Greetings, all! Mixing/Sound Quality question...

Greetings, all! I hope you’re all well!

I’ve got a couple of sound quality issues I’m trying to work through. Here’s my basic setup:

Running Win10 and Audacity 3.0.3.

I also have a Focusrite Scarlett 2i2 that I’m using. So obviously I’m recording this onto my laptop.

Microphones: AKG P170 for guitar, Oktava MK319 or Shure SM58 for vocals.

I don’t know that it matters, but the guitar I’m using is a Takamine EG334 SBC in standard tuning.

My problems are as follows:

  1. I’m having to record this in my garage. There’s not really any place else in my house to record it, and since it’s pretty much a “personal project” at this point, I really don’t want to shell out hundreds (or thousands) for studio time. My garage is already rather full, so I can’t really “sound treat” anything. I’m having to “make due” with what I have. Which leads us to…

  2. My absolute biggest problem is background “noise”. I remedied most of it simply by upgrading my microphones. Are there any other effects (other than those I note below), and in what order can I apply them, to reduce noise as much as possible? As I said, this is purely a “personal” project for now, but I haven’t decided if I want to sell copies or not, but that’s for a different post. But I still want it to sound good.

I’ve tried using the Loudness Normalization, Noise Reduction, and Normalization features, but as I said, I’m very, very new to recording/tweaking/mixing. I know the answer to “how much background noise is acceptable” is probably “none”. So if there’s a way to get rid of the background noise completely, I’m all ears. I kinda-sorta can get rid of some of it (though I haven’t written down how), but then when I raise the gain on that track pretty high, I still hear noise, although it’s in the -48db range. If I leave the gain at 0, it’s like the background noise (which is usually that ambient “hum” or some call it “rumble”). I tried turning on the -20db setting on my P170, and unless I’m doing something wrong, or don’t understand what that switch does, I don’t like how it sounds. It doesn’t sound quite as clear as it does when it’s set to 0.

  1. I’m still having trouble with the guitar sounding as “clean” as I’d like. I might have found somewhat of a “sweet spot”, but I would really like some tips on “mixing” or EQing a guitar. It’s still a little “boomy”, and I’d like it to sound more “natural”. Any tips here would be great.

  2. Is there a way on the vocal track I can “edit out” the “inhale” sound that happens right before I start singing a line or verse? I have a pop filter in place, but my MK319 is picking up the inhalation breaths.

I’d like my recordings to sound “raw” without sounding “raw”, if that makes sense.

All of these songs are probably just going to be me and my guitar for right now. No other instruments. Ultimately I want to get a vinyl pressing made for myself (hang on the wall), and maybe some CDs made for family and friends, before I decide whether or not to sell them.

Thank you for all your help! Have a great day and make terrific recordings!

My garage is already rather full, so I can’t really “sound treat” anything.

That’s not bad news. The worst thing you can have is clean, shiny, bare walls facing each other. Does it have a peaked roof?

Are we recording neighborhood noises, or is most of it local to your equipment and setup? Can you tell?

What is the door made out of? If you’re listening to the neighborhood, it might be coming in through the metal door. They’re not all that resistant to noises. In that case, you can hang a couple of furniture moving pads over the inside of the door. This can be a little bit of an exercise if the car door is the only way in or out.

Post 10 seconds of the work. Include a little actual music and then mostly silence so we can hear the “room.”

Export it as WAV (Microsoft) 16-bit. Scroll down from a forum text window > Attachments > Add Files.

Koz

No, the ceiling of the garage is flat.

Though I am pointing the mic toward my garage door, so that might be some of it. The garage door is metal, or some kind of composite that appears like metal.

I think some of it is my laptop fan, but there’s nothing I can really do about that. I’m not really getting people talking or whatnot, but if a car drove by I would pick it up. Most of the noise I’m concerned with, though, is just that natural, “ambient” noise. Like if you just went into your garage (say, at night when people aren’t likely to be out or driving through your neighborhood), and you plugged in your microphone and hit record for 30 seconds or a minute.

I’ll have to create a short sample and post it. So you can hear what I’m talking about. I’ll do that in the next couple of days.

Acoustic guitar tough, like a lot of acoustic instruments. It’s highly dynamic (a large quiet-to-loud range) so even when you maximize the volume it’s not that loud and the quiet parts are quiet-enough that noise is a bigger problem.

  1. My absolute biggest problem is background “noise”. I remedied most of it simply by upgrading my microphones

Microphones are linear so a more sensitive microphone picks-up more signal and more acoustic noise. It does help that all of your microphones are directional (cardioid) because noise comes from all-directions.

However, the higher-output from a more sensitive mic (like your condenser mics) will help to overcome any electrical noise from the preamp in your interface (a better electrical signal-to-noise ratio).

The same goes for your recording volume/gain level control. Turning-down the gain reduces the signal and noise together with very little effect on the sound quality.

usually that ambient “hum” or some call it “rumble”).

Effect → Filter curve → Manage > Factory Presets → Low roll-off for speech will kill the low frequency noise with little or no effect on voice or acoustic guitar.

I tried turning on the -20db setting on my P170

Don’t use that. It’s to prevent overloading of the mic’s internal electronics with loud sounds like a kick drum or loud guitar amp. The SM58 is a dynamic mic with no internal electronics and it’s virtually impossible to overload. (Dynamic mics are also a lot less sensitive.)

  1. I’m still having trouble with the guitar sounding as “clean” as I’d like. I might have found somewhat of a “sweet spot”

If you have a good reference recording you can try to match the EQ by trial-and-error. Also try experimenting with mic position and maybe research how to record acoustic guitar (some people use two mics). Then once you get the best sound, you’ll probably want to add some tasteful reverb. Dynamic compression (the compressor or limiter effect) can make a “louder”, “more intense”, sound but compression makes the signal-to-noise ratio worse and of course it reduces the dynamics.

  1. Is there a way on the vocal track I can “edit out” the “inhale” sound that happens right before I start singing a line or verse? I have a pop filter in place

The pop filter is for when air comes out of your mouth, especially with “P” sounds (“plosives”). Breathing is normal so although you might want to reduce it you may not want to eliminate it altogether. Try to back-up and/or turn your head to when you take a breath and beyond that it’s probably best to reduce it manually with the Envelope Tool.

When you mix “watch your levels”. Mixing is done by summation so you can easily get into clipping (distortion) even though the unmixed tracks don’t clip.

Ultimately I want to get a vinyl pressing made for myself (hang on the wall)

What does that cost? If you can afford that you can probably afford a couple of hours in a semi-professional studio. Of course the trick is to get the song(s) down so you can get it done in a couple of hours.

No, the ceiling of the garage is flat.

And the floor is cement, right?

I think some of it is my laptop fan, but there’s nothing I can really do about that.

There is, but you’re rapidly talking yourself out of recording in the garage. You can’t write a check for better kit and more processing and get rid of a bad room.

Do you have a car? Put the car in the garage, get in the car, and record there. More than one sound track was produced in the back of a car if nothing else worked right.

Koz

I’ll do that in the next couple of days.

Remember when you do that, include some natural voice or music. Do a short intro: “Here’s my garage background noises.” We can’t use 10 solid seconds of laptop fan noise.

Koz

Wouldn’t this be what the normalization functions are for? It appears that below you suggest not using compression, but are there other ways to kind of “even out” things?

Microphones are linear so a more sensitive microphone picks-up more signal > and > more acoustic noise. It does help that all of your microphones are directional (cardioid) because noise comes from all-directions.

However, the higher-output from a more sensitive mic (like your condenser mics) will help to overcome any electrical noise from the preamp in your interface (a better electrical signal-to-noise ratio).

The same goes for your recording volume/gain level control. Turning-down the gain reduces the signal and noise together with very little effect on the sound quality.

The “happy spot”, I think, for gain for my P170 on my Scarlett 2i2 is somewhere between 1/4 and 1/2 of full gain. I think it’s somewhere similar for my MK319.

Effect → Filter curve → Manage > Factory Presets → Low roll-off for speech > will kill the low frequency noise with little or no effect on voice or acoustic guitar.

At what point should I apply that effect? Alone, or in conjunction with some of the others I mentioned (Loudness Normalization, Noise Reduction, and Normalization), and if so, in what order?

The SM58 is a dynamic mic with no internal electronics and it’s virtually impossible to overload. (Dynamic mics are also a lot less sensitive.)

I kinda noticed that. I also noticed that, as you noted, the signal to noise ratio didn’t seem as good. The SM58 seemed to pick up more “ambient” noise than I had expected during vocal takes.

If you have a good reference recording you can try to match the EQ by trial-and-error. Also try experimenting with mic position and maybe research how to record acoustic guitar (some people use two mics). Then once you get the best sound, you’ll probably want to add some tasteful reverb. Dynamic compression (the compressor or limiter effect) can make a “louder”, “more intense”, sound but compression makes the signal-to-noise ratio worse and of course it reduces the dynamics.

I have some ideas, I can think of two off the top of my head. They’re from vastly different eras, and different musically. They’ll either really confuse you, or give you a better idea of what I’m looking for. If you can find it (and that’s a big if), try to find the Garth Brooks/“Chris Gaines” version of “It Don’t Matter To The Sun” and listen to the acoustic in that. Also listen to “Broken Halos” by Chris Stapleton. Also find “With Tears In My Eyes” by Hank Williams, and “Deep In The Heart Of Me”, the acoustic version, by Junior Brown. I think that’s kind of the direction I’d like to head. Simple, clean guitar.

The pop filter is for when air comes > out > of your mouth, especially with “P” sounds (“plosives”). Breathing is normal so although you might want to reduce it you may not want to eliminate it altogether. Try to back-up and/or turn your head to when you take a breath and beyond that it’s probably best to reduce it manually with the Envelope Tool.

I’ll check that out.

When you mix “watch your levels”. Mixing is done by summation so you can easily get into clipping (distortion) even though the unmixed tracks don’t clip.

The extent of my “mixing” in Audacity is recording the tracks, messing around with the normalization, and adjusting the gain until things sound close to what I want.

What does that cost? If you can afford that you can probably afford a couple of hours in a semi-professional studio. Of course the trick is to get the song(s) down so you can get it done in a couple of hours.

I’m not sure I’d be able to be done in a couple of hours. As far as the vinyl goes, I did find a place that would press 1. It’s $55 for 140g black vinyl, and $60 for 180g black vinyl. A paper full color jacket is free, but a premium one is $12 more.

And the floor is cement, right?

Yes, it is.

Do you have a car?

Yes, but it won’t fit in the garage. We’ve got too much stuff in there.

Wouldn’t this be what the normalization functions are for? It appears that below you suggest not using compression, but are there other ways to kind of “even out” things?

No. Normalization makes ONE adjustment to the whole file (or the whole selection), like adjusting the volume before the song starts. Regular normalization is based on the peaks which don’t correlate well with loudness so if you normalize all of your music, some songs will still be louder than others. There’s also “Loudness Normalization” which does target a loudness level, but it’s also ONE linear adjustment to the whole file, and with loudness normalization you have to watch out for clipping.

At what point should I apply that effect? Alone, or in conjunction with some of the others

It’s not too critical but it could change your levels so, before normalization. Usually it’s a good idea to normalize as the last step for the simple reason that other effects can affect the levels.

At what point should I apply that effect? Alone, or in conjunction with some of the others

Mixing is when you blend (mix) your voice and guitar. If you are recording guitar on the left and your voice on the right, you aren’t mixing (the sounds are mixed acoustically in the air) but you probably wouldn’t want to leave it that way.

I did find a place that would press 1. It’s $55 for 140g black vinyl,

Wow, that’s cheap! I would have guessed a LOT more! In the “vinyl days” it was a complicated expensive process and it wouldn’t have been economical to press one record.

Maybe get away with the car being outdoors … https://youtu.be/kO8Fl1IQVNA

Interesting. I thought you could apply “Loudness Normalization” on a “per track” basis. I didn’t realize it was a whole file thing.

It’s not too critical but it could change your levels so, before normalization. Usually it’s a good idea to normalize as the last step for the simple reason that other effects can affect the levels.

Ok. That makes sense. So if I apply the “low roll off”, when would I apply the “noise reduction”? After that I would presume.

Mixing is when you blend (mix) your voice and guitar. If you are recording guitar on the left and your voice on the right, you aren’t mixing (the sounds are mixed acoustically in the air) but you probably wouldn’t want to leave it that way.

Right now I’m not “panning” anything. I’m recording vocals and guitar to mono tracks. And I’m completely clueless on the rest. Even the information you’ve given me is brand new information. and I want you to know I very much appreciate it!

Wow, that’s cheap! I would have guessed a LOT more! In the “vinyl days” it was a complicated expensive process and it wouldn’t have been economical to press one record.

Well, I was a little wrong. it’s not a “pressing”. It’s done via lathe, apparently. I’m not familiar with that process, though. I’m not really planning on playing it, it’s going to mainly be for display purposes.

Maybe get away with the car being outdoors … > https://youtu.be/kO8Fl1IQVNA

Booth Junky is good. He has simple explainer videos on all the recording stuff.

Also. https://www.youtube.com/channel/UCHHf1h8k7MA6-AG8FXjnQSw

Writing that down.

The only thing I don’t think he covers is how to sound like you have three testicles.

There is one forum poster trying to do that with Effect > Change Pitch. It’s not going well and there’s a good technical reason for it. Mike’s natural voice has good velvety, low-pitch tones, and crisp, natural mouth sounds. Change Pitch moves the voice tones and everything else. You can’t go very far before it sounds like somebody threw a blanket over your head.

Koz

A car is actually not a bad place to record, especially when parked inside a garage or a quiet suburban road.

I was in the mood to get some coffee from the “mermaid” :wink: so used the opportunity to make a test recording for those that may still doubt the effectiveness of a car’s interior.

Attached is the clip, it is completely raw, I have not modified it in any way.
It’s in 44.1KHz as that is the highest my phone will record at, when using wav.
Microphone was the phone’s internal one.

At around the 7 sec mark, you can hear me moving around and at 11 secs, opened the door.
The noises you can clearly hear with the door open, are a car wash about 30 meters away and a small diesel truck at around 20 meters away.

Depending on frequency, there is a difference of between 16 and 20dB, with the door opened and closed.
I excluded the noises made by me moving around and the actual noise of the door opening.
That ain’t bad.

When you submit a sound test, it’s good to include a reference. The important measurement is not noise. It’s noise as compared to the voice or music.

“Car sound test, May fifteenth, twenty-twenty one.” [background sound]

Koz

The idea behind this recording was to show the attenuation from outside noises, not noise floor relative to wanted sounds or tonal characteristics of the car’s interior, although that will play a part in the response curves shown below.

By noise, do you mean the section when the door was closed?
If so, that is the attenuated sound of the car wash and truck.

My car was not idling, radio was off, as well as the air-con and was parked in an outside lot.

The two responses below further illustrate the attenuation as regards frequency, higher attenuation at higher frequencies which makes perfect sense.
The lower the frequency, the more it’s going to penetrate into the car’s interior as the doors, padding, etc have little mass.
The peaks at around 65 Hz are not from mains as we use 50 Hz.

Given the built-in low and high end roll off of the phone and the mic, which are constant, it’s quite clear from the responses below that they are the same noises.


The idea behind this recording was to show the attenuation from outside noises, not noise floor relative to wanted sounds or tonal characteristics of the car’s interior, although that will play a part in the response curves shown below.

True. But you can change outside noises by moving the car. If you had a poor voice test, then no moving the car is going to help and the recording technique may not be valid. Good to know that right up front. My pickup truck may not be a good studio candidate. Too much glass relative to the padded surfaces, but it would probably give a really good differential door position test.

I don’t need any excuse to visit a park for lunch. I’ve never measured the noise up there.

Screen Shot 2021-05-16 at 4.42.44 PM.png
Koz

But you can change outside noises by moving the car. If you had a poor voice test, then no moving the car is going to help and the recording technique may not be valid. Good to know that right up front.

For sure, I’m not saying that a car is a fantastic space to record but, given the OP’s lack of a quiet space, it may be preferable.
Of course positioning will make a difference, and one can plan for that, however, it’s the random noises that are the unknown.

The test I did will give the OP and others, a good idea of the kind of attenuation they can expect.

Too much glass relative to the padded surfaces…

Glass will certainly add to the reverb time and color the sound, it will be dependent on surface area and angle of the glass, ratio of glass to padded surfaces, interior volume of the car, etc
It’s easy to test though, no fancy RT60 tests required, just get in and belt out a couple of lines or riffs and see how it sounds.
It’s not all bad though, a bit of reverb and color is often nice on vocals and guitar.

I don’t need any excuse to visit a park for lunch.

Looks a bit hazy today, best stay indoors. :wink:

I’ve never measured the noise up there.

Where, Mount Lee?
My guess is that noise will be the least of your worries.
It’s all the hot air from the pompous celebs you have to be careful of. :smiley:

It does occur to me the worst problem might actually be wind. There’s a neighborhood street that might have good prospects. It’s off a busy avenue, but it’s a dead-end into a park. No traffic.

It’s all the hot air from the pompous celebs you have to be careful of.

Not to worry. They’re never found outside of their expensive cars. I wonder what the noise differential is inside and outside those cars.

Koz

It does occur to me the worst problem might actually be wind.

A good point, wind is pretty broadband…assuming of course you are referring to wind caused by uneven heating as opposed to… :smiley:
The windy season is coming up for me, will certainly do a test.

I wonder what the noise differential is inside and outside those cars.

My guess, considerably better, probably 8 - 10 dB more.
After all, I’m sure those very important individuals don’t want to be subjected to the noisy riff-raff outside.
Seriously though, it would very much depend on the materials used for the padding, and some of those cars would probably be bullet proof, which will add mass.

A quick search revealed some of the popular materials used:

Lightweight Composite Armor.
High-Hardened Ballistic Steel.
Ballistic Nylon.
Kevlar and Aramid Fibers.
Lexan.
Ballistic Fiberglass.

I guess a good start would be to try and find the absorption coefficient of those materials over different frequencies.
Interesting question Koz, when I get some time, will try and research it and come up with an estimate.

Of course the problem might be self-limiting because the owners would have to abandon their home sound studio to use it.

Koz