Hey everyone, I’m looking for someone with more experience than I do to see if they can find the right equalization for my voice. I’ve noticed recently that the speech that I have been making is not quite as crisp as I would like it to be. I’ve tried messing with EQ bass and treble but I don’t understand it well enough to make the right adjustments.
If you find a good EQ balance, please post it here.
I posted a speech sample in the attachments below. The audio sample has been normalized and compressed, so let me know if you need a raw sample. Also, if you notice anything else that’s off (besides the static noise in the background. Decided not to use noise removal ) please let me know.
I’ve noticed recently that the speech that I have been making is not quite as crisp as I would like it to be.
How are you listening to it? The speakers or headphone type could make a big difference. How is most of your audience listening to it?
You could be paused on the precipice (so to speak). Unless you have a perfect recording, any noises are going to go up and down with your performance correction. I imagine your buzz is going to go nuclear if you try to bring up the crispness.
Yes, please. A raw clip. Don’t help it. Just perform like you normally do.
You’re sure you didn’t slide a little noise reduction in there? That clip appears to have pumping noise which is a feature of the older Noise Reduction tool.
I don’t hear anything wrong with that clip the way it is. Most people are frantically trying to get rid of harshness, sibilance and too much crispness from microphones trying to “help them” be “professional.”
I would have no trouble listing to that voice tell me a story, and it very nearly passes ACX AudioBook conformance if you care.
Alright here is the new sample you requested. No normalization or compression.
Answers to your questions:
1: I normally listen through my Sennheiser HD 558 headphones. I understand that Sennheiser does things to enhance audio so I am probably listening with a bias.
2: The headphones do make a big difference. My voice sounds much clearer and more pronounced through the headphones than through speakers. My speakers seem to distort the audio, rather than continue to make it clear like this guy https://www.youtube.com/watch?v=_dcmkufzLto. His voice sounds clear through both the headphones and speakers.
3: Would not know about the audience but I would guess mostly through headphones.
Other things to note: I think I agree with you on the buzz. I have had difficulties where after I edit my vocals and do a noise removal to get rid of background noise, the static or background noise is not fully removed from the parts where I am speaking. Any pauses are fine (all static is gone) but when my voice starts up again you can hear the buzz a little bit. Any ideas on how to fix this? (eg: soundproof room, get a mic stand with a shock mount, etc)
-Have a blue Yeti (Standard)
-Sometimes the microphone seems to pick up too much. I’ve been messing with gain and microphone level settings on the computer, but I haven’t yet found that perfect level.
“A little soft” is putting it mildly. The level on that raw recording is WAY too low, by at least 20 dB.
I don’t know enough about the internals of the Yeti to say what the best combination of hardware gain control (the “gain” knob on the back) and the software volume control (the slider next to the microphone icon in Audacity). But you need to turn one, the other or both up to get the level reasonable. When you are speaking the “peak” bars on the Audacity meters (the blue lines) should be bouncing in the -8 to -3 region. And yes turning this up will make the noise louder too.
There’s many sources of noise to worry about: Acoustic noise in your environment, thermal noise in the microphone preamp (that’s the “sssss” sound), electrical interference (hum from the 60 Hz mains or whine from the USB data signals), and quantization noise (the limit to which 16-bits can represent a signal). The relative setting of the physical gain knob on the microphone and the software gain knob might make a minor difference to the amount of the thermal noise (The “ssssss” hiss), but unlikely to have much effect on any of the other sources.
Generally you shouldn’t have to worry about “quantization” noise UNLESS you record at very low levels, at which point it will get amplified to unacceptable levels when you “normalize” the recording. (Audibly quantization noise is pretty much indistinguishable from thermal noise both go “ssssss”).
Listening to the background noise on your test file I only hear the “sssss” noise, but it’s likely that was because the record gain is so low that the quantization noise is masking the other issues.
Some other basics to check: The Yeti has a number of “patterns”. For solo voice the best will be the “cartiod” pattern (the icon looks like an upside down heart). With that pattern set you should be talking at the side of the microphone (the “Blue” label should be facing you). That arrangement will make the microphone a bit less sensitive to background noise and echos, but not a lot. To get a good recording there is no substitute for a quiet well-damped room.
That’s way too low. When you record, you’re intended to watch the sound meters and hit the yellow region in Audacity 2.1.0.
And the blue waves should bounce around 50% like the recording I made as a noise test (attached). That’s a raw recording, too (but I’m not using a Yeti).
As above, make sure you’re talking into the blue logo. The Yeti works from the side, not the end. Set the Yeti for “Cardioid,” turn it all the way up, and get closer until the Audacity meters and waves work.
All microphone systems have noise. Your job is to make your voice way louder than the noise.
Low volume is a very common problem with USB microphones, and it causes lots of problems in post production.
That and there is another test you can do. Overload the microphone. Turn everything up and keep getting louder and louder and louder until the Audacity sound meters turn red and go all the way up (warn the neighbors). Good practices recording volume is a little lower than that.
Alright here is a third sample. I messed with the gain on both the microphone and the computer, and managed to get something alright. I think you’re right about the USB microphone catching lower levels, because even when I reach the range of -3 to -12db the levels on the timeline seem to stay below the -0.5 mark. If I raise it any higher the levels will jump up to the point where it’s catching too much. The audio still sounds strange through the speakers compared with the previous video I linked. I feel like it has to be something with the microphone or the USB connection to the computer (although I could be way off the mark).
The levels in that clip are just fine. Your voice sounds just fine too.
There’ s lot of extraneous noise in the bits at the head and tail of this piece. Possibly stuff being moved around on the desk, or a noisy chair. You need to avoid those sorts of things, I suspect without that stuff the ambient noise floor will be just fine, but can’t tell for sure.
You say “it sounds strange through the speakers”… Have you listened to your clips with good headphones? What sort of speakers?
There is a fair amount of low-frequency rumble in the recording you might want to try the attached 100 Hz rolloff and see if that helps. (On the small chance the issue it your sub-woofer reproducing the rumble and that’s what is sounding odd to you).
But otherwise you do NOT want to equalize for what “sounds good” on cheap computer speakers. The reason is you won’t hear what you are doing to the material outside the range the cheap speakers can reproduce, and the result will be awful for anyone with good equipment. LF_rolloff_for_speech.xml (299 Bytes)
You’re right about the noisy chair. Good call. Unfortunately there is not much I can do about that for the moment.
I have Sennheiser HD 558 headphones so yes I have listened through good headphones. The problem is, the speakers are not too bad either. Of course they’re no match for the headphones, but as stated in earlier posts, the clarity and crispness goes down through the speakers compared to other examples played through the same speakers. I’d like to find a balance between the two but if you say prioritize headphones I will do that.
( I don’t have the exact name for the speakers but hardware is Realtek HD Audio)
Low frequency rumble? Not sure what that means exactly. Also, the rolloff XML file you gave me, is that a plugin?
I’ll try not to equalize too much for the speakers. but I would still like to find a good equalization level to maximize the crispness and clarity of the speech. Any help on that front would be much appreciated.
If you do “Analyze → Plot Spectrum” on a section of your recording where you are not speaking, You will see that there is a lot of energy below 100 Hz. There is an upward trend from 7Hz down that I suspect is electronic “1/f” noise from the per-amplifiers in your Blue Yeti, but I suspect that a lot of the stuff between 20 Hz and 100Hz may be in your recording environment.
And sorry I should have explained it. The file is an equalization curve for Audacity’s equalization function. If you bring up the equalization panel (effect → equalization) then under “save/manage curves” you will find “import”, select that, and choose the LF_rolloff_for_speech.xml file (assuming you have already got it downloaded to your computer somewhere).
You probably won’t hear any difference after applying this filter to your recording, but it will make a significant difference in the measured noise level.
As for adjusting the “crispness” then I would probably just try a typical “treble” adjustment, put two points in the equalization curve at 1 kHz and 2Khz then add a bit of boost to everything from 1 kHz up. As they say in recipe books “season to taste”.
Note the first two seconds are freeze and hold your breath. Yes, we can hear your pants and shirt moving. I makes measuring the electrical noise of the system very difficult.
“That’s either microphone thermal noise, or your jeans could use a good washing.”
Beginners always underestimate how sensitive this stuff is.
We’re not entirely sure what causes it, but many recordings arrive with a lot of church organ/thunder/earthquake rumble. Most of it is too low pitch to hear no matter whose headphones you use. Steve’s rumble filter strips all that away and makes sure that your tools and filters are working just on your voice and nothing else. It’s not a normal plugin. It’s designed to be installed into Effect > Equalization. It’s run with the “length” slider all the way up.
Open the show or Generate > Noise> OK. It’s just to get something on the timeline.
Effect > Equalization points from about 4000 and up affect brightness or crispness. Experiment. I prepared a simple curve attached. That’s approximately what “voice” microphones do internally to sound brighter.
Hmm I think the rolloff does get rid of some of the ambient background rumble, so that’s good. Now I just need to figure out what’s up with my voice. It SEEMS like the microphone is catching too much of my voice (though I may be wrong), which is causing some strange distortion to these speakers that might not be able to handle it. Here is an edited sound byte (it’s a bit of a rush job so if you notice anything let me know). If you can, compare it through your speakers to a notable commentator or podcaster (eg: Markiplier). There is a significant difference in quality compared to his. I want to figure out why that is. If you really don’t think it’s important, and that I should continue to prioritize through headphones, I’ll do that.
I’m wondering if my hearing is off. Personally I think the Before sounds a little cleaner than the After. I guess I should clarify (to be sure), I don’t want my voice to sound like the other commentator, rather, I just want the same level of quality coming from my voice. The one thing that seems to be recurring is that through speakers, my voice comes off as dirtier, with a little distortion. Perhaps it’s the particular tone of my voice, or some issue with my microphone, I’m not sure.
On a side note, when I say speakers, I mean multiple types. I recently tried listening through the iPhone 6 speakers and got the same sort of distortion, although it didn’t sound quite as bad.
Also, I do appreciate what you guys have been doing for me. I completely understand if I’m not a priority.
It is cleaner. I’m going with my first comment. I would have no trouble listening to a story in that original voice.
The reference you posted as an example has a more forward, sharper sound than you. I guess it helps that he’s screaming into the microphone a lot of the time, but that’s him. It’s possible to accept the curve I made and change it so to hit it in the middle, be sharper than you but not as aggressive as him.
I can’t tell. Do you want me to talk you into using your natural voice? Count me in. Other posters would kill to have your natural, clean presentation.
We do note, though, that there’s no shortage of people wanting us to make them into actors. There’s no filter for that. They still hire actual actors to be the voices in animated movies, obviously not for their good looks.
Emmet in the Lego Movie was voiced by Chris Pratt… because he can act.
Haha that’s very nice of you to say. I don’t particularly want my voice to change, but, I want to figure out why my audio samples come off dirtier than other commentators. It’s no problem if you have decent headphones. (even iPhone earbuds make the sound quality alright, although not as good as my Sennheiser headset), but any speakers seem to be a problem, whereas for these other commentators out there, they don’t have the same problem (eg: I can listen to that Markiplier video with no distortion through my speakers, but my own audio samples sound bad when played through the speakers). I’m wondering if it’s because my voice tone is lower, and is resonating with the speakers more.
As for the equalization, most the issues that I’ve been having have been solved, or are in the process of being solved. So I guess the main point of this thread is now moot. I do appreciate the help, and if you have any ideas about the other problem, let me know.