Steps to improve a clip that is only vocals?

GoldWave general discussions and community help
Post Reply
treefitty
Posts: 2
Joined: Sun Nov 13, 2016 5:02 pm

Steps to improve a clip that is only vocals?

Post by treefitty »

Hello, I've used GoldWave before for modifying sound effects, but I've never used it to deal with recorded voice before. I've used some of the simpler tools like the normal EQ, pitch, reverb, echo, but I've never really messed with things like: dynamics, compressor, highpass, bandpass, and definitely not the parametric EQ (I can't figure it out)

From what googling a bit tells me, I need to use a lot of those tools to improve voice quality. I have a voice clip but it has a few problems:

1.) First, it needs to be a little louder. I recorded it with the mic as close as possible, but it's a basic headset, not a high quality one. I realize that the best thing to do is record with better equipment before you even go into the audio editing program, but I'm working with what I have. When I listen to my sound at 100% vol, it's still not as loud as e.g. a YouTuber's voice over, however, I feel like just increasing the volume is wrong - it starts to sound "blown out" even though it's the same volume as the YouTuber's voice's volume. If this is unavoidable because of mic quality then that's fine, but my main point here is that I do need to increase the volume a bit, as well as:

2.) Being that my mic needs to be close to my mouth to have the volume okay, SOMETIMES there is the screaming/hissing "S" or sibilance I guess it's called. It doesn't happen on every single S, but it happens more than enough times that manually copying and pasting everywhere it happens isn't a great option (it's 30min of voice). I've read things that say things such as, "boost 3-6khz" or "put a bandpass on 2-4khz where sibilance usually lies". Can someone actually explain the exact steps to accomplish this? I tried going to "bandpass/stop" and setting it to 2000-4000, and it made it super quiet and sounds like I'm talking through a can-and-string phone. What's the correct way to do this?

3.) I have a somewhat deep voice, it's not insanely deep but I noticed/read that a lot of male voice recorder's add bass to their voice to make it sound better. I know there is software/hardware that adds this on the fly, but what's the best way for me to do this with Goldwave? Just use the normal EQ? I saw the bass boost option but I couldn't quite get it to sound as bassy and nice as other examples, but, again, that could just be my mic quality limitation and if that's the case I can't complain.

Here is a voice clip:
The forum is preventing me from posting a link because I'm a new user so sorry about this but the "link" is:
https_instaud_io/Cjo (where the first _ is a colon double slash, and the second _ is a period.)

I picked a clip that has the hissing "S"... "SSSSave" and at the end of "Townssss". But then later, the "S"s in loads/saved/transported sound ok.

Basically, I'd like to know how best to accomplish these things and in what order? Which do I do first? Do I maximize volume and compress between each step? And also if there is any other thing one "should" do to a voice clip to make it sound better, etc. Thank you for any advice!
DougDbug
Posts: 2172
Joined: Wed Feb 16, 2005 3:33 pm
Location: Silicon Valley

Re: Steps to improve a clip that is only vocals?

Post by DougDbug »

I think the main problem is your headset mic. A podcast mic will probably solve most of your problems. The downside to a studio/stage mic is that you'll have to back-off a bit, and with your mouth further from the mic you'll pick up more room noise.

The trick to going louder without distortion is compression/limiting. Almost every commercial/professional recording has some dynamic compression. Try the Reduce Peaks preset and then bring-up the overall volume with the Volume Maximize effect.

There is an effect called a de-esser but you'd have to find a 3rd-party plug-in and I'd try reducing some of the higher frequencies with the equalizer first. (Bandpass & Bandstop filters are too extreme because they try to completely remove the sounds outside of the pass-band and you need most of the frequency range for natural-sounding voice. (You can get-away with removing the deep bass.)

Your voice sounds good to me and I don't think you need more bass.
DewDude420
Posts: 1171
Joined: Fri Mar 11, 2005 11:15 pm
Location: Washington DC Metro Area
Contact:

Re: Steps to improve a clip that is only vocals?

Post by DewDude420 »

I downloaded your audio sample and threw it though my usual software to see if it could be improved and what steps you would need.

Your audio benefits from some dynamic compression; however there are additional recording flaws that prevent it from being entirely effective.

For starters, there is a 60hz hum in your file...probably from the fact you used a headset. This means that anything below 60hz in your file is going to be gone. This also means you'll have harmonics of that 60hz that come in to play.

Your hissing S can be somewhat mitigated by de-essing; but not entirely. They work best with a whistling type of noise since essentially what you are doing is picking the frequency band that contains your essing and applying additional dynamic range. But your bigger issue is "popping". You can't really get rid of the popping digitally. You have to eliminate that before recording.

When I record with a headset; I never put the microphone in front of my mouth; I always put it off to the side...that way the air leaving my mouth doesn't strike the element. Most professional setups have pop-guards...they're additional things in front of the microphone that prevent those impulses.

You also have some asymmetrical pattern to your recording; which may indicate an additional DC offset.

My copy of your voice is at http://dewdude.ath.cx/voicetest.wav - you can improve it...but some of the flaws of the recording become more apparent after processing. I should add this is much higher compression than I'd typically run; I was just doing a "quickie" to see what steps would need to be taken. I think if you start by improving the quality of your recording; you'll be ok.
treefitty
Posts: 2
Joined: Sun Nov 13, 2016 5:02 pm

Re: Steps to improve a clip that is only vocals?

Post by treefitty »

Thank you both for the help. So from what I understand, a better mic will make it sound a lot better overall, but also I should be careful of popping. I messed around and found that putting the mic down helped a lot, but the exact position is pretty difficult, a couple millimeters too far down and the voice starts becoming too quiet/distorted/weird. I guess as long as I have my headset mic, that balance will be hard to get exactly right, but I gather I can improve it somewhat with position.

Additionally you both mentioned compression will help as well, is there an exact science to compression? I tried: ReducePeaks, and then MaximizeVolume, but, the overall waveform didn't change too much, and I think that's because in the overall file, the volume differed too much (probably because of the peaks made by popping). So, I had to do ReducePeaks and MaximizeVolume multiple times on my recording (I actually re-recorded it per your advice to minimize popping, but it still happened very occasionally). So I did something like:

ReducePeaks,
Increased volume with EQ,
ReducePeaks,
ReducePeaks,
MaximizeVolume(default),
ReducePeaks,
MazimizeVolume(90%)

And after all that the waveform started evening out a little more, but, because of how many times I had to apply those filters, I felt like what I was doing was incorrect or "newb-ish", is that so? Is there a more streamlined way to apply those things or does it usually always take multiple passes before it starts sounding alright?
DewDude420
Posts: 1171
Joined: Fri Mar 11, 2005 11:15 pm
Location: Washington DC Metro Area
Contact:

Re: Steps to improve a clip that is only vocals?

Post by DewDude420 »

I honestly don't have much experience with Goldwave's dynamic compression.

I did my processing in Adobe Audition with a combination of tube-modeled compression, multi-band, and hard limiting. De-essing was done in multi-band.

These are much more aggressive than Goldwave's; so depending on various settings you may need to apply them several times.

Nothing wrong with doing something a bunch of times if you back off the processing level. I often do noise reduction in 10 or more passes....but at very gentle amounts.
DewDude420
Posts: 1171
Joined: Fri Mar 11, 2005 11:15 pm
Location: Washington DC Metro Area
Contact:

Re: Steps to improve a clip that is only vocals?

Post by DewDude420 »

Also...try extending the mic so it is beside your mouth as opposed to below. When I do things with my headset the mic is just to the side of my mouth.
Post Reply