Does Noise removing twice badly impact the actual rec?


I am wondering, since I am using a C03u Samson mic and I notice an annoying background static.

I noticed that when I apply the noise removal twice (24, 0, 300, 0.15), I get an excellent result and the background noise significantly reduces and barely noticeable.

My questions are, if I apply the NR effect twice:

A. Will I negatively impact that actual recorded audio clip?
B. if so what should I then do? Can you suggest a better technique please?
C. What is your opinion regarding the settings that I am using for the NR effect? (24, 0, 300, 0.15)?

Thank you very much!

Each application of noise removal damages the audio to some extent,
(also the higher the settings for “noise reduction” and “sensitivity” the more the audio will be damaged : it’s a trade-off )

Below is a “before-after” using the audio you posted previously with the noise reduction settings …

noise removal.gif
After Noise Reduction then apply a noise-gate to reduce/remove the noise between phrases.

Thanks a lot for replying, can you please help with the attachment?

OK I will then apply the NR once, but I am having trouble with the most appropriate settings with a minimum damage to the recorded vocal parts.
Would you please help me and tell me please what are the steps and settings to remove the background noise?

Here is what I am doing, I open the clip (please see the attached), I raise the gain by 36, the background noise is there. I apply the NR but still it is not
dead silence (still background noise).

Your time and efforts are truly appreciated.
raw b4 editing (922 KB)

You’ve left out a step – select a “noise only” portion of the recording, do Effect > Noise Removal and click “Get Noise Profile”.

– Bill

Hi BIll, thanks a lot, but the audio clip that I uploaded is raw, no effects were done from my side, yet.
I am lost, because I dont know what are the settings that I should use for the noise gate, and I also tried the settings that you suggested in regards to the NR, but it did not do much impact.

Please walk me through, thanks again!

To totally remove noise in “silence” so it is “dead silence” you need to use a noise-gate plugin, see step 5.

Suggested work flow to reduce noise from “raw b4 editing II”…

  1. Amplify on default settings, (i.e. to 0dB)

  2. select a region for the noise-profile which has no voice, e.g. 1.5sec to 6 secs.

  3. Apply Audacity’s native Noise Reduction with the settings: 10dB, 6dB, 1000Hz, 0 seconds , remove .

  4. Optional : select all of your track and put the code below in the Nyquist prompt to remove the faint 1KHz whine + 10 harmonics , ( I’m nitpicking here : you may not be able to hear the faint whine, if so don’t bother with his step).

    (setq mysound s)
    (setq q 10)         ; set the base Q for the filter
    (setq iter 10)     ; set the number of iterations
    (setq freq 1000)  ; set base frequency

    ; start the DO loop
    (dotimes (i iter mysound)   
    (setf mysound (notch2 mysound (* freq (1+ i)) (* q (1+ i))))
    )                     ;end of loop

code to remove 1KHz whine.gif

5. Select any conspicuous unwanted noises between phrases, e.g. 11.40-11.60sec, press “Ctrl” +“L” to silence them.
Then select the entire track and apply a noise-gate to squelch the noise between the phrases.
Steve's noise-gate plugin.gif
[continued on next post …]

  1. select the entire rack and apply a limiterto reduce the variation in volume, and make the track generally louder [this is optional].
    Steve's limiter plugin on default settings.gif
    the final product …

I was able to get rid of much of the fan noise with a simple notch at 171Hz Q=10. I don’t think anything that far down is likely to impact the voice. Then you can go into the gymnastics to get rid of the hiss. Gentle rain in the trees hiss is less objectionable than the fan hum and you may find you need to do less work.


Hello friend, seriously thanks to your time and awesome efforts.
I copied the same steps to the letter, but for some reason I am getting the same results. Your clips has NO flickering in the output level meter.
Mine still does.

I cant explain it, sorry to bother you again.

I can see that you did a great job on my clip, but the one I attached is no where near yours, why?

Best Regards for the awesome work that you guys do!
raw b4 editing II.rar (992 KB)

The noise on "raw b4 editing II.rar (991.54 KiB) " during he first 4.5 seconds is below -70dB …
''raw b4 editing II'', noise is below -70db = noise is inaudible.gif
that very low level of noise is inaudible. If you change the "level reduction " on the noise gate to -100dB it will make the silence truly silent, flat-line, however there won’t be an audible difference from the current -70dB level unless you amplify it by an unfeasibly huge amount, (the voice will be clipped if you amplify the track by more than 1.4dB).

I disagree : I can’t hear any difference. Maybe your computer is adding some noise when it plays back audio. Attached is 5 seconds of proper flatline silence : if you can hear noise on the attached WAV your computer / soundcard is adding it.

Hello, I know that you are the experts and I value your opinions.
But, if you play the flac file with 36 gain, the output meter is silent, was the noise gate “level reduction” placed at -100dB?
because that was not mentioned in the 1-6 steps (step 5).

You said that you dont agree and that there is no difference between the flac file and my uploaded clip (rar).
I feel that my clips is not like the flac file in terms of noise elimination. here is how i can prove it.

Play both on audacity, the flac is dead silent in the quiet parts, my clip has fluctuations in the output meter. If my computer has a problem then why did the flac file have no fluctuations whatsoever?

I wouldn’t go that far

There is no need apply 36dB gain, doing so is unfeasible: if you apply more than a couple of dB gain the peaks and troughs of the audio with speech will go off the chart and be horribly distorted by clipping. In step 5 the I set the “level reduction” on the gate to -22db which took the noise below -70dB : that’s good quality. If you must have flatline silence push the “level reduction” on the gate to the extreme of -100dB. However having flat-line silence between phrases can sound weird : if you leave a long pause listeners will think they’ve been cut off, leaving a tiny hint of noise between phrases is preferable to removing it completely which sounds unnatural.

Where noise has been squelched out completely [flatline] sometimes a little noise is added for listener comfort …

i.e. total flatline silence between phrases is not desirable.

That is truly great and did help me take my recordings to the next level. Thanks a whole lot!
One last question, now all that takes great care of what I am recording,
But i will eventually add clips from others into my spoken clips.
I am talking about an intro music for my podcast, and friends who are giving a short introdcution
To the show. Is it a good idea (no negative impact to my
Recording) to copy and paste my clip, the music, and the intro, export to WAV
And apply levalator to the combined clips??

Thanks and best regards!

If the music has been professionally recorded then applying dynamic-range-compression, (a.k.a. compression, a.k.a. levelator, a.k.a. limiter), to the music is not advisable unless you want to make it louder than normal … Loudness war - Wikipedia

On Audacity you can have many tracks running simultaneously, say one track with your voice, another track with another person speaking, and a third track with the music. You can independently adjust the volume of each track in the mix , and apply dynamic-range-compression only to the tracks with speech and leave the music track with the dynamic range the musician intended.

Hello friends, sorry for not getting back sooner, I was testing and applying the 1-6 steps, and I squeezed the notch filter as step 6 (Koz’s advice: 171, 10), and made the limiter step number 7. The end result is attached. Please tell me:

Q1: What do you think of the outcome? Does is sound quite good?
Q2: Should I add any other effect?
Q3. Do I need apply Bass (increase or decrease)? if yes how much?
Q3. Do I need apply Trebele (increase or decrease)? if yes how much?

Thank you so very much!!
Truly friendly and your help will never be forgotten!
done with fx 1-6 do i need bass or treble.rar (618 KB)

You’ve not done all of Step 5 : silence any conspicuous unwanted noises (spikes) which occur between phrases …
silence spikes by selecting them and pressing 'Ctrl' +'L', then apply noise-gate.gif
, then apply the noise gate.

BTW the equalization (the treble&bass content) seems fine to me.

thank you very much! here is an updated and adjusted clip, and hopefully i did it right, your feedback really matters.
btw does ‘Noise gate’ only work if I remove ALL the ‘any conspicuous unwanted noises (spikes) which occur between phrases’ or doing and silencing 2-3 is enough for the noise gate to take effect?
done with fx 1-6 do i need bass or treble 2.rar (846 KB)

The noise gate will still function if you do not manually remove these noises, but it will perform better if you do.

If you set the threshold of the noise gate higher, say -30dB instead of -50dB, it will remove more of these noises automatically, but there is then a danger the beginning and end of words will be chopped off.

ahhhh ok thanks, I will keep that in mind. btw so the last audio clip that I attached is good to go? is it good enough in terms of recording and quality output
for the audience online/CDs?

It’s quite good, however here a few minor clicks that could be fixed using Audacity’s “repair”.
IMO it needs a touch of de-essing … De-essing - Wikipedia
Also the track is mono, people prefer stereo. It is possible to add a pseudo-stereo effect with Audacity which also smooths the transition between the slight reverberation of the room you recorded in with the artificial dead-silence created by the noise-gate …

CD is old-hat, online is where it’s at : essentially no up-front costs.