A few beginners questions about Audacity

Audacity 3 Help Forum Windows
1

Hey everyone, my first post on this forum ! :slight_smile:

I have a couple of noob questions about using audacity, it would be great if you could offer me the answer for them.

  • First of all, i have a very cheap microphone which i am pretty sure is mono only. Uptill now i have been recoring in stereo and logicly this recording was louder then mono because stereo basicly recorded 2 same mono channels and when you play 2 of them at once as oposed to 1, its louder. Is this the same as if i were to record in mono and then just copy this channel and paste soo there would be 2 clones ? I think it is the same.

  • I have a bunch of recordings containing human speech but they are quite flawed due to bad microphone quality and the noice around. Microphone as it is produces a lot of noice/hiss and such, but if you add cars, trucks, wind, birds and other such things…you can barely understand the speech. What are the options of clearing such recording to make it less pleasant to hear and understand speech easier ? I know it is not as easy as it is on movies where they have birds and human speech on same recording and they simply click “remove bird singing” and that i would probably have to try different things.
    But where would be a good place to start ? I tried playing around with normalize, amplify and level. I tried to see what the difference in each of those is but to me it didnt really sound different ( if i compared them to eachother ) and was very similar to general Db boost. This recordings are not all the same, in some for instance you can barely hear the car since it is very far away and in those it would be best to just remove noice, in others the microphone is basicly next to a running car and here it would be more importaint to remove the car noice.
    I tried noice removal but i wasnt very lucky. All i did was remove almost everything and got a bunch of robotic sound. In recordings where speech is louder i had more luck with noice removal, but on recordings where “other noice sources” are louder then the speech itself, removing noice did more damage then good.

  • Right now i am recording at 32bit 44,100Hz and i export either to Wav 16bit or Flac 24bit ( should i notice the difference in those 8 bits ? ). Since i can only export to 24 bit, is it worth recording at 32bit, does 24bit conversion simply cut those 8 away or does it use them somehow ? I have been wondering if it is worth to record in 32bit anyway since my microphone is of such low quality and 32bit vs 24 or 16bit makes the recording a lot larger and editing slower. As far as exporting files goes…flac 24bit ( or 16bit ? ) is the best ?
    I dont suppose there is some quick microphone trick i can use to improve its recording, mostly that background noice bothers me. I even tried recording in complete silence and it records noices i cannot hear at all.

This is all for now, i hope you will be able to answer my questions. I am having lots of fun with Audacity, great program indeed !

2

cars, trucks, wind, birds and other such things

Hiss is difficult, but wind, birds, etc. are impossible.

i have been recoring in stereo and logicly this recording was louder

Logical, but not quite right. Music systems play a whole copy of one mono track to each speaker. They don’t divide it half and half. The volumes between that and stereo are the same. There is a way to damage stereo where it only plays in stereo, but given most healthy systems there is no difference.

Audacity always works internally at 32-bit. The reason to use higher bit rates is not always so the music sounds good. It’s also so you can do production/effects/filters without the music falling apart.

I even tried recording in complete silence

Most people don’t hear their environment noises until they try to record live voice or music. “Is that my refrigerator making that noise?” One environment problem that’s not noise is echo.

We can’t fix that. People still rent studios for a reason. There are people at work who hung blankets inside a closet to get rid of recording echoes. I use furniture moving quilts.

574Ă—1000 51.1 KB

Audacity is not a particularly good forensics program. We can’t rescue your show from the grave like they can on the TV shows. Notice that not once in all this did I trash your microphone. It’s been my experience that most microphones can do a credible job and what kills them is everything else. I wrote a list of guaranteed ways to kill your sound show.


– The Four Horsemen of Audio Recording (reliable, time-tested ways to kill your show)
– 1. Echoes and room reverberation (Don’t record the show in your mom’s kitchen.)
– 2. Overload and Clipping (Sound that’s too loud is permanently trashed.)
– 3. Compression Damage (Never do production in MP3.)
– 4. Background Sound (Don’t leave the TV on in the next room.)

3

Even with a great studio environment, a microphone will sound low quality if the microphone input is low quality.
The microphone inputs on most PCs are rubbish and will record a lot of hiss along with your show. An external sound card or an external microphone pre-amp plugged into a “line level” input can make a vast improvement, even with an inexpensive (but not totally rubbish) microphone.

In that situation the car is not “noise” it’s a performer in your show. Low level background hiss or hum are “noise”. TV’s, cars, dogs barking, birds singing, cockerels crowing, pneumatic drills, and anything else that is not a low level constant sound are “performers”. Noise can be reduced but taking out a performer is like trying to remove sugar from a cup of coffee.

4

How do i reduce the noice ? I tried that noice removal and it basicly removed everything and replaced it with occasional robotic sound. Settings were probably wrong and too severe…
My sound card is low price SB audigy value 2…i will keep both microphone and soundcard for now, and try to learn some editing, because like i said, i am already stuck with a bunch of recordings ( i even have a few where i am playing a piano and while i can always perform again, it would be nice to at least practice on those recordings ).
On those recordings things are occasionaly soo bad that you cant even understand anything…not sure if simple noice removal can help here. Studio is offcourse an option but as such, for future, not for recordings already made.

5

How do i reduce the noice ?

You very probably don’t. I think we should have called it something else because our “noise” is different from everybody else’s. Noise is a constant background drone like an air conditioner or maybe computer fan noise or even low microphone hum. It’s not “Whatever I Don’t Want.”

The noise has to be very much lower than the show and never change during the performance. That kills car noises, dogs barking and leaving the TV on by accident. Those are now performers in your show and we can’t stop that.

soo bad that you cant even understand anything

That’s when you include some of the performance here so we can listen to it. There are certainly really evil things you can do to recorded sound by accident and we may be able to help. This is an instruction pack on how to send sound clips to the forum.

https://forum.audacityteam.org/t/how-to-attach-files-to-forum-posts/24026/1

Koz

6

Thank you for your help, i am sorry for my late reply,

I found some time to play with audacity and it indeed is a great tool, yet easy to use. I found out that cutting certain frequencies helped a lot. For instance car driving nearby can be almost completely removed if you cut frequencies under 300Hz with highpass ( before that i used lowpass under 300Hz to see what i am about to remove, and basicly you cant even hear speech ). Same with high frequencies, over 5.000Hz there is basicly no speech, just noice. I found out though that if i remove frequencies over 5.000Hz, recording changes a lot, becomes kind of empty…even though it doesnt affect speech or voice, it is kind of life-less.

I have a few questions if you dont mind helping me.

  1. removing frequencies with high and low pass. What is this rolloff option which you can choose from 6 to 48dB. I played with it and the best i can tell is that if you have a noice that produces 48dB on your recording, if you choose 48dB rolloff, the noice will be removed, if you pick 6dB, it will only be “silenced” for about 6dB. I also noticed that sometimes when i use highpass 1000Hz ( removing everything under 1000Hz ), the line wont be completely flat ( lets say i also removed all noiced of 1000Hz and over, therefor should have complete silence ). But if i used it like 3 more times, the line will be completely flat and there will be no noice. I guess those noices were over 48dB soo it took a couple of hits ?
    When i am cutting frequencies under 600Hz, do i accidently damage frequencies over 600Hz aswell ? Or is this simply cutting everything under 600Hz and wont affect Frequencies over 600Hz at all ?

  2. I have been wondering. Lets say we have 2 noice sources, both producing exactly 500Hz but one of them having 30dB and the other having 50dB. Now amplyfying it both would change nothing. How about if you set it soo that everything over 30dB is cut ? This shouldnt affect the noice producing 30dB at all, since it will just cut everything over 30dB. And after this editing it would be as if having both sources producing 30dB, and in turn you will hear the first one much better then before, right ? I tried leveler, normalizer and amplyfier but i am not sure which one exactly will do what i want. It is importaint that i dont want this leveling to lift first source from 30dB to 50dB, because any amplifying of it will affect its quality. I want this just as if different balls were bouncing on the floor and you would put a big wooden piece horizontaly over them at the height of 1m. It wouldnt matter if 1 ball would be able to fly 2 meters in the air, since the wooden piece is there, ALL balls will only reach 1m. I hope this isnt too confusing, i really have to work on my explaining skills…

  3. is it possible to have like an equalizer, but a really big one, with lets say 50Hz difference. And you would play a recording and u would click on button for 200-250Hz and it would instantly turn off all frequencies between 200 and 250Hz, while still recording it. It would be really easier to compare and to make the best choices. Because if you edit it and then listen to it…or even if you produce 2 recordings, one with certain frequencies cut and 1 not…and listen them one after another…its just not the same. I have this kind of equalizer on my creative sounds card program but it has way to few sliders. Basicly minimum would be 1 slider for every 100Hz, every 50Hz would be even better.

  4. audacity has noice removal where you get a certain pattern of silence and then it will remove all of the kind from the recording. Does it also have any recognition capacibilities ? Lets say human voice produces certain patern and you make the program remember it. And then it will show you all parts of the track where it found this patern ?

Thank you for your help, once again, what a great program audacity is. :slight_smile:)

7

I think that these pages answer question 1:
http://manual.audacityteam.org/o/man/high_pass_filter.html
http://manual.audacityteam.org/o/man/low_pass_filter.html

Sorry I don’t quite understand what you mean in question 2.
When talking about signals, 0 dB is the maximum, full scale level (the waveform is the full hight of the audio track).
Signals that are less than this have negative dB values, for example a signal that is half the hight of a track has a peak level of -6 dB.

Question 3.
Audacity does not do “real time” processing. It is not possible to process sound with Audacity during recording (or playback). Processing in Audacity is only done on tracks after recording has been stopped.
The Equalization effect allows much more precise filtering control that the high-pass and low-pass filters: http://manual.audacityteam.org/o/man/equalization.html

Question 4.
No, Audacity does not have that. Voice recognition is a very complex task and is only usually found in specialist voice analysis software.

8

Question 2: I think what you’re after is some kind of Limiter and/or Compressor effect.

–
Ragnar

9

Thank you for your replyies.

I dont know the correct terminology but the best i can describe point two is like this. Lets say that on the recording there is a speech and sound of a car. The sound of a car is twice louder then the speech, therefor you can hear it better then you can hear speech. I used Decibels to describe loudness because thats the only unit i know. And i just put into a comparison 30 and 50 dB. If you were to place a limit on how loud things on a recording can be…this would kind of make both speech and car sound of equal loudness. Soo if speech is between 0 and 30 dB and car is between 0 and 50dB…and i simply cut off everything over 30dB…in theory i will still hear both car and speech, but speech will sound louder then it did before, because all the car noice between 30-50dB will be gone and at the same time it wont cut off voice because voice is only between 0 and 30dB. A limiter, yes. But it is importaint that i know how to set it to make sure i dont do any changes to the speech part. Basicly i need some kind of normalization but instead of making silent parts louder ( which means amplify and lose quality ), i want to make loud parts quieter. :slight_smile:

10

Thanks audacityfanboy, I think I understand what you mean.
Unfortunately it won’t work - or at least, it won’t work if the car and the voice occur at the same time.

The problem is that where the car and the voice occur at the same time, the recording is not two sounds, it is not “a car sound” and “a voice sound” but is one sound that has been created by a car and a voice. The car and the voice have become inextricably mixed like adding milk to a cup of tea. You can reduce the amount of milk in the tea cup by pouring some of the tea away, but then there is less milk and less tea. You can’t just pour the milk away and leave black tea in the cup.

11

Ok soo basicly low pass and high pass is still the best option ? Because even though they are merged into 1 sound source, they still have different frequencies, for instance car is mostly under 400, voice is mostly over 400. Soo if i completely remove, or maybe just lower the sound of those frequencies, it should do the trick in a similiar way that i wanted, right ?

12

Yes, filtering may help.