data and change the track title before confirming.
Now I edit my feed file using a text editor. It’s not hard once you can copy from somebody else... Take a look at mine, hh.xml. Save it to disk and replace my stuff with your stuff and delete the extra items and you should be fine.
If you think a little more details anywhere would help, let me know and I’ll add some.
#Podcast
1. Click Removal (Threshhold 200; Max Spike Width 20)
2. Compressor (Threshhold -12dB; Noise Floor -40dB; Ratio 10:1; Attack Time 0.2s; Release Time 1s; Make-up gain)
3. Truncate Silence (Level -25dB; Duration 0.3s; Truncate Detected Silence; Truncate to 0.3s)
I’m sure there’s more I could do. Sometimes the loudness goes up and I haven’t found a good way to normalize it. If you have some good ideas, please let me know. For now, I think having a quick process that always produces acceptable results is better than endless tinkering because that way lies madness and burnout.
If you listen to it, watch the playback level in the top right. I try to keep my stuff around -6db. If the three steps at the beginning didn’t do that for you, you might think about using the Amplify effect? In general, I try to avoid tinkering but there you have it.
Next, cut out all the single clicks using Ctrl-X. Also, mark sections that are too loud and use Amplify with -3 dB to bring them down. Once you have done this, keep using Ctrl-R to repeat this later. This takes a while, I know.
The rest stays: Generate Silence at the beginning, import the jingle, Amplify such that it is at around -9 dB, Fade In the first second, Fade Out a bit before the greeting comes in. Export using variable bitrate, quality 7.
Now use *Truncate Silence* (Level -25dB; Duration 0.3s; Truncate Detected Silence; Truncate to 0.3s).
Play the podcast, cutting all the stuff you need to cut. And if there’s a section that’s too loud, use *Amplify* just on that section.
At the end, run *Compressor* (Threshhold -12dB; Noise Floor -40dB; Ratio 10:1; Attack Time 0.2s; Release Time 1s; Make-up gain).
How To Use A Compressor: Threshold, Ratio, Attack, Release & More
Frotz left me the following comment:
When I do the editing for Internet Office
Hours,
I now do a few different things every time. I bring in every vocal
track for each panelist, and then apply compression to each using
the following settings:
Threshold -17.0, Noise floor -35, ratio 3.0:1, attack .1, release
1.0, and enable "Make-up gain for 0 db after compressing".
Then I set the stereo position for each panelist, and listen,
adjusting the gain manually via the slider for each track if
necessary.
I clip all the vocal tracks to our claps, and then Sync-lock the
tracks and start editing.
Once the edit is done I bring in intro and outro tracks and adjust
everything to fit and then export to WAV. Since the intro and outro
clips aren’t expected to change, I’ve already applied the compressor
settings above to the tracks.
I start a new project, import the WAV, and run compressor with the
following settings:
Threshold -15db, Noise floor -35db, ratio 2.0:1, attack .1, release
1.0, and uncheck "make-up gain" and listen. If it sounds good I
export to variable MP3 (110-150kbps joint stereo), otherwise I might
undo and then redo w/ the "make-up gain" checked and then do the
export if everything sounds acceptable.
In earlier recordings I had to mess around with the noise floor on
my track as my recording environment was noisier.
I’d like to figure out an EQ setting for vocal tracks but since many
of us keep changing our recording setup I don’t feel like I have
something that can be applied indiscriminately.
For recording, on my side I adjust my input level so that my peak
falls between -24db and -6 (in the Recording toolbar, I turn on live
monitoring and adjust the slider until my speaking falls within that
range and try not to move out of position for my desktop mic (which
is harder for me than it probably should be).
– Frotz
I suspect that this leads to more dynamic range (see episode 49): some passages are louder, the quiet parts where I trail off really are more quiet. Now the question is: do I want this in a podcast? Maybe not. Perhaps the strict amplitude increase with a cap, making everything about the same is the better approach for this particular use?
Episode 51:
1. *Compressor* (threshold -17.0; noise floor -35; ratio 3.0:1; attack .1; release 1.0; check make-up gain for 0 db after compressing)
2. *Truncate Silence* (level -25dB; duration 0.3s; truncate Detected Silence; truncate to 0.3s)
3. manually edit it
Episode 52:
1. *Limiter* (type: soft limit; input gain (dB) mono/Left: 6.00; input gain (dB) right channel: 0.00; limit to (dB): -6.00 hold (ms): 1.00 apply make-up gain: No)
2. *Truncate Silence* (level -25dB; duration 0.3s; truncate Detected Silence; truncate to 0.3s)
3. manually edit it
