How to EQ Your Dialogue from Start to Finish: Recording Step 2

You've worked hard with your microphone placement and recording levels to capture a great sounding spoken word recording.

However, you find that it just doesn't sound as clear and crisp as you'd like it to, especially compared to other dialogue recordings online...

Let's continue from last week's post, Recording Step 1, where you have your dialogue recording in hand - just the recording with no processing.

I'll show you my exact voice recording processing settings, specifically focused on just the EQ settings, from a voice-over for video project.

But keep in mind, these are not a preset to be used on any dialogue recording.

Instead, think of them as guidelines or principles to navigate your EQ corrections and enhancements.

With these EQ settings, I'm molding or sculpting my raw voice recording audio (being aware of its flaws) to get a bit closer to what my reference voice recording sounds like.

Finding the Flaws

So, how do you become aware of the flaws or good qualities of your podcast, sermon or voice-over audio to guide the EQing process?

I'm so glad you asked as the answer may not be the most obvious.

You start to become aware of the good parts and of the deficiencies of your voice recording this way:

You begin to listen and compare your raw recording to a good sounding reference recording as detailed in my last post, How to EQ Your Voice Recording from Start to Finish: Recording Step 1.

The reference voice recording becomes the guide for your own voice recording and points out the good qualities and the problem areas.

Of course, don't forget to volume match the reference audio to your own dialogue audio to be able to make a helpful comparison.

All of this is something you can start doing today!

Starting From the Top

Continuing, the EQ settings we'll cover are below in the channel strip plug-in.

Look closely at the four sets of knobs in the center of the plug-in from top to bottom - orange, blue, white and black:

First, let's look at the high frequencies (the two orange knobs at the top) that go from 1.5 kHz all the way to 17.0 kHz.

Sonically this covers some important areas for the intelligibility, clarity and sheen of the spoken word recording.

I'm doing a gentle, broad boost at 14,000 Hz (14.0 kHz), to "open up" and clarify a dull sounding top end to my raw voice recording.

The lavalier mic I was using emphasized more of the mid range and low end of the voice, so the top end was lacking.

My exact settings are: a 1.5 dB boost at 14.0 kHz using the shelving function (see image below).

Here is some of the thinking behind these settings:

• 1.5 dB boost - I didn't want to overdo the boost since you can introduce hiss, noise and harshness by boosting too much in the top end:
  
  
• further, any boost on the top was a compromise as it also added to the sibilance issues I have in the 4.6 to 4.8 kHz area...
  
  
• so, you could boost less here to avoid having to deal with more sibilance down the line...
  
  
• but, I chose more boosting for the upper end clarity and had to tackle the more severe sibilance as a side effect.

• 14.0 kHz shelf - in this EQ, as in other Neve EQs, the shelves sound very "musical" and "natural" in the high frequencies:
  
  
• the gentle shelf boost at 14.0 kHz will extend a touch below 14.0 kHz as well as boost more dramatically above it, having a broad effect on the upper frequencies that sounds great.

As always, there are many ways to EQ and it would be just as valid to go the route of cutting more than boosting.

For instance: you could leave the high frequency area alone and cut out other trouble areas in the lower frequencies that could then bring out the high end.

Don't be afraid to experiment!

Moving On Down

Continuing, let's move to the high-mid range frequencies (the three blue knobs below the orange knobs) that go from 0.8 kHz to 8.7 kHz.

Sonically, the upper mid-range frequencies are the most important area for the intelligibility and clarity of the voice recording.

Remember that a clear sounding voice recording is really the ultimate goal of what we're striving for with all our recording and post-processing moves.

With that being said, I chose to do a cut in this area rather than a boost.

Interestingly, the only boost I ended up doing was in the high frequency area (orange knobs) out of all four EQ areas.

This leads to a helpful general principle of EQing voice recordings: cut before boosting for the best results.

The reason for the cut is because of the most troublesome, out of balance area in my audio - the sibilance issue I mentioned above, which I accentuated a bit when I opened up the top frequencies.

My exact settings are: a -1.5 dB cut at 4.6 kHz using a 3.5 Q on the bell filter (see image below).

Here is some of the thinking behind these settings:

• -1.5 dB cut - again, my sibilance spikes terribly in the 4.5 to 4.8 kHz area, so this is the first of multiple steps to deal with the sibilance:
  
  
• however, I didn't want to make enormous, broad cuts here to retain the frequencies in the general area around the sibilance...
  
  
• these frequencies give that all important clarity and intelligibility to my voice recording...
  
  
• I can use a dynamic EQ or de-esser down the line in post-processing that will only cut in this area when the "s" sounds spike above a certain threshold...
  
  
• this cuts the problem sibilance and preserves the good upper-mids we want.

• 4.6 kHz using 3.5 Q on the bell filter - 4600 Hz is the general mid-point of my sibilance problem (found using a free spectrum analyzer plug-in):
  
  
• and the 3.5 Q wideness of the bell filter was wide enough to hit some of the other sibilant frequencies above and below 4600 Hz...
  
  
• but, the Q of the bell filter wasn't so wide that it cut out those important upper-mid areas that I wanted to keep.

Listen, Twiddle and Sculpt

You can improve your audio for video, podcasts or sermons by using an EQ and following these simple steps even today!

Remember to listen and compare your voice recording to a good sounding reference recording first.

This helps you know where you want to go with your EQ sculpting.

Next, you can start with the upper frequencies where you can go for more sheen and clarity if your dialogue recording needs it.

Or you can leave it alone and cut lower frequencies that will allow the top end to shine through.

Recall that it's all too easy to overdo top end boost and end up with harshness.

Finally, deal very carefully with the next area of upper mid-range frequencies from around 1000 Hz to 4000 Hz since they're the most important for the clarity of your speech.

Remember, in general you'll get better results by cutting problem areas before boosting throughout this EQ process.

Have YOU used these steps on a recent voice-over for video?

How did they help? Let me know in a comment below!

Cheers, Bryan
Your Dialogue Editing Mentor