Streamline Audacity Recording & Editing for Voice Overs

Dear Forum Members,

I’m Using Audacity 2.2.2 in a Windows 8.1 platform

I welcome tips anyone might like to offer on streamlining the recording and editing process when recording voice overs using Audacity, as I feel that the way I do it is quite clumsy and time consuming.

Perhaps the solution is obvious (to everyone except me), but although I have searched extensively online and read the relevant FAQ points, I am still at a loss.

My current recording process is:

1. Set levels: speak some test text and adjust in order to not red-line (no problem here).
2. Press the record button and begin (also OK).
3. When I take a BREATH periodically, I use 2 techniques:
   a) Stop speaking briefly to leave some silence, take a breath and then wait a bit to leave some more silence - this allows me a clear waveform target to delete during editing, as well as keeping the breath a safe distance from the text being recorded.
   b) Stop talking briefly to leave some silence, pause the recording, breathe, restart the recording, wait a bit more to leave some more silence, then start talking again - this also allows me a clear waveform target to delete during editing, and a gap from the text.
4. When I make MISTAKES, I immediately clap my hands close to the mic (to purposely red-line the meter) and then I leave a few moments of silence before proceeding with the recording - thereby facilitating the waveform localization of the glitch for deleting during the editing process, as well as a safe gap.
5. I then perform noise reduction using that effect as per normal instructions (works very well).
6. Then I listen to the entire recording to see if I missed any glitches (breaths, clicks, etc.). If I find any, I copy a bit of the silence and paste it over the glitch.

Well, that’s it in a nutshell. I offer my thanks in advance for any orientation.

Best regards,
Kevin

Good thing you’re not reading for audiobooks. The process is longer. Those people have very strict technical quality standards.

No, you’re doing what the rest of us are doing. Somebody put together a de-breathalizer. tool. I’ll see if I can find it.

No doubt you figured out already to re-read a whole sentence or two when you fluff. It’s killer difficult to split a sentence with a correction.

When you get to the end of a raw reading, do you Export a separate WAV for protection? Put it in a separate place and don’t touch it. Edit a copy. If Audacity or the computer goes into the dirt while you’re editing, it could take the reading with it. If you have no clean backup, you’ll be reading the whole thing again.

I’m not clear what you want Audacity to do. Do you want Audacity to stop you from making mistakes? That would be a popular button. I wrote an April Fools story once like that.

New filter that can turn the trash you shot into a professional voice track!!

The instant you stop to edit a show, you hit the 5X rule. Editing will take five times the length of the show.

Do you know what Noise Reduction is taking out? Is it something you can prevent at the recording step? I have an insanely quiet third bedroom and if I record at night with the best microphone, I don’t need noise reduction.

If you do use Noise Reduction, don’t overuse it. It’s a common mistake to try and reduce background noise too much. This tool used to be called Noise Removal. Too many people expected it to remove noise. All of it. To zero. That makes voices honky, tubey and wine-glassy and it’s rough to get good quality open voice sound. That can’t be fixed later. Once you do that, it’s all over.

I’ll look for the de-breathalizer.

Koz

Oh. I think it’s still sticky tape and paper clips experimental.

https://forum.audacityteam.org/t/debreath-in-audacity-proof-of-concept/48934/1

How close are you to the microphone? Can you turn away or look down and then gasp?

I think there’s room here for a cough button. On-air broadcasters frequently have a cough and belch button that mutes the microphone as long as it’s held down. At least once during every announcer’s career, they accidentally put a book on the cough button. If you have a green engineer, you could be off the air until the director of engineering got there.

I wonder how hard that would be to program… P Pause won’t do it. You have to hit that one twice and it will throw the story timing off.

Koz

The store-bought DeBreath plugin is [u]much[/u] better,
but I can’t get it to work in Audacity in Windows (Vista 32-bit), but apparently it is possible on more modern Windows & Mac.

Hi Koz,

Thank you soooo much for taking the time to write such info-packed replies. Here are my responses:

Re: Good thing you’re not reading for audiobooks. The process is longer. Those people have very strict technical quality standards.
Got it. Will leave that challenge/opportunity for a later stage.

Re: No, you’re doing what the rest of us are doing.
Oh my… just got to buckle down then.

Re: No doubt you figured out already to re-read a whole sentence or two when you fluff. It’s killer difficult to split a sentence with a correction.
Roger that.

Re: When you get to the end of a raw reading, do you Export a separate WAV for protection? …. Edit a copy.
I make a copy of the raw reading Audacity Project and then edit that. I will start exporting a separate WAV of the raw reading as well.

Re: Do you want Audacity to stop you from making mistakes?
Well… I didn’t expect that it could. But wouldn’t that be nice?

Re: The instant you stop to edit a show, you hit the 5X rule. Editing will take five times the length of the show.
Roger that. Stiff upper lip…

Re: Do you know what Noise Reduction is taking out? Is it something you can prevent at the recording step?
I can’t figure out if it is removing the room noise or noise in the equipment - although the room sounds reeeal quiet to me, so I suspect the equipment. When I am in the studio room (a walk-in utility room that is somewhat sound conditioned) with the equipment all ‘fired up’ and the mic in the “sweet spot,” I note that there is noticeable background noise. When I turn down the “sens” (sensitivity) knob on the audio interface (I am using an AKG P220 microphone and a Roland UA-25EX Audio Interface - both quite used) the background/hissing/white noise goes way down, but then my voice is not very audible. This leaves me wondering if the audio interface (or mic) is a naturally noisy (internal static) one.
After doing some online checking, I think the culprit might be the audio interface’s internal noise.
Other possible sources are the rather old connection cords and the laptop’s fan. (BTW, everything is running off of the laptop’s battery power, so there are no power cables plugged in to wall sockets to cause that type of interference.) I will also move the audio interface away from the computer.
Regardless, after I do the recording, I currently use the standard Audacity Noise Reduction setting and the result sounds great to me and, so far, my clients are liking it.

Thanks for the link to the de-breathalizer.

Re: How close are you to the microphone?
About 7-8" (18-20 cm), what I understand to be the recommended distance, and it gives me a pretty good “sweet spot.”

Re: Can you turn away or look down and then gasp?
Well, I experimented with that and it didn’t go too well.

Thanks again for all your help.

Very best regards,

Kevin de La Tour

I can’t figure out if it is removing the room noise or noise in the equipment

Record a 20 second sound test and post it here (This is all optional.)

http://www.kozco.com/tech/audacity/TestClip/Record_A_Clip.html

We can do sound analysis and give you an opinion. Export the work as WAV (Microsoft) and do nothing to it before you send it.

The fuzzy rule is to maintain good volume into Audacity and if it’s too noisy then you start to worry about who’s making the most noise.

Your voice should peak on occasion about -6dB to -10dB on the bouncing sound meter. That works out to half-way and a bit less on the blue waves.

That’s optimum.

Notes:

The AKG is a side-fire microphone. You should be talking into the grill just up from the company name. -20dB Should not be selected.

If I got the interface model right, I would not be shocked if you had to run the “sensitivity” control all the way up to get good volume. That’s very common on home style equipment.

Too high volume causes instant and permanently fatal sound damage. Too low volume may be fixable in post production and everybody thinks it’s their fault. Only one of those conditions causes you to send the interface back.

Koz

Hi Koz,

Thank you for your meticulous reply.

To my naked ear, my studio space is “dead quiet” when I am ‘holding my breath and not moving.’ But when I get the volume up to a suitable level, the consistent background noise fluctuates within about -40dB and -50dB on the bouncing Audacity sound meter.

What I have tried thus far to eliminate the background noise:

• set up the studio with a SAMSON C03U USB mic (to eliminate the possibility of the noise coming from the AKG P220 mic or the Roland UA-25EX Audio Interface
• tried the USB plug in all 3 USB sockets
• turned off the overhead light
• unplugged the wireless mouse

The background noise when I am ‘holding my breath and not moving’ is identical when I use the SAMSON C03U USB mic. This would seem to leave me with 3 possibilities:

1. The USB cord needs to be replaced (will check that)
2. I have some knobs/sliders/buttons in the wrong position.
3. That’s just the way things are.

I will send the WAV files as requested.

BTW, I do always keep the bouncing Audacity sound meter in the yellow at most (in peak moments).

Thank you so much for all your help.

Best regards,
Kevin de La Tour

Hi Koz,

I have uploaded the WAV file for analysis as requested.

As well, I tried using a different USB cable with both the AKG P220 and the SAMSON C03U mics, and was able to determine that the USB cable is nots the source of the background noise.

Also, I have reread the Roland UA-25EX Audio Interface manual and it seems I have all the knobs and buttons in the right positions.

Re: I would not be shocked if you had to run the “sensitivity” control all the way up to get good volume.
In order to get a ‘suitable’ recording level, I have the sensitivity level of the Roland UA-25EX Audio Interface at three o’clock (there are 10 tick marks on the sensitivity knob, three o’clock puts it at the eighth tick mark).

In case it is relevant, I offer the following notes about my mic placement: subsequent to reading articles and experimenting, I angle the body/shaft of the mic downward (in a SE orientation) so that the mic is ‘dangling’ from above and to the left of my head, and I speak over the top edge of the mic. This seems to be as effective as speaking directly into the grill, but eliminates the need for a pop screen (I’m not using one and consciously curtail my 'p’s and 't’s anyway), and allows me to see the script on the laptop screen.

Thanks again for your assistance.

Best regards,

Kevin de La Tour

Nothing wrong with that. I ran it through straight Audiobook Mastering and this came out. No noise reduction.


https://forum.audacityteam.org/t/audiobook-mastering-version-4/45908/1


I have a production complaint. Being an instructor doesn’t mean you have to go slow. This is what it sounds like to me, and the clip doesn’t go beyond 20 seconds. A whole class of this would be deadly.

Koz

I can’t deal…

Koz

That may seem like I just speeded it up and added music.

Nope.

The phrases are Effect > Change Rhythm, so that the timber, pitch and expression remain constant, but the words go by faster. I did that twice, 10% faster each time. I know that’s not 20%, but I’ll do the math later. It finally started sounding like natural speaking rhythm.

Then, to accentuate the instructional rhythm, I added spaces between the phrases. So even though the speech is faster, the overall duration isn’t.

Please note I removed the word “Now” about 2/3 through. That suggests you’re trying to make it up as you go and you need the spacer to gather your thoughts. You should be the teacher, no gathering needed.

That’s “Sandstorm” in the background to push the upbeat tempo and yes I know that’s trite “exercise music,” but I couldn’t restrain myself.

Koz

Hi Koz,

Thank you for all your input!

Re: Nothing wrong with that. I ran it through straight Audiobook Mastering and this came out. No noise reduction.
& RAW_SAMPLE_20SEC-AudiobookMastered.wav
I gather that the sample I sent would be acceptable to ACX. Please forgive me if I sound too picky/demanding, but both my original file and the ‘acceptable’ file you ran through Audiobook Mastering sound really noisy to me.

I have uploaded my original file after running Audacity’s Noise Reduction effect with the following parameters (which I believe are standard):
Noise reduction (dB): 12
Sensitivity: 6.00
Frequency smoothing (bands): 3
To me, this file sounds groovy: quiet (zero background hiss) and my voice sounds the same (no quality loss) as the original recording. Isn’t that preferable?

Now, although the “noisy” file I had first sent is apparently acceptable to ACX, I am still distressed at the -40dB to -50dB background hiss that I perceive - and that Audacity is faithfully recording.

Presuming that there are no defects in the equipment I am using - both the AKG P220 & Roland UA-25EX combo and the SAMSON C03U, as well as the cables - I can only conclude that my studio is somehow noisy - even though it sounds entirely quiet to me. After all, we are supposed to aim toward a -60dB noise floor, correct? Do you have any recommendation or am I analyzing this incorrectly?

Re: RAW_SAMPLE_20SECMasteredAndTemporallyCompensated.mp3
Sorry, I didn’t get the point of that file.

Re: Barbells-Master.wav
I think I’ll go off to the gym now…

I thank you kindly for all of your input and hope I have not been too picky with what I see as a problem, but I do want to provide the best voice over product that I can.

Very best regards,

Kevin de La Tour
https://www.linkedin.com/in/kevin-de-la-tour-873a27161/

both my original file and the ‘acceptable’ file you ran through Audiobook Mastering sound really noisy to me.

I use Audiobook conformance as gold standard since their specifications are stiff and difficult to meet. Are you turning up the volume between paragraphs? You’re always going to find noise if you do that. The general goal is to hear little or no noise at normal listening volume.

If you need to get better than that, then you may need Noise Reduction or even noise gating.

To me, this file sounds groovy:

That’s what counts. If you find settings that produce good work, write them down and use them.

the “noisy” file I had first sent is apparently acceptable to ACX,

Not exactly. I ran your file through the AudioBook Mastering process. It succeeded and passed the ACX Audio Standards. It’s important I did it all without Noise Reduction. That’s difficult and not many people can do that with their home system. You can certainly add noise reduction at the end and get better yet.

Please note that one goal is to use as few corrections as possible. You may run into bookkeeping problems. Yes, you can apply a handful of corrections to all your readings, but you have to remember to do them all every time you read and in the right order.

it sounds entirely quiet to me.

That’s because it is entirely quiet. As I said up the message, it’s very difficult to meet audiobook standards with just a few volume corrections, and your system is good enough to do that. Leave your system alone.


This is editorial comment and has nothing to do with Audacity.

Sorry, I didn’t get the point of that file.

That’s what your presentation sounds like to me. It sounds too much like you’re trying to talk to a four year old even without the speed change.

I can’t watch most new user instructional videos and I suspect most other people can’t either. If you watch the successful YouTube channels, they all talk about 10% to 20% fast. I can’t prove this, but I suspect that’s the speed most people read the printed word.

The second clip I posted has the paragraphs reading roughly 20% faster with the spaces between paragraphs longer. The stupid music is optional.

That sounds more comfortable to me and something I could listen to for an extended time.

I can’t tell if that’s your normal speaking style, but if you’re acting like an instructor, it’s my opinion you should stop.

Koz

A note about Noise Reduction.

Listen carefully to your reading before Noise Reduction and then after. Listen to the “S” sounds and make sure there are no ssss boosts or gritty sounds in the words.

One symptom of too much noise reduction is voice that sounds like you’re talking into a wine glass or milk jug. This is also the cellphone sound and for a lot of the same reasons. Cellphones use stiff noise reduction.

And you’re only going to hear any of these quality problems if you’re listening on a good quality sound system or headphones. If you can hold your sound system in one hand, that’s probably not the best idea. Avoid laptop speakers.

Koz

Hi Koz,

Thanks again for your replies, they have made things very clear.

Re: Leave your system alone.
Roger that.

Re: That’s what your presentation sounds like to me….
Cognizant of all that, but followed the client’s instructions to the letter - and the customer is king, right? Got a 5-star rating and a r-a-v-e review from that job, so…

Re: A note about Noise Reduction….
Great input, thanks. I use headphones.

I think we have taken this analysis to completion and your input has been extremely helpful. Thanks loads!

Best regards,

Kevin

r-a-v-e review from that job, so…

So you win. The crazy-making clients are the ones that have no idea what they want.

Re: Leave your system alone.
Roger that.

I can’t stress that enough. Most new users trying to read commercially would kill to be able to record with that quality.

I use headphones.

Which ones and why?

Koz

Hi Koz,

Re: I use headphones. / Which ones and why?
Just $10 Phillips headphones, because that’s what I have. Interestingly, Bill Dewees downplays the importance of high quality headphones, saying he uses ear buds, positing that VO artists can get too “into” hearing the dulcet overtones of their own voice through headphones…
And so far, so good, it seems. Perhaps good quality speakers later…

In case it is useful, I devised a somewhat ‘hacker’ method of speeding up my editing process:

1. After completing the VO recording and running the Noise Reduction function, if that is the case (and perhaps removing some very obvious glitches and mis-takes) - and, of course, making a backup copy of the totally raw and this ‘semi-raw’ Audacity Project,
2. Make another copy of the ‘semi-raw’ project and name it “COPY FOR PASTING SILENCE,” for example.
3. Open both the original ‘semi-raw’ project for editing, as well as this “COPY…” project (as you know, you can have 2 or more Audacity Projects open at the same time).
4. Since we have a good 10 seconds of pristine recorded silence at the beginning of the project, as one goes through the recording and wants to “wipe clean” little noises, such as breaths, mouse clicks, etc., from the gaps between phrases/sentences/paragraphs, simply measure the duration of the entire gap or the ‘dirty’ part of the gap (e.g., 2.75 seconds), flip back to the “COPY…” project select the first 2.75 seconds of pristine recorded silence, copy it, then flip back to the project being edited and paste it over the ‘dirty’ gap/portion.
5. And Bingo! You now have a pristine 2.75 seconds of aural suspended animation where there once was grit!
   I find that doing this for any gaps/portions that need attention (which is most of them) really speeds up the editing process.
   Hope this is helpful.

Best regards,
Kevin

Bill Dewees downplays the importance of high quality headphones, saying he uses ear buds, positing that VO artists can get too “into” hearing the dulcet overtones of their own voice through headphones…

That’s a new user problem, yes. The flip side of that is the way I present which is flat and expressionless with the risk of putting everybody to sleep. I’m rated slightly better than warm milk.

You can get into a religious war with earbuds. “Most people are listening on earbuds so that’s what I should do.” I don’t agree. Don’t follow a manufacturer’s opinion into the abyss.

You can get into trouble with earbuds. They change tone depending on how they seat in your ear and they don’t always form a tight enough seal to keep your headphone sound out of your microphone. As an aside, I use custom earbuds for hiking because I can’t stand the sound of the ones that came with my iPod.

Perhaps good quality speakers later…

You may be beyond the need for those. Those can be valuable if you need to find out why your noise is so high.

[poster] “I can’t pass noise.”
[me] “I hear air conditioning rumble in the background.”
[poster] “What rumble?”

After completing the VO recording

Good idea. A version of that by simply recording extended Room Tone at the beginning and use that instead of juggling multiple shows. It’s faster and doesn’t take up as much system resources. I shot a broadcast presentation in Los Angeles and the producer (in some far-off land) recommended I record a healthy chunk of Room Tone before anybody said the first word.

We have a difference of opinion about the raw recording. I Export a WAV (Micosoft) 16-bit stand-alone, common sound file when finished presenting. If I’m feeling obsessive, I’ll copy it to a thumb drive or other location.

Scan over the forum for the people who can’t open a Project after they recorded it. Yes, there was a recent bug which made damage much more likely, but Project failures happened before that as well. WAV files are pretty robust as long as you don’t go much over about three hours stereo (six hours mono) in one go.

Koz

Hi Koz,

Thanks for your comments.

Re: "You may be beyond the need for those [speakers]. Those can be valuable if you need to find out why your noise is so high.

Got you. And I’m still trying to find out why my noise is “so high.”
Is the system picking up cosmic microwave background radiation?(?!)
I applied to work with www.voicebunny.com recently, for example, and they require a noise floor at or below -60db. They also specify “Please do NOT use any compression, noise reduction, or other kinds of post-production effects.” Whereas I find no alteration in my voice per se in the raw and noise reduced recording, they don’t allow noise reduction and require -60db in the raw recording… Ouch.

If I submit a voice over directly to a radio station (for example) will I be able to run noise reduction on my recording prior to sending it in?

Re: “A version of that by simply recording extended Room Tone at the beginning and use that instead of juggling multiple shows.”

1. I used to do that, but I personally find it easier and faster to juggle between the 2 shows as I don’t need to return to the beginning of the show to copy some silence and then scroll back to the location I want to paste to.
2. I will follow your recommendation of making a WAV of the raw and finished recording.

Best regards,

Kevin

Hi Koz et al,
I’m trying to start a new topic, but can’t seem to do it (can’t find the “new topic” button).
If someone could send a lifeline it would be much appreciated.
Sorry for my apparent denseness.
Best,
Kevin