Take the 2-minute tour ×
Sound Design Stack Exchange is a question and answer site for sound engineers, producers, editors, and enthusiasts. It's 100% free, no registration required.

I'm creating a podcast with two voice tracks and some intro/outro music, and I'm trying to decide on a general workflow for editing and mastering. Specifically, I'm not sure if some of the effects should be done before or after combining the voices to a single track, and if the music should be combined with the vocal tracks before mastering.

Editing

  1. Create a new project in Audacity
  2. Import vocal tracks from audio interface
  3. Remove any mistakes, long pauses, "ums", etc from each individual track
  4. Remove background noise from individual tracks
  5. Run high pass filter effect to cut everything above 11kHz
  6. Run low pass filter to cut everything below 80Hz
  7. Add compression to each track
  8. EQ each track
  9. Manually adjust the volumes of each track so they sound similar
  10. Pan the host's voice track a little bit to the right
  11. Pan the guest's track a little bit to the left
  12. Combine tracks and export to a lossless format

Mastering

  1. Start a new project in Audacity
  2. Import the combined vocal tracks
  3. Add a little bit of reverb to it
  4. Add the intro/outro music
  5. Run a normalize effect
  6. Check how it sounds on earbuds, make minor tweaks
  7. Export to MP3

Is anything out of order, missing or unnecessary? Do you have any other general suggestions?

share|improve this question
    
That looks like it should work fine for a podcast –  Rory Alsop Oct 7 '11 at 11:48
1  
Why the high/lowpass filters in Editing steps 5 and 6? –  Warrior Bob Oct 11 '11 at 21:23
1  
I don't understand all the reasons behind it, but I've seen it recommended a few times. I think part of it is that the MP3 encoder is going to ditch frequencies that the human ear can't hear anyway, and that the encoder will be able to create a better sounding file if it has less data to analyse from the start. I've heard recommendations to throw out anything below 80hZ or above 11khZ. What do you think? –  Ian Dunn Oct 12 '11 at 14:31
    
the high/low pass filters are useful here - gets you better quality at the same bit rate mp3. –  Rory Alsop Nov 1 '11 at 11:53
add comment

migrated from video.stackexchange.com Feb 14 at 15:19

This question came from our site for engineers, producers, editors, and enthusiasts spanning the fields of video, and media creation.

2 Answers

up vote 4 down vote accepted

I currently do a few podcast and have some input on your workflow.

1) You have your high-pass and low-pass terminology backwards. A high-pass filter cuts low frequencies, a low-pass filter cuts high frequencies. Not a big deal, just remember that "low-pass" is only letting the lows pass, and blocking (filtering) everything else.

11 kHz is far, far too low. Your podcast will sound like an old timey telephone conversation. Unless there's a really good reason, you should be leaving the high end alone. Don't cut highs for the sake of cutting them, EQ them down a bit if the track is unnecessarily harsh.

I do usually cut below 60 Hz to cut out mic boom and rumble.

2) I don't pan on a podcast, nor would I recommend doing so. Work with a mono track.

3) I edit out "umms" and pauses after the track is mixed down.

4) Don't add reverb - there's no reason to. The reason one would add reverb to any audio signal is to simulate being in an acoustic environment, which is sometimes desirable for music. Listen to any radio or television broadcast - no reverb. It will make it sound artificial.

share|improve this answer
    
Thanks for the advice :) Can you explain why you wouldn't recommend panning? My thought was that a subtle difference between the speakers would sound a bit more realistic –  Ian Dunn Nov 19 '11 at 22:45
    
It may sound more realistic in some listening situations, but I would counter that simulating a real conversation is not the intent of a podcast, the intent is to deliver content and information. Even small differences in panning, especially in back-and-forth conversations, are incredibly noticeable to listeners wearing headphones. I only have anecdotal evidence to offer, but the few times I've seen feedback on panned talk segments, they've been overwhelmingly negative. Listeners can distinguish between talkers based on their voice; panning serves as more of a distraction than a cue. –  Michael Nov 21 '11 at 14:13
    
Hmm, ok, I've done it before with instruments (e.g., rhythm vs lead guitar) and just assumed the same would hold true for vocals, but you're probably right. –  Ian Dunn Nov 22 '11 at 2:27
    
Panning is indeed very important in stereo music recordings, and again that's to help simulate listening to music in real life. The rules for what works best in a podcast/spoken word situation are usually different than music recording. –  Michael Nov 22 '11 at 15:42
    
I think that panning is more necessary with music because the instruments are generally played at the same time and panning helps to distinguish between the different instruments and prevent them from taking up the same space in the audio spectrum. The voices in a podcast are (usually) not talking at the same time, so panning them would be more of a distraction rather than an enhancement. –  Friend Of George Apr 5 '12 at 18:12
add comment

Just one suggestion, try to get clean audio as much as possible and then Normalize/Amplify your sound before any of the hi-pass low-pass and filtering adjustments. This will eliminate the tiny, computerised sound that over process and re-processing cause. In other words, do steps 4-8 before 9.

Try using a stage mic (dynamic) over a studio mic (condenser) if your recording room is not studio quality.

Also, have a look at Levelator http://www.conversationsnetwork.org/levelator, I have found that in most cases, it eliminates steps 4-9 altogether.

share|improve this answer
    
Great tips, thanks :) –  Ian Dunn Mar 15 '12 at 0:57
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.