Single Word Too Loud or High Pitched, Sounds Wrong

[GuppyTJ]

Happy Holidays! Sneaking in a bit of editing before the family rolls in…

What is your advice for fixing a single word that is too “loud?” The word “diagnosed” in the attached clip sounds wrong to me. I don’t know if the problem is the amplification or the tone or frequency or all of the above. I do know the problem is me not reading it properly the first time. Haha!

For background, I am recording my book to become an ACX audiobook . From an editing perspective, the file this clip was cut from has been edited for spacing, declicker plugin ran and 36Audiobook-mastering-macro ran. ACX Check passes for all 3 criteria. There are 5 similar instances of this same problem in this 23 minute Introduction.

I tried to fix the word “diagnosed” in this clip using Amplify, Compressor, and Pitch but it never sounds good enough. I think probably what I need to do is re-record the sentence, then copy the single word out and replace it in the original. And try to not do this anymore and when I do, re-record it in the moment.

My equipment is:
Recording in my soundproofed walk-in closet.
Maono ProStudio 2x2 Lite USB Audio Interface PS22 Lite
Maono Professional Large Diaphragm Microphone Set PM500 Series Professional Audio Innovation
Maono Microphone Boom Arm BA37
Maono Au-MH601 Studio Headphones

If the answer is “re-record” that’s perfectly fine. I was not really able to find the answer in this forum so thought to ask.

Thanks in advance for the help, grateful for you all!
Theresa

[kozikowski]

See most of the blue waves are sharp and pointy or “expressive”…except that one blob in the middle with the flat tops and bottoms? That one is overloaded. Your voice exceeded the ability of the system to record it. You would think that would just stop getting louder, but no. Once a system does that, it also starts making up its own sounds and they can be sharp, sizzling, and crackly. So nobody is shocked you don’t like the sound of that word. It’s not all your voice.

You can reduce the overall sound of your recording and get rid of that during the performance. And no, you can’t fix it later.

It’s possible your system is misadjusted. Do a sound test. say “AHHHHHHH” and just keep getting louder and louder and get closer and closer to the microphone. We expect the blue waves to get taller and taller until eventually, they all sound crunchy, but they should do it at 100% or “1” on the blue waves.

If the waves get flat tops before +1 and -1, then there is something wrong with your recording system.

Koz

[kozikowski]

There it is. I knew I had that illustration.

This was a microphone adjusted wrong. It’s a single microphone recorded on a stereo (two blue wave) system and then forced to mono (one blue wave) later. They don’t much like doing that. They reduce the voice volume by half “to help you.”

Koz

[kozikowski]

One more note. Never, ever blow into a microphone, but you can usually sing or yell as loud as you want without breaking anything. Koz

[Trebor]

As Koz mentioned the word “diagnosed” is clipped which shows up in the waveform (flat-top) and spectrogram.

But IMO the resonance is the most outstanding problem, as it occurs constantly,
e.g. at ~380Hz, visibly this is most obvious in breath sounds.

Resonance can occur by sound reflected by nearby surfaces, but can also occur via sound leaking from headphones being picked up by the mic, (a/k/a “headphone bleed”).

[GuppyTJ]

@kozikowski

Do a sound test. say “AHHHHHHH”

Koz,

I did the sound check you recommended and you are right. My system is misadjusted. I was toping out far below + 1 and -1. I think that portion of my problem is I had the Audacity Recording Meter toolbar slider set to -12 (89%) and it was too low which was causing parts of my recording to be clipped. I have everything set to Mono so I think I’m good there, no Stereo.

I continue to struggle to understand, then properly configure and test the multiple volume controls: 1) the Windows 10/Manage Audio Devices/Recording/Microphone/Properties/Levels, 2) the Audacity Recording Meter toolbar slider and 3) the microphone dial on my Maono ProStudio pre-amp, 4) my microphone placement and 5) my voice volume. I have looked for details on how best to set them all, and never finding the full answer, I have fiddled with them for about 2 weeks, clearly, unsuccessfully. I thought it was set in the 20 second test of my work I posted earlier, but it was NOT. There are peaks in that 20 second test that were also truncated. This just goes to show how much I’m fumbling about and frankly, I don’t think it’s ever been set up correctly.

At this point in the below new sample, the pre-amp is still set to just below max and my Audacity recording meter toolbar is set to 0 and so is the Windows 10 setting. My microphone is hung upside down on the boom arm with the front facing me and is the hang 10 sign away from my mouth and a little to one side. But now because the Audacity recording meter is set so high, I can’t pass the ACX floor noise, even after the 36Audiobook macro is applied. It sounds like it’s because the microphone itself is making too much mechanical noise because it’s set too high. UGH.

I am considering returning this condenser microphone and getting a dynamic. I can’t tell if it’s my equipment or me bumbling around and not being able to figure out how to set all these interconnected parameters: preamp volume, Audacity volume, microphone placement and how loud I speak. Did I say UGH yet? I’ve been working on this for 2 weeks and I’ve learned a lot, but I still feel do not have the formula for my environment set correctly.

Much thanks,
Theresa

[GuppyTJ]

@Trebor

But IMO the resonance is the most outstanding problem, as it occurs constantly,
e.g. at ~380Hz, visibly this is most obvious in breath sounds.

Thank you, Trebor. I can hear it now that you point it out. I’m not sure why I can’t hear things until someone expert like you points them out. I sort of could feel that it was wrong but I just don’t know what good vs. bad sounds like. And in those graphs, I truly don’t know what you see and how you interpret them. Is it the yellow in the spectrograph you see that is the resonance? That’s why you are an audio engineer and I’m a girl struggling to get her audiobook read properly.

Here is a new version. Is it fixed? All I did was move the Audacity Recording Meter bar from -12 to 0. It seems to my ignorant, uneducated ear that the resonance is gone.

Thank you for your kindness,
Theresa

[Trebor]

@GuppyTJ

I’m a girl struggling to get her audiobook read properly.

Having listened to other examples of Audiobooks I may be nitpicking:
some of them have noticeable resonance too, and they are best sellers.

Perhaps I’m too accustomed to highly-processed audio.

There’s a plugin called Sonible Pure:EQ that automatically adjusts EQ and dynamically reduces resonance, and de-esses, (30 day free trial).

[GuppyTJ]

@Trebor

some of them have noticeable resonance too,

I listened to several of these, especially those that are read by a female voice. Is the “Hotel du Lac” one of them with a lot of resonance? I ask because I want my ears to learn what resonance sounds like so I can avoid it in my own work.

Theresa

[GuppyTJ]

@Trebor

There’s a plugin

I’m afraid more plugins and processing will make it sound more processed. Plus, I can’t tell from the analytics what the information is telling me and therefore, what the appropriate changes should be. I end up down a rabbit hole and more lost.

I’m trying to get my system set up correctly and learn to read well. I figure if I can get that much done, I should be good to go. Obviously, at least for me, easier said than done.

I appreciate the help.

Theresa

[GuppyTJ]

@kozikowski

the microphone

I know microphone is not everything and it’s possible it has nothing to do with the microphone. Still, I am using a condenser microphone and see that dynamic is recommended for audiobooks. I really do have an extremely quiet environment in the backwoods country on a dead end with truly zero traffic , flip the circuit breaker to turn the frig off, heat off, husband takes pups on a 2 hour hike so it’s just me in my closed up walk-in closet.

Here is what I have

Amazon.com: MAONO AU-PM500 Professional Wide Diaphragm Microphone : Musical...

Here are 2 other options I’m considering. I’m trying to stay with all Maono equipment.
Option 1

Amazon.com: MAONO Dynamic Microphone, USB/XLR Podcast PC Microphone with...

Option 2

Amazon.com: MAONO XLR/USB Dynamic Microphone Kit, RGB Podcast Mic with Software,...

Thoughts? Thank you again and again for your help.
Theresa

[GuppyTJ]

Here is a new sample.

This is a raw read with volume on the preamp turned down a little bit, Audacity recording meter set at 93%. Condenser mic detailed above at hang ten distance and a little off to the side.

This is the same read with 36Audiobook macro applied. While it passes all 3 tests, I can see where the “diagnosed” word is again truncated, the top flattened out. I think this means I still do not have the microphone set up right. I am not sure if I should turn down the mic volume on the pre-amp or talk with less variation in my words, move the mic further from my mouth, get one of the dynamic microphones linked above or something else? I am unsure how to properly test this even, to figure out which it is. Guidance on how to proceed would be so much appreciated.

Thank you,
Theresa

[kozikowski]

My personal standard is to pass noise by at least -65dB (higher number is quieter). I know the actual ACX standard is -60dB, but that’s how an obsessive engineer works. I applied Noise Reduction at the gentle 6, 6, 6 settings and re-tested.

Your background noise (Room Tone) is a gentle Surf/Spring Rain In The Trees and easy to ignore.

In My Opinion, you can submit that. I can listen to a story in that voice. I don’t hear any “gritty” words or other damage and your theatrical presentation is very much improved. You don’t sound tense and frightened any more. That’s the idea. Telling someone a fascinating story over cups of tea.

The classic Noise Reduction is a little involved to use and the program developers made it more interesting by hiding the controls—and then changing the Audacity version.

I’m using 3.6.1 for these tests. Installing 3.7 is on my radar.

Drag-select a portion of clean room tone. This is what the two-seconds at the beginning is for. Recording that is where you hold your breath and not shift around in your noisy chair. It’s environment and microphone internal noises only.

Effect > Noise Removal and Repair > Noise Reduction… > Get Noise Profile. The panel closes.

This is where you let Noise Reduction “sniff” the noise so it knows what to process.

Select the whole chapter > Effect > Noise Removal and Repair > Noise Reduction > 6, 6, 6 (top to bottom) > Reduce > OK.

You still have to pay attention to all the other requirements for submission. Chapter headings, Retail Presentations, etc.

This is also where other not-obvious rules apply. Everything has to match. There is no changing the microphone between chapters three and four, for example.

I need to find my audiobook reading technique instruction. Centuries of digging presenters out of trouble has given me ideas how to prevent troubles.

As we go.

Koz

[kozikowski]

Found them. The first part is how to read assuming nothing goes wrong. You never see this because something always goes wrong.

`

Read your chapter with wired headphones in a quiet, echo-free room. Repeat whole sentences if you make a mistake.

Stop.

File > Export the work, errors and all, as a WAV (Microsoft) file for protection.

Edit the work to get rid of word-o’s, stumbles, stutters, tongue ticks, lip smacks, and other errors.

Tools > Apply Macro > 36Audiobook-Mastering-Macro.

ACX-Check. If it passes and you like the way it sounds, Export a WAV (Microsoft) file as Edit Master and go on to the next chapter.

`

Read Your Chapter Detailed Notes

— Read your chapter with wired headphones…

Highly recommended. Once you get accustomed to your own voice coming back to your ears, all your volume, emphasis, and expression errors go away—or give you obvious notes that you need to fix them. Wireless headphones/earphones can be complicated and cause problems.

— a quiet, echo-free room.

Echoey, Recording-In-The-Bathroom Voice can’t be fixed, and the noisy Metrobus going by can’t be removed.

I can do very respectable recordings in my garage. The boxes make terrific soundproofing, but I have to do it at night for best neighborhood background sound and I still have to pause periodically to miss the Metrobus.

— Edit the work…

Edit the work to get rid of of reading errors. ACX told me my submitted test file was Practically Perfect in Every Way, but they couldn’t deal with my mouth noises. Tongue Ticks and Lip Smacks are hard to fix and have the story still sound natural. Good luck.

There’s a Post Production editing job what should not be entirely post production. If you make a mistake, stop reading, leave the recorder running, look back to the next even sentence or phrase and read the whole sentence again with the correction. Continue on. When you get to the end of the chapter, Export a WAV Microsoft Protection File.

Only then then roll the performance back and remove the broken sentence and do other corrections and edits. If you wait until next week to correct the error, you will never get a matching tone, pitch, emphasis, and rhythm.

— Tools > Apply Macro…

36Audiobook-Mastering-Macro is a one-step collection of tools that includes a rumble filter, a loudness setting tool, and tip and peak corrector. It’s possible your chapter will come out the other end sounding OK and passing the ACX audiobook sound standards. I actually did this once with a test I recorded on my phone.

— ACX-Check

ACX-Check will tell you your chapter sound standards: Peak, RMS (Loudness), and Noise. Note, last I looked, the on-line ACX Audiolab will not tell you noise.

Noise can get you into trouble. ACX-Check does not like Blackness of Space, Dead-Zero Sounds. ACX uses that as an indicator that you have over-processed your chapter. Heavy noise reduction can cause vocal tone distortions.

It’s not unusual to need a little push to pass ACX Noise. Classic Noise Reduction can be applied at the gentle 6, 6, 6 settings.

Koz

[GuppyTJ]

@kozikowski

I applied Noise Reduction at the gentle 6, 6, 6 settings and re-tested.

Thank you Koz, for the technical help and the positive coaching. Glad I’m sounding better. The manual above is helpful as well.

I see that you applied the Noise Reduction 6, 6, 6 to the second sample, the one I already applied 36Audiobook to. I did the same to make sure I could get the same result, (see attached screen capture) so I understand this and can absolutely do Noise Reduction 6, 6, 6 for each of my chapters. I am using Audacity 3.7. Terrific, thank you.

So… one last thing (yeah, right…), what about that “diagnosed” truncated word? It still sounds bad to me in the 36Audiobook macro applied version and thanks to you, I know why it sounds bad to my ears. I see that you said above you would be comfortable submitting this, yet this truncation happens pretty often. Now that I’m aware of how it happens, I can see it in my waveform other times as well in the longer recording.

Would you say I should not worry about this too much and try to adjust my
performance to avoid this but otherwise, march forward with recording? I did go ahead and buy the more expensive dynamic microphone above. My husband has a book too, so once I’m done recording mine, he will be recording his. His book is fiction with a lot of action and characters so he’s going to run into even more high and low amplification. Maybe a dynamic microphone will help.

Thank you,
Theresa

[kozikowski]

@GuppyTJ

yet this truncation happens pretty often.

Reduce the gain/volume setting of the microphone or microphone system very slightly. Or back up an inch or two and try it.

36Audiobook will master anything. The slight volume dip shouldn’t make any difference at all.

As above, I know what traditionally caused that, but not with the more modern systems. That brings us back to my global recommendation of not recording on the computer.

Zoom H1n audio recorder.

Once you start recording on the computer, you can’t use the computer for anything else that might affect the sound. No Skype, Zoom, Meetings, Multi-Player Games, etc. They all change computer sound settings without telling you.

Koz

[kozikowski]

OK. Different angle. I saw the Maono instruction manual and they said the microphone volume knob turns colors as your volume goes up. White for normal volume, pink for high volume, and red for clipping and sound damage.

How do they perform those colors?

Overload/clipping is really important so what many microphone systems do is flash it really bright and then hold it for a split second to make sure you see it. Is that how this one works? My impression is most of your voice should be white and some pink.

Do that sustained "AHHHHH thing again and force it to overload. What does the red light do? Ignore Audacity for a while.

If you’re listening to the PS22 with your headphones, your voice should sound crunchy during the red light events (clipping distortion).

I’m betting the PS22 overloads at a different volume than Audacity does (flat blue waves). We need to find out why that is.

The other knob, Mic2-Inst? Should be all the way off.

Can you read the instruction manual clearly? This is what it looks like here.

Koz

[kozikowski]

I looked up the PM500 microphone, too. It’s a side-fire microphone. You talk into the grill just up from the company name.

Koz

[RandomWanderer]

The methods mentioned here are useful, but if you have less time or want an easy method, you can try a simple noise gate or a loudness normalization

[GuppyTJ]

@kozikowski

Do that sustained "AHHHHH thing again and force it to overload. What does the red light do? Ignore Audacity for a while

You are a genius and a saint for looking up the manuals for my Maono set up. Yes, I read the manual and had the INST GAIN (Mic2-Inst) dial to zero. BUT, after many tests of AAHHHHH, and your ideas to focus on just the pre-amp lights, maybe I have this figured out.

It seems the mic volume is controlled entirely by the PS22 pre-amp. I believe it was set too high. It does turn red when my AHHHH gets really loud and exceeds the volume set by that dial. I really can’t hear the crunchy sound you described but I think it’s just that I don’t know what crunchy sounds like.

The only thing the Audacity recording meter toolbar slider seems to do is is CAP the volume at whatever the slider is set to. So, if I put the Audacity recording meter slider at say, 89%, it truncates (flattens the top) of any sound that exceeds that. If I’ve got the Audacity recording meter set to 89%, the waveform NEVER goes red to show clipping. Instead, it just flattens out the waveform at 89%. If I set the Audacity recording meter slider to 0%, it goes red when 0% is exceeded.

So, at this point, I turned down the PS22 mic volume on the pre-amp and put the Audacity recording meter slider to 0.

Here is that result as a raw read. I didn’t try to read it well, just wanted to get something to test so don’t mind the lack of good performance, please. I sound tired and squeaky, you can hear a click between sentences, that’s me turning the headphone monitor off to hear if that changed the sound and was creating the resonance Trebor heard earlier. You can hear a mouse click at the end too. So ignore all that, please.

Here is that raw read with just 36Audiobook macro applied. We can see that the “diagnosed” problem word is not flattened too badly. It’s not chopped off but seems to be reduced. It doesn’t sound terrible anymore either.

I think it’s fixed. Yet, I am getting the dynamic microphone, due to arrive today. I’ll try that one using these testing methods you taught me and see if the dynamic mic gives me more flexibility. The closer I get my mouth to this condenser mic to increase the volume, the more breathing it picks up. But I think you explained it above, that it’s ok for the volume to be lower, as the 360Audiobook macro easily fixes that. And the lower volume means the noise floor is lower.

THANK YOU, Koz, for going above and beyond to help me. I don’t know if someone compensates you for helping all of us. But regardless, do you take donations? I’d like to send you a thank you. You looked up the manual for my equipment and have spent a lot of time coaching me from a tech and voice performance perspective. I am better for it by big leaps. Tell me how I can send you some $$ as a simple thank you and recognition, for your patience and expertise.

Theresa