compressor EQ and limiter

After I've done all the narration, I make sure to go through the following editing.

Noise Cancellation (Absolute MUST to have decent quality)

Equalization (look on youtube for walkthroughs of which settings give you a good EQ curve for voice)

Normalization

Compression

Then re-do the Normalization.

Quality comes out much better than the raw recording before edits.

Theres a pretty good Google doc that gives a general tutorial (+ links to required FREE extensions) on how to get your audio to sound nice post-recording but it leans towards a more bass-ier voice, so take it for what you will.

https://docs.google.com/document/d/1fmS0Wp6EantjSmQiBnPWyCCOSNpiawjgCOdUJknKbnE/edit?usp=drivesdk

1. Get a mic that is not the Blue Yeti (I love the Rode Podmic USB)
2. Get it about three finger widths from you and point it at the corner of your mouth.
3. Have a windscreen/pop filter
   Check out Mike's Video https://youtu.be/WsmMMKRZp5g?si=dUsIEp7HjXEkvOOi

Moderator Required full disclosure: I am the head of Podcasting at Podpage and the founder of the School of Podcasting.

Define "professional". If you are talking about why your voice doesn't sound like you THOUGHT it should sound, no.

That is just what you sound like. We all went through it.

Noise Reduction

Did you have excessive background? Why? You should eliminate the source of the background noise before you record.

EQ

Bass Boost Treble Boost Low Rolloff for Speech

This advice makes no sense. If you are boosting the "Bass" frequencies (20hz - 200hz) then you will offsetting them when you boost "Tremble" (2khz - 20khz) and then you will be removing them with low Rolloff speech preset [cutting everything below 120hz).

https://www.masteringthemix.com/blogs/learn/understanding-the-different-frequency-ranges

btw, the human frequency range is between 80hz to 3500hz .

https://flypaper.soundfly.com/produce/eqing-vocals-whats-happening-in-each-frequency-range-in-the-human-voice/

All in all the those settings are wrong and for a shitty recording to sound listenable but not necessarily good.

Normalize

The normalize tool and the amply tool work the exactly same way and do not work as people believe. They raise the volume of the whole selection based on the highest peak volume NOT average volume. Essentially normalize is a worthless tool not needed, just use amplify.

Compressor

Compression is art not a science. When it comes to compression for Speech you have to determine what you want to achieve. Eliminate peaks? Use a limiter instead.

https://youtu.be/likk4fbEblg?t=145

Even out the sound, make the louder parts quieter and sound at same volume as the quiet parts of the audio?

First determine the volume range of your audio,the highest peak in your audio and lowest volume of your audio.

Then turn off make up gain in the compressor. Set the threshold around 5db below your peak with Ratio of 2:1. Speech really does need a lot of compression. In fact a good recording really shouldn't need any.

Do not worry about attack or release for speech it's really not needed.

If you on linux you just use Barry Satan Maxizmizer, it's original name was Stupid Compressor. On Windows/Mac The Leveller setting in the Distortion effects tool works very similar to BSM but not as aggressive.

Low Roll-off for Speech is good for most voice recordings as it removes (unwanted) low frequency rumbling noises. Unless you have an exceptionally deep voice it is unlikely to do harm. As it is a high-pass filter (a low-cut filter), it also removes (unwanted) DC-offset (if present).

A little bit of dynamic range compression, and / or peak limiting is often beneficial, though skilled voice-over artists are often good at maintaining a consistent level, reducing the need for compression. I've had the pleasure of working with some very good voice-over artists, and they make the engineers job (that's me) so much easier.

Other than that, any other effects depend on the voice and the recording.

It is very easy to overdo effects. I would highly recommend minimising the use of effects for voice recordings.

For home studios, processing voice recordings is largely about compensating for deficiencies in the equipment and studio environment, For example, if your microphone is a bit dull sounding, then you may need to boost the high frequencies a bit, but that will also tend to boost high frequency hiss which may then require noise reduction, whereas if using a very good microphone and pre-amp, the recording is more likely to be bright enough to not require any high frequency boost, and the level of hiss from mic + pre-amp are likely to be low enough that noise reduction is not required.

While it is possible to make a good recording with modestly priced equipment, it is easier with really good equipment, (and impossible with bad equipment). I worked with low-budget equipment for many years and had to do the best with what I could afford. The up-side was it taught me a lot about sound and recording.

Probably the most important and underrated requirement for good quality audio recording is the room. The recording room needs to be extremely quiet and not at all echoey. Extraneous noises and echoes are near impossible to get rid of without damaging the sound.

There are no "best settings" that work all the time.

The settings which would most do what you want depend on your individual voice, the particular recording environment you're in, the hardware you're using, and what you want out of your final sound. Anyone just using the settings shown on some YouTube video or something else without actually understanding them is just importing nonsense which may or may not apply.

The only way to discover what the best settings for you are is to experiment with every setting. Record yourself, listen back, tweak things, and do it over again. Do this a whole lot, tweaking different things each time. This is how you learn what the different things actually do, which ones have an effect you happen to want, and how it all comes together.

https://forum.audacityteam.org/viewtopic.php?t=45307#p113447

Lots of good advice here. I do not like Audacity's Noise Reduction in general. If it's used lightly it can be ok, but if you need a lot of noise reduction, it really harms the sound quality.

So first reduce the noise generated by your gear (expensive). Then reduce the noise heard by your gear (inexpensive). Then use as little noise reduction as possible.

I have had much better experience with Audition's noise reduction but I've left Adobe on non technical principals and can't recomend, but it suggests there may be other better plugins out there to explore.

The main thing you have to do is find and remove the sources of noise.

Treat the room, even if you have to use couch cushions, mattresses and moving blankets. Or move your mic into a more sound isolated space, like a closet full of clothes in the center of the house.

This is assuming your mic and interface are of good enough quality that they're not introducing noise. If that's not the case, replace them with something better.

Once you do those two things you'll be able to use much less aggressive noise reduction settings and it won't alter the sound of your voice nearly as much.

Try running a real time EQ plugin. Use a super narrow bandwidth with the level boosted fairly high. Start at one end of the frequency spectrum and slowly sweep the bandwidth to the opposite end of the spectrum. Listen for areas where the offensive sounds/background noise seems to get louder. Once you find the worst area, turn the boost way down. This isn't a great method to rely on, but in certain situations it can be helpful.

I don't know if this will help with your specific problem and I'm sure I'm not explaining it properly, but if you want to look into it try searching YouTube for videos on "EQ sweeping" or "Seek and Destroy EQ technique".

Noise reduction is a destructive process. So less is more.

The best place to invest time and money is to learn how to fix your audio before you even hit the record button.

That means identifying sources of noise and figuring out what you can do to reduce or eliminate them. Another thing that beginners miss is how to treat your recording space with “broadband” sound absorption. I like to say that I’d rather use $100 mic in a properly treated room than a $5000 mic in a closet.

If your recording is good, then you don’t have to do much post production at all.

That being said, Izotope RX has a bunch of great audio restoration tools for cleaning up your audio.

Supertone Clear is a nice plugin for eliminating general background noise.

Accentize has a bunch of noise and reverb removal plugins for professionals. They are used in TV and Film a lot.

With any kind of noise reduction I would do a little at a time. If I wanted to reduce noise by 12 db. I would do 3 passes of 4 db instead of one at 12 db.

I've recently started using audacity with a Sennheiser profile USB mic and a pretty quiet room to record audio narrations up to an hour long per session. My workflow is :

1. Noise reduction (not too much, sample a few seconds of "silent" recording first and then apply only once)

2. EQ. This one takes a bit of trial and error as every persons voice is different so you might have to just see what sounds best.

3. Normalise (I just use the default level of -1db)

4. I then apply noise gate and then listen through and remove any clicks or pops. This takes the longest amount of time.

5. Final step before exporting, I do Loudness Normalisation to -16db to -19db depending on thr recording itself.

Its quite a few steps but thats what I find yields me the best results.

But also if anybody else has any feedback or tips on my workflow that would be great.

No idea what they mean by “Input gain”. Set the project sample rate to 48 kHz, and resample each track to that if necessary. There's an option in the Tracks menu to mix stereo down to mono. When you export, you can set the precision. As for normalisation, that's a strange requirement that I don't see the point in, but once you have the mono track, use the amplify effect so that it peaks at -6 dB.

Without knowing what kind of mic and how good it normally sounds; or what you are plugging it into; a mixer, then usb out into a computer running Audacity?
I would say, in general put a compressor on the whole voice track (2:1 ratio for starters) and then use the Bass/Treble effect to probably boost the treble +3db while reducing the bass -3db and that should give extra crispness and bring the syllables out to make it clearer when speaking

Thanks I will use these - needed simple steps like this otherwise shit just goes over my head

instead of bass/treble boosting i'd suggest reducing mids instead. you see additive eq tends to introduce a bit of distortion in those areas, it's better to reduce "undesired" range rather than boost "desired" one, and then normalize overall gain as normal

You don't need to record to -16, you need to normalize to it.

I use scoring behind dialog on my show, and find that having voices about 4 dB louder than the scoring makes it so I can hear both at about the right level.

If you're having several people talk at the same time, or if you're having trouble keeping the spoken bits at the same volume across sections, perhaps because of variable volume during the performance, I find that using a tool called a broadcast leveler can really help flatten things out. It's like a multi-stage compressor but built to keep things within a fairly narrow range of volumes. I use tri-leveler 2, it is good and also free

Here's my setup in my incredibly noisy room;

Compressor

Bass Boost EQ @ 6db

Trebble Boost EQ @ 3db

Normalize

Noise reduction between 10 and 17db (yes my room is super noisy)

Try using this as a bassline and changing different steps to see what sounds best

– M

Honestly hearing it will probably make it easier to help you, but here are some basic things you could do.

​

A noise gate on the dry track if the room is very noisy. This should help a little when you begin to compress the track, which we will get to. Also make sure of the input gain of the mic - ideally you want the input level to be between -12dB to -6dB. Too low and the noise will overwhelm the voice, too high and it will clip and distort, particularly noticeable once you compress.

​

Now onto EQ. A basic thing you should always be doing to spoken word is to put a high pass filter and cut EVERYTHING below 70-100Hz. This is dead sound that only causes mud. It will also improve the clarity, and again, will assist when you compress. Cutting between 300-500Hz if there is still mud in the voice and cutting 800-1000Hz if the voice sounds particularly mid-ranged.

​

If the S sounds in your voice are overwhelming or loud, look into de-essing. This is an easy to follow method to get SpitFish to work in Audacity. It's a plug-in that will be all you ever need for spoken word de-essing. Ideally you want the S sound to NOT overwhelm straight from the source, meaning you want to put a beanie on your mic and could also try a pop shield. Pop shields are great as they also help you maintain a level distance from your mic.

​

Next, compress compress compress! A decent amount of compression will help for clarity in the voice. Medium timed attacks and medium to slow release. Adjust the threshold based on your track (this is where hearing it would help!) but you can try anywhere between -4 and -10 on the threshold, until the quieter bits are properly audible and the louder bits like saying S don't kill your ears. You will find that this will really increase the overall volume. This is where we used the makeup gain to adjust the overall volume. You could also use the track volume to get the overall level more appropriate.

​

Lastly, some good reverb to top it off. Light room reverb on spoken word can make all the difference. This is however a very "ear focused" thing so I recommend slapping some reverb on it and adjusting it until you get something you like. "Vocal rooms" type reverb can be applied and they sound really good, similar to how a glass vocal booth may sound. And "Cathedrals/Churches" type reverbs are also great for spoken word, as it adds a nice adjustable tail of sound to what you are saying. I like to use the latter in gameplay videos, albeit very very lightly! Good luck, just mess around, reverb is awesome.

​

Just some things I generally do for any spoken word track I have, before I make adjustments from there. :) Hope this helps.

Blue yeti owner here, have tried many things with audacity and tbh never really see a difference from the original audio. Could just be a noob though

Now this isn't Soo much an Audacity thing as it is a recording thing so I hope it helps.

While nothing you can really do will completely eliminate reverb and weird noise from the recordings there are things that can help. If you're able to try getting a foam mattress topper from somewhere like Walmart and put pieces of it in the ceiling corners. Majority of the time sound likes to bounce around and amplify there specifically.

If you can't do that try throwing a blanket over your head. You'll look a bit odd but it Helps muffle the sound from echoing back into the mic. This one depends on what kind of video you're making. If you're doing a gameplay video skip this.

As far as Audacity goes. Alot of people can give you their settings for their voice but that may not work for you. Use other people's settings as a base then slowly tweak until you figure out what works