Odd Audio Effect I Want Gone

I have been recording audiobooks a bit at a time at a radio station I work for, and decided to set up my first home system. Using it, though, I am getting an odd effect. It is especially evident when words start with vowel sounds, and feels like the start is being cut off - almost an “anti-clipping”.

I tried adjusting the compression going into my computer, and high compression did seem to minimize the problem, but not eliminate it. My setup is as follows:

Yamaha Audiogram6
Dell laptop with Intel Celeron, Windows 8.1 (64bit).
Audacity 2.0.6

In the clip I’ve attached, my only post-recording processing has been to amplify the track a little. Compression on the Audiogram was all the way down, to the left. You can hear the effect at 2 seconds in - at the beginning of both words in “and over”.

Any thoughts?


Thank you for the system details.

In the clip I’ve attached, my only post-recording processing has been to amplify the track a little.

How little? We warn people not to do anything to the clip before posting. We had one poster make the clip louder “so we could hear it” and it turned out they were recording sound just above the noise floor. Not surprisingly, they were complaining about noisy sound.

That and there is a target clip format.


Did you decide to do this because of someone else you like who’s doing this? What prompted the move? I will warn you the longest thread on the forum is Ian in Hollywood who wanted to record audiobooks from his apartment in Hollywood. That’s it. A year later and 39 forum chapters and he’s on his way.

As a first step, it’s strongly recommended that you turn off all processing in the sound pathway. Everything, including Windows Enhanced Services.


This can get crazy enough that we may recommend not using the computer at all to record your voice. I do work on a Zoom H4 and Olympus stand-alone recorders. There was a recent performer on a Zoom H4n. One of my favorite music performers works with a classic Zoom H2. All the computer problems vanish when you do that.

In some circumstances, we recommend recording at the local radio station.

As we go.


Your processing, distortion and noise floor are unstable and moving. You need to find and kill all your little processing “helpers.” Something in your setup is trying really hard to “help” you make a personal call or conferencing connection.

For one example, the noise in the middle of the performance is -55, but the noise at the end is -70. That’s not real life. Something is processing your work.


I’m not just making that up. “Overprocessing” is a common ACX AudioBook rejection.


Another note. I know your interface has a 24-bit, 192KHz sampling service, but if your computer can’t handle a real-time USB connection that fast, you may get holes, glitches and stuttering in your recording. The recommendation is connect at plain, ordinary 44100, 16-bit, Stereo (Audio CD Standards) and get that to work perfectly. Then, if you’re wearing your adventure shirt, go for the higher standards.


IMO it’s too bassy : needs to boost the treble …

In Audacity 2.1.X real-time equalizer adjustment is possible with a free plugin , see video here … https://forum.audacityteam.org/t/version-2-1-1-0-doesnt-remember-equalization/39205/5 [ Windows ].
That allows you , by trial & error , to rapidly obtain the optimum equalization for your voice.

Thanks for all your help thusfar. Sorry I was MIA yesterday - my “easy” day turned into the busiest of the week. At any rate, here’s what I’ve done.

The stock driver/hardware combo for this computer (Realtek), unfortunately, does not give me options to tinker with its settings. I was able, though, to disable noise reduction for the built-in mic. I did have the mic disabled already, but tried that anyway. No change.

Though my computer said that my driver did not need updates, I found there was a patch. That didn’t help me, though.

I had an unopened copy of Cubase AI-4 booting around (came with my Audiogram6), so I contacted Steinberg and got activation codes. I had not planned on learning a new system as I have used Audacity on and off through the years & am also familiar with Adobe Audition, which feels similar to me. However, I thought it may be worth a try to see what that sounded like.

Long story short, The audio was not improved any by running it through the ASIO4ALL driver & Cubase.

Another option will be to hook my Audiogram6 up to my gf’s desktop. On the plus side, I know her machine is more powerful than mine and has a better sound card. On the other hand, the best location for me acoustically is our garage attic - there are no electronics there, it is isolated from house noise, and has a sloped roof & dormer windows to break up echoes.

I did find that if I record a bit hotter than I feel comfortable it lessened the effect, which points to you being right - that there is some sort of noise cancelation going on that I’ve not yet found access to.

Regarding my reason/goals for recording, short term it’s simply a creative outlet. Mid term, I wouldn’t mind eventually making a little pin-money to upgrade my mic or get a dSLR that’s capable of doing decent video.

Thanks again for all your help.


ps: Love the “Morgan Freeman” version of my voice!

I only skim read this. Are you using Koz’s recommendation of recording at 44100 Hz?

Are any audio effects coming from the Audiogram? What mic are you using (make and model number)?



And what do we note as common thread running through all these changes? I’m not sure I trust a mixer whose top selling point is all the live special effects it can do.

I note that I get two different mixers when I search. There is a super simple Audiogram6 with like three knobs on it and then there’s the one with tons of effects.


Which one have you got? And yes, do post the info on your microphone.

On the other hand, the best location for me acoustically is our garage attic

Exactly correct. Non-parallel hard walls can work almost as well as heavy soundproofing. Exactly parallel hard walls cause problems. I had an office which would sustain a three minute ringing echo from a simple hand clap.

Mid term, I wouldn’t mind eventually making a little pin-money to upgrade my mic or get a dSLR that’s capable of doing decent video.

We like the Canon 5D. There is a story behind that camera. The Original 5D came out with movie features, but not 24-frame movie speed. The Entire Planet screamed as one WHAT ARE YOU THINKING?? and Canon quickly brought out a “New and Improved” 5D which did. Many “movies” “filmed” on a 5D.

It may be very small pins. Everybody Knows you can create your own videos. One of our Respected Hollywood Directors cranked out a show using iPhones just to see if he could so it.

He could.


Just a side note. Nowhere is it written that your equipment can’t be broken. See how small you can break up the system and test each segment.

Processing isn’t the only thing that can produce sound like that. You can also get that if the electronic biases are off or mid-point of the blue waves is just flat missing. That will also sound much better as you get louder.

Make it worse. That’s another troubleshooting technique.

Try recording quietly from across the room. I expect you to have a noisy, hissy voice. I bet your voice vanishes altogether.


If you apply [“high”] dynamic-range-compression before cutting-back the low-frequencies [ via equalization ]
the parts of speech which include strong bass-notes , [ e.g. vowels ] , will be made selectively-quieter by that compression. The musical-equivalent of this phenomenon is called compressor-pumping : the volume of the lead-melody dips every time there is a loud bass-note.

To avoid that , cut back the bass (effectively boosting the treble) , before applying compression.

There’s no symptoms of driver-problems or audio-driver-based “enhancements” on your recording.
Equalizing first , which you can now do in real-time in Audacity , then compressing will improve intelligibility ,
( to avoid the selective-attenuation of bassy bits ).

There’s no symptoms of driver-problems or audio-driver-based “enhancements” on your recording.

That’s where we part company. Nobody can get a -70dB noise floor in a casual recording right out of the gate. I can’t do that with a good room and good quality equipment. The signal is being “helped” somewhere.


Maybe it was hardware after all - after moving things around and disconnecting/reconnecting equipment to try to pin down things, The main issue I was worried about seems to be okay to me.

I’ll attach a test of ambient room noise & my speaking voice. My setup is: Samson R11 → Yamaha Audiogram6 → Dell Inspiron → Audacity.

The Yamaha, by the way, is just a simple USB audio interface with no equalization. The only pots for each line are compression, gain & a level for that line. I have compression set in the middle at the 12 o’clock position. At the moment, money is very much a consideration - are there any good entry level USB interfaces I should be keeping an eye out for?

Thanks for all the EQ tips, by the way. I currently have no external EQ, and need to just take a day to sit down and play more with Audacity’s EQ functions.


The built-in equalizer in the latest version of Audacity [2-1-1] has a bug : it “forgets” the equalization settings , which makes tweaking the equalization frustrating. If you have a Windows computer , my suggestion is you install an equalizer plugin , if you do that in 2-1-X you can get real-time adjustment to equalization , see … https://forum.audacityteam.org/t/version-2-1-1-0-doesnt-remember-equalization/39205/5 , ( a facility which Audacity’s built-in equalizer does not offer , even when bug-free ).