Audio Advice

[topaz hazel]

A place to ask audio questions and get help from other creators

[woeful mango]

im not in love in how the voicoever sounds here- any tips as to how to clean it up? rerecording is not feasible at this point 🙁

https://youtu.be/Opaee6duWIw

[radiant void]

Thanks again for opening this up @topaz hazel

Specs:

• Shure Sm7dB (1" from mouth off axis)
• MVX2U

Software:

• Davinci Resolve Studio
• Izotope 11
• Shure MOTIV Mix (MVX2U control)

Current Setup:

• SM7dB: BYPASS OFF, +18 dB, phantom 48V ON
• Track/timeline: VO track is MONO, project/bus is STEREO
• Recording levels: loud phrases peak around -6 dBFS (avg roughly -18 to -12 dBFS)
• Final: normalized/limited for YouTube

Problem:
My recording is crisp and clear, but it doesn’t feel as forward / present / “up-front” as my reference channel (Voices of the Past). It comes across slightly behind glass / backgrounded, even when loudness is correct.

Ask:
What’s the most likely cause of that missing “forward presence” in this chain (mic technique vs noise reduction artifacts vs compression/EQ choices), and what settings/workflow would you recommend to get closer to that intimate, mid-forward narration style that Voices of the Past has perfected?

[radiant void]

My 2c : Audio sounds crisp and clear, the lowest hanging fruit would be to trim out the breaths and mouth clicks (where possible).

[topaz hazel]

Hard to say without knowing what you're aiming for. What don't you like about it specifically? Tonality, breathing, clicks/pops, something else?

[topaz hazel]

Do you have a voice sample?

[woeful mango]

I just think it sounds too “boomy”. I want it to sound like a podcast. But if you don’t hear anything that’s glaringly wrong then it’s probably fine

This video is heavily dependent on voiceover quality so I want to make sure it’s perfect

[topaz hazel]

I think it sounds nice and full, personally! Definitely has that podcast feel, imo. Biggest thing I'm noticing besides slight occasional mouth clicks which takes surgical time to remove is an ever so slight aggressive taming of your sibilances. Just on the verge of giving yourself a lisp but it's definitely not a video killer. Definitely way ahead of most channels, that's for sure.

[woeful mango]

Happy new year everyone! I just want a second set of ears here

Does the voiceover sound bad at all?

No BS. Here's how to make your first $1M investing.
https://youtu.be/O0p1Kaeya_Q

[topaz hazel]

Nope! Sounds great to me 😄

[woeful mango]

Thank you so much man!

[wise gulch]

So in the last few recent weeks I have been dabbling with hardware for audio processing in which I bought a dbx286 as I felt it was the easiest on my wallet and I like the results I saw around the internet. Now my issue is I wasn't sure if I was going to be able to use this for both doing voiceovers and trying my hand at voice acting or audiobooks stuff like that and also use it for gaming content so I have one set up and not multiple. So what I found out recently was the fender quantum lt2 interface that I got is underpowered for what the dbx puts out. So I am going to exchange that for a focusrite solo Gen 4 most likely since that is fairly equivalent to the fender model. The next issue that I'm trying to resolve is closing the gap between using this as a semi-professional home studio setup and also gaming content like I do in Indie horror games and to not clip when I scream using the setup. But here is an example of where I'm at right now and this is just a yap that I did to myself as a sample recording to hear the tone and what not

I will say two. The only post-processing this audio has is just a single band compressor to help smooth everything out and then a little bit of makeup gain because of the compressor used. Otherwise everything is the mic and the processing unit itself

[keen sinew]

So I've been working on my new documentary and I've used the same settings as before so I wnated to ensure that this sounds good to you guys. I'm using Adobe Audition with the effects of noise reduction at 85% with 50 dB, vocal enhancer (low tone), mastering (make room for vocals) and a delete silence process that shortens my silence by 500 ms.

@wise gulch @topaz hazel

[topaz hazel]

Killin it!

[topaz hazel]

Not bad! Only thing that popped out to me is a background hiss thats coming in and out.

[wise gulch]

definitely a huge improvement from where you started! Maybe a apply a slight DeReverb too.. like 5% worth? up to 15% max. I hear a bit of room reverb (i have the same issue with my audio at times. most the time you cant hear it with other noises going on (game, music, ambinet sounds, etc). but my raw audio in solo listening - you can hear it.

[hasty ferry]

I'm in need of some audio advice, if anyones willing.

What other information may be needed besides this information? -

Microphone: Samsung Q2U

OBS settings: Gain = 10db; Limiter = Threshold: -2db and release of 60ms

Editing Software: Kdenlive
I use a homemade box with towel dampeners to ensure recording is done well

Type of content is Gaming/Narration (non-streaming), so most of effects are added in the editing software

I want to get the audio to the crisp professional level sounds (like you'd see with highly professional set-ups

Attached is also a recording of the current output from OBS*

By chance with this information, is there additional effects I'd need to add that I can set as a preset for future audio narration for editing?

[wise gulch]

What post-processing effects did you do to your vocal audio for this sample, if any?? I'm listening in my car stereo and it almost has this like digital swirl effect. Typically I feel that happens when too much of an effect is applied

[hasty ferry]

i didnt use any just the standard OBS settings I have of Gain = 10db; Limiter = Threshold: -2db and release of 60ms

I use the gain because it always comes out to quiet when I don't

Samsung Q2U's are known for being too quiet, I learned the hardway

[wise gulch]

I'll have to assume that you are connected via USB and your PC mic volume is set appropriately as well

[hasty ferry]

It should be, I ensure each time that its set to input/100% volume, etc.

[wise gulch]

Sometimes depending on the system and mic and all that 100% volume on the PC is not always the best.

[hasty ferry]

Eek, it automatically set it lower but the voice comes out way to small when its like that. And, the gain needing to be applied degrades the quality, im very tone/audio deaf but I can still hear it pretty bad when its like that :/

[wise gulch]

I'm pulling up to my house so I can get on my PC and kind of go through what I know as far as the basics. Sean has a lot more in-depth audio knowledge so if he sees this he can by all means correct? Anything I say cuz I am not an audio engineer. I just know what has worked for me and what I have seen. A lot of other creators do and I have done a lot of research, but that doesn't mean I've comprehended that research 100% well enough. So stand by me

[hasty ferry]

I really appreciate it! I'm happy to put in the work necessary and have struggled with audio since my first video, so I am grateful for any information that can assist!

[wise gulch]

i'll be back... i have to run out for a bit. forgot to do something before coming home. ugh!!!

[hasty ferry]

Haha no worries! I know how those things go 😄

[wise gulch]

okay, i'm back

[hasty ferry]

Welcome back!

[wise gulch]

Now I'm having audio problems with my new focusrite interface. Ugh!!

[wise gulch]

hey bud, sorry about all that... I sorta got my stuff figured out - but still unsure if a cable issue, or something else. incoming rant stay tuned...

[hasty ferry]

Hopefully you can get it figured out! Audio is a wild subject on content creation because it seems like it should be so simple, but its like the most complicated 😭

[wise gulch]

well, in this little rant was to test the units for the little sudden frequency scratchy noises that happen, and audio dips out til i fiddle with things

[wise gulch]

you are absolutley sure no effects processing is on before or after for this sample? no, graphics card noise cancel stuff, or anything like that?

[hasty ferry]

The only thing truly is the gain and limiter in obs. Unless the Samsung q2u has stuff automatically applied I'm not aware of 😬

[wise gulch]

only other thing I can think of is it's just Discord being horrible with the file shared.

what are your other OBS settings for audio?

[hasty ferry]

recording audio bitrate settings are at 320

[hasty ferry]

Are these what you mean?

[wise gulch]

yes.. sorry for all the questions and such. I just like starting from the top. (as best as possible)

What about export settings from your editing program

[wise gulch]

have you tried applying the Noise Suppression filter in OBS to your effects chain. As long as you're not screaming loudly in your recordings, this could be beneficial for your Voice Overs with the set up you have.

I can't have my Noise suppression on because when i scream from playing horror games - it cuts my mic out.

[hasty ferry]

No worries! Happy to answer, I think I did send the export settings? I didn't use kdenlive when sending over the sample audio

[hasty ferry]

Where would this be in the flow? I have had in on OBS before but wasn't confident in it.

Right now the stack is 1st gain, 2nd limiter*

[wise gulch]

ok okay.. the the audio sample up above was straight from OBS, and NOT from the Kdenlive ?

[wise gulch]

personally I'd go
Gain
Noise Suppression
Limiter

then do a test read of something.

[hasty ferry]

Yep!

[hasty ferry]

Awesome! I'll do another test recording when I get home from work today and send it here (with the new stack)

[wise gulch]

Awesome! cant wait to hear it

[hasty ferry]

I appreciate the help a ton!

[hasty ferry]

Awesome, I'm back - here's the new clip. Don't mind my nerdy banter at the end XD

[wise gulch]

the bane of my existence ...nice

i'm assuming this is with noise suppression on, yes?

[hasty ferry]

Yep

[wise gulch]

sounds WAY better in my opinion.

i'm assuming too, the area you record in is not sound treated at all, or very little.

[hasty ferry]

Very little, I don't have the biggest budget, but did my best with a box and towel the mic sits in. The room has some sharp corners

[wise gulch]

i'm just wondering if the box you have is causing some "boxiness" in your voice. BUT - I feel with some EQ you could reshape the audio a bit. IDK Kdenlive software, and what effects tools they offer

One common suggestion I see out there is record your VO in a closet, that has clothes hanging, and that will help the reverberance situation too

[hasty ferry]

They have a lot and can accept custom ones. I have Audio Equalizer (avfilter). Has central frequency, width type, band width, gain, and Mix settings

I'm in a unique place because the closets are small and wouldn't work for a recording 😭, the advice I got is the box helped the audio quality a TON

From where it was previous*

[wise gulch]

oh yes, I didnt mean it was causing anything necessarily bad, I can tell it cuts down on room noise for sure. you just have to correct the boxy sound is all. a little EQ and you are probably good in my opinion

[hasty ferry]

Oh that's good to hear haha, how does EQ work? I'm not familiar and don't trust my ears to do very well 😅🤦‍♂️

[wise gulch]

this is a good guide for what the area of frequencies and such are targeting.

let me download kdenlive and see if I can help get you close to the EQ shape.... Sean is actually way better at the EQ than me. Well, he's the goat with all audio really.

[hasty ferry]

So with the box region would that be where I'm aiming for?

I have got a good process of volume down where I always aim for my audio to be around the 10 mark here (where mouse is), but am unfamiliar with equalizing

[hasty ferry]

Thank you! Kdenlive is open-source. I appreciate you being willing to even download it!

Sean and you are GOATs in my opinion haha

[wise gulch]

So yeah, if you're gain or levels are correct. I would aim for and average of -12 with the occasial spikes to -6. I personally do an avg of -18dB with spikes to -12. and correct with compression in post. ....etc etc.

the EQ will come in kdenlive software. ...not in your levels as you just showed above

......

you could do the EQ with OBS filters too... but I would refrain from that. as it will be baked into the raw audio, and it's best to EQ after, IMO

[wise gulch]

Okay... not gonna lie...........the effects area in kdenlive is horrible. but im trying here. lol

[hasty ferry]

Haha I'm glad to get some validation here. I always wondered why the audio effects were different then what I was looking for

[wise gulch]

yeah, this EQ is making me question my sanity

[wise gulch]

i'll try to mess with the audio a bit more tomorrow. I'm not the best with EQ in post tbh

[hasty ferry]

eek yah, it does have the option to download other effects into the system, maybe thats the route i should go

[hasty ferry]

Would EQ be the last effect needed or is my audio needing a lot more work haha?

[topaz hazel]

Sorry friends, been swamped with work lately 😅 Severlight's a beast

[wise gulch]

lol, I'm trying man... i'm trying. ....glad to see you poppin up in the chats again! missed ya bud

[topaz hazel]

d'aw missed all you fellow youtube nerds as well

[hasty ferry]

Whenever you get the chance (of course once work settles) would you also be able to check out the test audio?

Also Severlight is the GOAT, I do like how it sounds with the noise suppression the EQ is stumping me but I think that's common haha

[topaz hazel]

Sure thing! Probably wont be able to until Sunday. Working audio for a One Republic concert this weekend

[wise gulch]

I feel there's two common ways to go about this... either EQ before (pre-record)... or EQ after (in post)

So you could try for Sh#ts n grins to EQ from OBS filters - which will be baked into the record of course. This is what I used to do for my game play thrus. As it was easier for me to set and forget really. I did COUNTLESS test recordings of the same paragraph so I could compare each one to my liking.

The 3 different EQ's I tried were:

OBS 3-Band EQ

Reaper plug-ins

and

TDR Nova

I went with TDR Nova Plugin for OBS as it was way better than reaper (over rated, IMO)

https://www.tokyodawn.net/tdr-nova/

with TDR Nova I was able to either EQ in PRE or in POST (as you probably can with Reaper as well. Assuming Davinci, Premiere, and other editing software have VST plug-in capabilities

[hasty ferry]

I tried looking into this, I'm a little confused 😕. How do you know if the EQ is set correctly? I feel like I'm borderline audio deaf

[wise gulch]

well, how are you going about doing the EQ part?

Pre - or - Post?

[fallen axle]

@wise gulch @topaz hazel what can be improved?

[wise gulch]

Sounds good to me, maybe just apply a little bit of dereverb filter. Like 5-10% worth. Too much will degrade the audio

[fallen axle]

besides the reverb can anything else be improved or is it perfect overall? were u able to listen to the full thing?

[wise gulch]

I do hear a hum as you are speaking, a noise suppression filter should clean that up, and sometimes the volume spikes up a bit.

basically if you can't do much about the hum from the source, i.e. a fan, a PC running, an AC unit in the windows, grandma's fridge that has been running since the great depression but refuses to get a new one because they don't make 'em like that anymore and doesn't want any service guy coming to fix what she has because she feels like they are eyeing her stuff and looking to steal her candy dish.... if that's the case, then before you start speaking let your recording go for a few seconds to record the floor noise/room noise... then start your speaking/Voice over. then in Post you can use a noise reduction or noise suppression filter to remove all that noise that comes thru when you speak.

[fallen axle]

hum as in the background noise? which timestamp are you talking about that whitenoise whirring? correct?

[wise gulch]

at 3:36, the volume jumps

for a short bit

and yes, there's a background noise hum as well during the parts you are speaking

[fallen axle]

ou can use a noise reduction or noise suppression filter to remove all that noise that comes thru when you speak. do i do this for each audio file i import or do i it for one and it works for all of it?

[wise gulch]

yes... do you have audacity?

just use the filter in Post (after you record) ...

or

use a Noise supression filter like in OBS before you record

[hasty ferry]

I would like to preferably do post production (however the kdenlive options isnt the greatest as you saw), but I see the option in pre with effects for obs 😅

[wise gulch]

I would try out the OBS filter, the 3-Band first. Could be all you need before going more advanced.

you can monitor your voice in real time as well in OBS. in your microphone audio properties

click here first to get to the Properties shown above (whatever you call your Mic Input though)

[hasty ferry]

Just the flat 3-band? There's no setting adjustments?

[wise gulch]

3 band is, bass(low), mid, and treble (high)

[hasty ferry]

But I guess im confused because i thought you set the decibal levels with EQ filters? Does 3-band auto-detect/set those?

Again, I dont trust my ears lol

[wise gulch]

@hasty ferry best way I could explain it.

[hasty ferry]

This is really well done, I'll see if I can get something decent moving forward 😁. Would you still be willing to listen to the finished product as verification I got it right?

[wise gulch]

no need to ask, send your sample recordings anytime man

[topaz hazel]

Apart from what Sever has already said, there are a few times when your peaks are getting crunchy. Keep an eye on your levels when you lean into words/speak louder and make sure you're not clipping

Yeah the two biggest things with this one are room echo and background hum. Resolve can solve these pretty easily, especially that background hum. It's nice and consistent so should be pretty easy to mitigate with a basic denoiser. Just be careful not to crank denoising too far or you'll get artifacts.

[hasty ferry]

Peaks getting crunchy, but that should be resolved with some of my EQ additions soon right with the 3 band?

[topaz hazel]

Nah separate issues. Clipping and tonality are different.

The sweet spot is green meets yellow on most digital meters

[wise gulch]

https://tenor.com/view/thank-goodness-youre-here-ella-payne-house-of-payne-mayjor-payne-s9e2-gif-25145113

[wise gulch]

@hasty ferry if you look at my meter as im speaking, im hovering arounfd the -18 to -16 on average with occasional spikes to around -6. Not my ideal as i dont remember if i did any post processing to this clip. but when i speak at a normal level i target -18 (persoanlly) BEFORE any processing is applied.

wait... technically this has processing... it's running thru the DBX unit first

[topaz hazel]

Uuu did you get the DBX channel strip?

[wise gulch]

yeah, i have the DBX286s running to the Focusrite solo gen 4.

[topaz hazel]

I love that thing! Great sound for on the cheap

[wise gulch]

yeah man, for how cheap it is, it produces such a good signal and tone.

for the DBX, only think i dont like is the Compressor design. I thought about getting a separate compressor as well. but this will do fine as is.

I basically do way less post processing now with this set up... only draw back is it makes a super sensitive mic so - only post process i do is a Reaper Gate (for my mouth smacks, pops, cracks, snorts, and farts at threshold of -40 to -38), and single band compressor to level everthing out

[topaz hazel]

Beautiful setup! I wonder if anyone has made a de-fart module before? 😂 iZotope de-fart

[wise gulch]

omg, that would be hilarious. Sounds like a SNL type skit

[hasty ferry]

Is there anyway to actively prevent this or is it more of you have to keep an eye personally on it?

[wise gulch]

there are countless ways of setting up your mic for proper levels to prevent clipping/distorting, and preventing getting into the red as Sean mentioned in the chart...

i would turn off any processing you have set up, and test out your levels with your normal talking. be sure to have proper mic placement as well - as this will be crucial down the road. if you set up your mic initially - let's say at 3 inches away from your mouth - then you record with the mic at 8 inches away it aint gonna sound the same.

so keep a consistent mic placement everytime after you set up your levels

[wise gulch]

Whelp, I finally bought one, a dream mic of mine.. the RE320.

[topaz hazel]

HECK yeah!! Those are beautiful! I use its cousin, the re20. EV mics sound fantastic

[wise gulch]

yeah, i was torn between the two really... price was a big factor as well as i got the RE320 for $160.

I think the RE320 will suit my tone as well

[topaz hazel]

Duuuuude thats like half off, what a steal! Yeah the only real difference is the 320 is a tad brighter and gives a bit more air to the voice. They sound nice

[wise gulch]

yeah i think i should be able to tame that with the dbx a bit, smooth it out a tad

[wise gulch]

It's amazing.

[wise gulch]

both mics running thru the DBX, no changes on setting except in post to bring levels up to match each other. the best i could anyway

[wise gulch]

My weakness is EQ... Here's me talking to myself about different settings and options I have with my setup. I really don't know what to exactly do for adjusting EQ

[topaz hazel]

Both def sound great, that's a fantastic setup! You nailed it, a lot of that low end can lead to a broadcasty sound. You'll always have a little bit of that sound due to the mic, the EVs are classic radio mics. If you're going to edit in post later I'd experiment with leaving the high pass off first and then you can play with filters on an EQ in post to find that surgical spot. 80 is a great catch all for HVAC or to help reduce plosives but with your timbre of voice it's on the line of eating some of your natural low end. (You have a great full body voice ).

Number one rule of EQ is to not use it just to use it. Go in with a problem in mind or you'll swiss cheese your recordings. You'd be surprised how little I use in film projects that are recorded properly.

[wise gulch]

thank you for having a listen... and yeah, at times i feel I don't need to EQ anything and let the mic do it's work. and the DBX. Maybe I'm so used to EQ'ing my old mics that I feel I need to now as well.

this is how is all sounds with gaming. the dbx settings, and only an additional gate, and a single band compressor in post. WARNING, jump scare moments fyi lol

[topaz hazel]

Dude that sounds fantastic! With the dramatic lighting and game genre your mic setup is about as perfect of a fit as you can get

[wise gulch]

did I show you what I did with my lighting? my mod i did?

i took my old scrap matt board for framing photos, and made them into like, partitions to make a soft box grid

a lot less light leak, and more focus on the subject. i.e. me.... I still need a better lens, one that is less light hungry. and i am making another light as well to fill in on my left side (picture right)

[topaz hazel]

Duuude tight! Saved yourself at least 60 dollars by making it 😂 looks great too!

[fallen axle]

is that color graded too?

[fallen axle]

where is the room echo coming from? also i am SORRY for the late response! i just saw it now

[topaz hazel]

Different per room. Common issues are emptiness (large rooms without carpet, couches, bedding, etc). High ceilings. Reflective windows.

[fallen axle]

im recording in the same room i do not hear any echo? do you have a specific timestamp in the segment? so i can try to listen in more closely

[fallen axle]

i do have wood floors but theres a lot of stuff in the room

[wise gulch]

i actually dont remember.. i typically use LUTS though

[topaz hazel]

Right in the beginning. "But..." you can hear it. Its not a killer, I wouldnt stress about it.

[fallen axle]

Dang.. sean you have some really good audio ears! 🤣

i listened to that at least 5 times before and i couldnt catch it

[wise gulch]

i guess my reply to the echo or reverb wasn't valid 😢

[fallen axle]

im sorry.. i didn't see your reply then where was it?

[wise gulch]

.

@fallen axle

[fallen axle]

Ohh