This tutorial will detail how to put together a final interview after you have already recorded the initial capture. When you do the original capture, typically you'll make sure the guest understands they can stop and restart whenever they want to. This makes sure you can always get the best, most eloquent version of whatever they're trying to say, and you can streamline the interview while making it sound better. However, to save the guest time, and to encourage a natural flow to the conversation, you generally won't be doing the same in your questioning. So this will show not only how to process the audio of the interview to make it sound good, but also how to re-record asking questions and add them back in post-production.
- Process Interview Audio
- Put together a cohesive interview with re-recorded questions
- Make an interview sound crisp and clear, and like everybody is super smart and well-spoken
First, upload the raw file everywhere (shared drive, OneDrive, backup, etc.)
Next, comb through and delete all the parts where they said they wanted restart, long pauses, long ums, etc. (and upload/backup this version everywhere)
Next, capture a noise print from the file, and process the noise reduction for the entire file.
The last step for the processing the interview audio is ENCN, which goes as follows:
- Equalizer - Parametric EQ - Loudness Maximizer
- Normalize - 98%
- Compress - Dynamics Processing (de-essing or maximizing clarity)
- Normalize - 98%
The process is:
- Go into the effects rack, filter and EQ, select the parametric equalizer. Select "loudness maximizer" and apply.
- Go into the effects menu, amplitude and compression, select "Normalize(process)" and normalize to 98%.
- Go to the effects rack, amplitude and compression, select "dynamics processing". Set it such that segment one (compressor) has a ratio of 2.03:1, and threshold of above 20.27; and segment 2 (expander) has a ratio of 1.13:1 and a threshold of 20.27. Save this setting as a preset if you haven't already (this is the best for maximizing clarity), and apply.
- Go to effects to the effects menu one more time, amplitude and compression, select "normalize (process) and normalize to 98 again. (and upload/backup this version everywhere)
This process should take about an hour for a 20 minute interview. After you've done all that, we can begin making clips of each question & answer, recording questions, and putting everything together to sound like a cohesive interview.
Putting together the interview:
After you have the audio processed, you'll want to re-record those questions you asked in all kinds ineloquent ways. The best way to do this is:
- Listen through the interview, and simply write down each question you ask.
- Re-write those questions so they're worded perfectly
- In a single recording, just record you asking each of those questions, one after another. It might be helpful to include some conversation touches, like adding "yeah that makes sense" right before you ask the question, or changing your intonation so that it sounds like you're responding immediately. Leave about 5 seconds in between each question so the next part is a bit easier.
- Apply the same technique for vocal processing as you did with the interview. Noise reduction, the ENCN.
- Make sure you record these in the same room that you recording the initial interview and with the same mic settings, and also use the same noise print for the noise reduction as you did for the interview. This will make it sound more like a single recording.
Once you've record all your questions, put the interview and the questions into a multitrack session on different tracks. From here, it's pretty simply. Just use the razor tool to replace the questions in the interview recording with the questions you just recorded. Listen through the recording and for each question and answer follow these steps:
- Using the razor tool, cut the original question out of the interview audio.
- Using the razor tool, cut the new question out of the new questions audio
- Put the new question where the old one was (in a different track)
- Place the tracks so that there's an appropriate amount of time in between each question being asked and answered.
- Put a simple fade on the beginning and end of each track wherever you made a cut to make it sound more like a single recording.
As a quick tip, you can quickly change to the razor tool by pressing R, and back to the selector by pressing V. This will be useful for when you cutting stuff and moving thing around one after another.
Once you've done all that, you should be able to export and save the multitrack mix-down as a .wav file, and it will sound like one crisp interview, recorded all at once. Upload/backup that .wav everywhere necessary, and it's ready to use in the final mix-down of your full podcast, or for whatever other purposes you need.