Setup for 'Broadcast Voice'

[Mark Nicholson]

Hi,

I am a total noob here, and the first project I want to try out is to take a single channel of spoken voice and give it that NPR, broadcast voice. Can you give some pointers where to look for guidance on this? I'm sure there must be tons opf material on this, but my mind is still blown by the sheer scope of Cakewalk.

- Mark

[CJ Jacobson]

7 hours ago, Mark Nicholson said:

Hi,

I am a total noob here, and the first project I want to try out is to take a single channel of spoken voice and give it that NPR, broadcast voice. Can you give some pointers where to look for guidance on this? I'm sure there must be tons opf material on this, but my mind is still blown by the sheer scope of Cakewalk.

- Mark

How your voice sounds depends on these things:

1.) Your room - Your room plays a HUGE role in how your vocal recording sound. Tune and treat your room so it is as flat as possible with no standing waves. The radio stations like NPR have their studios treated

2.) Your mic - The mic you decide to record with. The RE20 is a industry standard in the broadcast industry. So if you wanted that sound, I would use that mic.

3.) Your recording techniques - where you decide to record in your room and how far away you are form the mic decides in how your voice sounds in the recordings. The proximity effect is a useful tool. Learn it and experiment with it.  Since we are on the techniques, the position of the mix to yuor mouth should not be directly on axis. It should be slightly off axis and maybe pointing between your nose and mouth

4.) Your mic pre - The mic pre-amp accounts for about 5 to 10 percent of the sound. You can get away with using a mic preamp in a average sound card form MOTU or Focusrite in the price range between $200 to $600

5.) Effects - You should be using EQ and compression to get the vocals sounding just right.

6.) The way your voice sounds. If you do not have a 'radio' voice that is close to what you want it to sound like, then hire someone that does. 

[Bill Ruys]

Nothing at all wrong with CJ's comments, but having worked professionally as a VO artist, I have some differing opinions...

The design principal behind the Electrovoice RE20 was to, as much as possible, eliminate the proximity effect, so if someone was going to learn how to use it to their advantage, the RE20 would not be the mic to learn it with.  It's a great mic, but probably has less of a proximity effect than nearly any other cardioid microphone on the market.  Strangely, one of the most popular microphones in the VO studios I've worked in is the Sennheiser mkh 416 - and it's a shotgun mic!  As someone who has recorded music for over 30 years, this struck me as very odd, but the VO world is not the same as the music recording world - they play by their own rules.  The professional VO producers I have worked with have actually discouraged me from using proximity to bass-boost my voice for the commercial jobs I have done over the years - they prefer a more neutral recording they can EQ themselves.  Of course, if you're doing your own thing just for fun or recording your own podcast, go ahead and beef it up!  The same goes for compression - heavy compression will get you that larger-than-life, in your face sound, but if you are working with a producer, they won't much like you doing that - so the question to ask is: Is this solely for you, doing your own production, or are you trying to get professional work?  Because the way you attack this will be very different depending on the answer to that question.

I would also add that standing waves are a problem, yes, particularly for the listening environment, but not so much the VO recording environment.  That's why many of the worlds leading VO artists can get away with recording in vocal booths.  If you wanted to find the worst possible environment for room nodes, modes and standing waves, you couldn't do much worse than a vocal booth.  In truth, early reflections are a far worse problem for VO recording, because they cue your brain to the fact you are in a small space, so absorption is the best place to start in dealing with this.  Standing waves cannot be corrected with typical foam panels and even diffusors as they happen at frequencies to low for most acoustic panels to absorb.  Standing waves are mostly to do with room dimensions and geometry.  Starting out trying to fix standing waves is the wrong place to start for a hobby VO studio.  Deal with early reflections first, then isolation.  Then look at standing waves, but realise you can't fix them with foam, you will need to analyse the room and fix with room geometry and bass traps.

My tips would be:

• Set up a reasonable recording space, with good absorbent treatment.  If you can use a larger room, resonances due to standing waves, will be low enough that you can EQ them out - but don't get too obsessed with this.
• If your room acoustics are less than stellar, go with a large diaphragm dyamic micrphone (like the afformentioned RE20), or a Shure SM7B.  These are going to hear less of the room than a very sensitive condenser microphone.
• Learn how to use a gate, subtly.  Learn how to use a compressor by experimenting and you will soon understand why you needed to learn now to use a gate first
• Practice, practice, practice.  Your skill and style are more important than anything else.  Get people to critique your recordings.  When I first started in VO work, my producer pointed out the little things I was doing wrong, like not sounding the last consonant in a word I was over familiar with.  I couldn't even hear this until he pointed it out to me.
• In Cakewalk, learn how to slip edit and apply the FX you need (as per above).  Then learn how to save your FX chain so you can add it back to new projects quickly.

The reality is that, for VO work, you don't need to become a Cakewalk guru right away.  I have over 25 years experience in Cakewalk, but with VO work, you are not usually going to need to deep-dive into it's feature set.

[bitflipper]

Start with an industry-standard microphone such as the EV RE-20, a solid boom and spider (that's the elastic web that isolates the mic from vibrations), then apply all the acoustical absorption you can afford.

The BBC has long been the leading authority on setting up voice studios, having been the first to do so (going back to the 1930's) as they set up remote facilities all over the world for their world service. Lots of authoritative information is freely available online from the BBC on acoustic treatments. Many of their techniques involved a creative but science-based DIY approach, as they figured out how to use locally-sourced materials (e.g. bamboo diffusers) where things such as rigid fiberglass weren't available. If you're in the U.S. or U.K. then rigid fiberglass isn't hard to obtain, even if you have to special-order it from your local hardware store. You'll need lots of it.

NPR voice tracks are heavily edited to minimize noise (e.g. breath noise, lip smacks, air-conditioner wind, traffic) and then heavily compressed to assure consistent loudness. There is a whole category of audio editing tools aimed at post-production treatments. You'd do well to invest in iZotope RX8 Advanced. Yes, it's expensive, about a grand IIRC. But it contains many time-saving features geared specifically for VO work.

[John Vere]

I think you guys probably just scared the $__t out of the OP. They were probably just looking for a preset on Prochannel   

[user 905133]

1 hour ago, John Vere said:

I think you guys probably just scared the $__t out of the OP. They were probably just looking for a preset on Prochannel   

Maybe, but I appreciate the insights!

[Byron Dickens]

6 hours ago, CJ Jacobson said:

6.) The way your voice sounds. If you do not have a 'radio' voice that is close to what you want it to sound like, then hire someone that does. 

Kinda like "How do you get Angus Young's guitar sound? 1) Get a Gibson SG and plug it straight into a Marshall Plexi. 2)  Hand it to Angus Young.

[Craig Anderton]

2 hours ago, bitflipper said:

NPR voice tracks are heavily edited to minimize noise (e.g. breath noise, lip smacks, air-conditioner wind, traffic) and then heavily compressed to assure consistent loudness. There is a whole category of audio editing tools aimed at post-production treatments. You'd do well to invest in iZotope RX8 Advanced. Yes, it's expensive, about a grand IIRC. But it contains many time-saving features geared specifically for VO work.

There's plenty of great advice in this thread. As to specific editing techniques, I wrote an article about editing speech and narration that describes how to deal with breath noises, plosives, etc. It's aimed more at video people who need better audio so the images show the techniques applied to Vegas software, but a waveform is a waveform, so it's easy to translate the tips to Cakewalk.

I do a lot of narration for instructional videos, which is a bit different from "radio voice." However, here are some tips.

• Use a really good pop filter. I spent the $300 for a Pauly pop filter, which may sound outrageous (and it is!), but it's the best I've ever used by far. It gets rid of the pops but doesn't color the mic sound.
• One of the reasons room acoustics are important is because you don't want to be too close to the mic. The voice level drops off rapidly as you move further away from the mic, but the level of the reflections remains the same, so they become more prominent. As mentioned previously, moving closer to the mic isn't necessarily the answer, because of the proximity effects.
• Those acoustic shields that go around mics, like the Primacoustic VoxGuard, are a mixed blessing. If you don't set them up just right, they can make the sound worse. But they can also reduce computer noise and room reflections. I use one all the time, but you can't get too close to the mic or the vocal acquires the "sound" of the shield. 
• I do a ton of processing to clean up my voice. My favorite feature in iZotope RX7 is the mouth de-clicker, it's saved me hours. 
• I'm not a huge fan of compression to get an even level, because even a little bit adds artifacts. The phrase-by-phrase normalization technique mentioned in the article referenced above is my go-to for dynamics control. It's no different than just changing the level, so there are no artifacts. Then I need to apply only a little bit of limiting to control the peaks and give consistent dynamics, which nonetheless sound natural.

 

[Mark Nicholson]

Hey guys, fantastic replies, thank you.

Let me give you some more on me. I have a Tascam Model 12, which has fine preamps, EQ etc. It can be used as a DAW controller with Cakewalk, but I'm not up to that level yet. I record myself on the Model 12, and simply transfer files.

I used to do it cheap and cheerful with a Zoom H4n, but the Tascam and Cakewalk has opened my eyes to what is available. I'm definitely not a professional studio, and am aiming for the standard of 'competent amateur'. I'm looking specifically for advice on the microphone/pop filter side (I have done a lot on the room front already, probably as much as I can do), and also for what to do once the WAV file hits Cakewalk. I'm aware of, but have not yet used, the VX-64 Vocal Strip plug in, but I'm a bit overwhelmed by the sheer scope of what's available, and am looking for help in just cutting down the available options to the ones that I really need for voice production.

Thanks again, Mark

[msmcleod]

@Mark Nicholson - if you're just starting out, to be honest any decent mic will do.

@bitflipper mentioned the BBC... a lot of the presenters are doing their interviews etc from home nowadays.  I saw one a couple of days ago (forget which show it was, but I think it was the news) and the BBC interviewer was using an Rode NT-1.  This is a great general use LDC mic, and will work well with your Model 12.

Assuming you've got a half decent mic, you can get about 95% of the way there with a noise gate plugin, a compressor, and some EQ.

If you just want to use what's already "in the box", try the following:

1. Put the "Gate" style dial at the top of the ProChannel chain, and adjust it so it removes any background noise while you're not speaking. If the "attack" portion of your voice is being altered, you've turned it up too much.

2. Have the PC-76 compressor next, and start with one of the vocal presets to try to even out your voice. You don't want anything too harsh here. If you've got the PC-2A, or the CA-2A you could try using that instead. Try altering the input until you get the sound you want, remembering to adjust the output to compensate for any drastic drop in volume due to the compression.  If you can't quite get what you want here, or you're finding it a bit too harsh, try going back to the preset and adjust the dry/wet control.

3. Have the ProChannel EQ next, and pick the "Vocals - Broadcasting Podcasting Dialog" preset.  For this preset, adjust the low frequency gain to taste.

In the past I've had great results with the voice-over preset in Nectar Elements - this was an early version though, so I'm not sure if Nectar 3 Elements has the same preset ( I think it has a wizard instead, which is probably better tbh).

[Mark Nicholson]

Thank you Mark!

I have a related question. I recorded some of my voice and used MAnalyzer to look at it. Averaging over ~3 seconds gives me a spectrum like so:

Now, when I start equalizing, what is my goal? Do I want to see a flat spectrum? What would a good spectrum look like?

[msmcleod]

4 hours ago, Mark Nicholson said:

Thank you Mark!

I have a related question. I recorded some of my voice and used MAnalyzer to look at it. Averaging over ~3 seconds gives me a spectrum like so:

Now, when I start equalizing, what is my goal? Do I want to see a flat spectrum? What would a good spectrum look like?

An analyzer may give you an indication as to where frequencies are dominant (or not),  but that's really only any use when you know what you're looking for (i.e. you can hear something bad, but not sure where it might be). It can also be useful if two instruments aren't fitting well together, as you can identify where their common frequencies are.

But in all honesty, visual tools can lead you down a rabbit hole where you're cutting/boosting frequencies for no good reason at all, and in the end it just sounds awful.
 

My best advice is... use your ears, not your eyes!
 

[Mark Nicholson]

Thanks again Mark

[Max Arwood]

Please read- 

https://current.org/2015/06/a-top-audio-engineer-explains-nprs-signature-sound/

According to this article Neumann U87 is the most used mike.

RE-20 is very popular and is a  great VO mike but so is a Shure SM7 or 7b.  Shures  are also frequently used.  Heavy compression is a must but that adds noise between the syllable/sentences,  You might be able to gate some of this out but you might have to do a lot by hand (or with RX)  I usually use RX and volume envelopes. I also like to add a little saturation to make the voice pop out a little, especially if it is paired with a music bed.  Saturation / tube distortion will give the voice a big/full sound.

[msmcleod]

If you want to keep cost down, the Shure SM58 sounds very similar to the SM7 / SM7b, albeit lacking some of the detail in the high end.

You could maybe compensate by using EQ, or a better solution might be to mix it with a condenser mic.

[Alan Tubbs]

I usually use a gefell  930 for vo.  I’ve had success with the warm 87.  I’ve tried others but dynamics don’t have the detail I want, but that might be as much about the room as anything.  It is treated but a little dull and dynamic mics don’t shine with moderate speaking levels.  I also have used  a nice crown pcm for vo.  It was a little too much  surface and little depth - a bit too hard tho it sounded radio friendly.   the condensers are more expensive but do double duty on music instruments and singing.

[Craig Anderton]

I use condenser mics, because they cut through better on phones and tablets. I'm not sure that there are "rules" any more, it's every voice for itself

[Mark Nicholson]

That's a very good point Craig. I'm doing this voice work for my YouTube channel, and phones/tablets are the majority of the views.