Audio Recording / Editing Tutorials

[insufficient]

TUTORIAL 1 - RECORDING

and welcome to the first tutorial of this series. This first tutorial assumes you are using Microsoft Windows; if you are a Mac user, please comment and I will be happy to help.

This tutorial as it stands is not finished, but I thought it wise to share what I had written so far so suggestions as to improvements could be offered. If there is anything you do not understand, please let me know and I will endavour to correct it . Also if I have missed anything out or if you have any other questions, please reply.

Any modern-day photographer is able to perform amazing feats with their captured images, when sat in front of a computer. Ask them, however, about the art of their craft, and they will always claim it is 'in camera'. Every serious photographer realises that no amount of post-processing can make up for a poorly-lit or poorly-staged shot. Likewise, in sound recording, a clean original recording, and a high-quality digital capture of that recording, are both key to a good end result.

It would probably be best if I defined these two terms in context at this stage -

recording - The physical process of transferring sound onto some medium (cassette, mobile phone, etc.)
capture - The process of turning that recording into digital data that can be understood by a computer. For digital devices such as mobile phones, digital dictaphones and so on, this process happens simultaneously with recording.


We can divide sound recording/capture into roughly three categories:

1. Live recording of sound onto some medium eg. recording a lecture onto tape or mobile phone, dictaphone etc., while you are physically present at an event.
2. Analog to digital capture eg. transferring audio from a cassette or other analog source into your PC.
3. Digital capture or transcoding eg. recording a live web-stream of a lecture

I will address these three scenarios one-by-one .


Scenario 1 - Live capture

There are some fairly simple key principles to keep in mind. Sticking to as many of these as possible will help maximize your sound quality.

• If possible, try to get your recording device as close to thelecturer as possible. If there is a stand where the speaker places his notes, etc., ask if he minds you placing your device there.
• Failing that, place your device on a surface that you know will not move. If you are in a lecture theatre, then a desk may be suitable - make sure no-one is sitting at it and likely to rest their arms there, tap on it etc. If you are in a masjid, the carpet next to you is a perfect spot and will do a good job of muffling any noise present.
• DO NOT place your device next to the PA system speaker - recorders are usually designed to record human speech-level audio. If you can hear the lecturer clearly, then so can your device. Your recorder will 'clip' audio which is too loud, which basically means that upon playback you will not be able to hear the distinct sounds of human speech, if you are recording from a source which is too loud.
• This is probably the most important point - DO NOT FIDDLE with the microphone after recording has started. It may be tempting to do so, but the noise created from movement will ruin your recording. It is better to have a universally quiet recording than one which changes volume in the middle, or one with lots of movement noise. The former is easy to correct on computer, the latter two are not.

Of course, the above points deal with approaching an event as an outsider. If you are involved from the beginning, you may have a chance to record directly from the facility's sound system. This falls under the second category I described above, so will be explained later on in the tutorial


Scenario 2 - Analog to digital conversion

This category can be further divided in two -

• Capturing from a microphone.
• Capturing from a tape or other analog device.

The difference between the two is minor, but extremely important. Most computers have at least three audio 'jack' sockets, and they are colour-coded. For capture purposes, there are two which interest us -
PINK - Microphone. Always use this when plugging in a microphone. If you plug 'line-level audio' (see below) into this, it will be 'clipped' and unusable.
BLUE - Line-in. Use this to plug in your dictaphone, radio, or cassette player, etc. You can use a standard 'line cable' with a 3.5mm jack on either end to connect the 'line out' or 'headphone' output of your device to the 'line-in' on your PC. Do not plug a microphone into the line-in socket. Instead, use the pink microphone socket.

It is important to note that some modern laptops do not come with both a line-in and microphone socket, just one of the two. In this case the socket is 'auto-sensing', which means you can plug any device in and the socket will configure itself to match.

The next step is to set the correct sound level for the device you are using. to do this, double-click on the speaker icon at the bottom-left of your screen. Click ???->????. Enable 'mic boost' if you are using a microphone, and mute the inputs which you are not using.

Leave the volume controls open and load up the recording software of your choice (Windows' built-in recorder will do the job well enough - just go to the start-->???-->????-->???). Start playing your line-in device or speak into your microphone for a few seconds. Stop the recording and then play it back. Adjust the volume and repeat the process until you seem to be recording at a reasonable/comfortable volume.


Scenario 3 - Digital capture

This is the easiest of all the capture methods. However, there re a few extra steps involved if you are using Windows Vista or Windows 7.

The idea is that, instead of capturing from a physical socket on your PC, you will record from the 'stereo mix' - in effect, the sound which gets output through your speakers. To this effect, you will need to close down any applications (eg. MSN messenger) which are likely to make any noise.

Next go into


--------------------------------------------------------------------------------

I will, keep adding to this tutorial until it is finished. I am not sat at a Windows-based PC at the moment, so am unable to take the screenshots which will explain settings for recording.

If there is some information which someone wishes to know, please reply and I will give it priority.

[insufficient]

***mod please delete***

[insufficient]

***mod please delete***

[insufficient]

I will be posting some tutorials soon. I apologize for the roughness of what is about to follow - if you have any questions, corrections or notice missing gaps, please reply and I will endeavour to fill in the gaps to make the tutorial into something useful for our purposes inshaAllah.

[insufficient]

Part 2 - EDITING

2.1 - PREREQUISITES FOR EDITING

Audacity http://audacity.sourceforge.net/download/

Download from the link above. This is the main audio-editing application. You will need to download version 1.3.12 beta or (if you're feeling adventurous) the latest nightly build.

LAME http://audacity.sourceforge.net/help...all&i=lame-mp3
Installation instructions found at link above. This is an mp3 encoding library, you will need it to save mp3s.

FFMPEG http://manual.audacityteam.org/index...#installffmpeg
Installation instructions found at link above. This is an import/export library, you will need it to import most common audio formats into Audacity

Plugins
The following plugins will not be used in this tutorial. However, they are sometimes very useful. They may form the subject of the next tutorial.
Compress Dynamics http://pdf23ds.net/software/dynamic-compressor/
Noise Gate http://forum.audacityteam.org/viewto...=30996&start=0

[insufficient]

2.2 - A QUICK INTRODUCTION TO AUDACITY
...Pending...

2.3 - THINGS TO BE AWARE OF

• The human vocal range lies approximately between 300Hz and 3000Hz.
• You can use 'spacebar' instead of clicking the play button in Audacity. Playback will begin from the position of the cursor. you can set the position of the cursor by clicking anywhere on the track.
• Clicking-and-dragging on the track sets a selection. Any filters you apply will then take effect only on this selection. You can click and drag the edges of a selection to fine-tune its boundaries. If you hit 'play' with some audio selected, only that selection will play.
• You can zoom out either by using the relevant toolbar button or holding the ctrl key on your keyboard and scrolling your mousewheel downwards. You can zoom in by scrolling up; the zoom will be in the derection of your mouse cursor.
• I am assuming you input files are mono. If they are stereo, use 'split stereo to mono' in the dropdown to the left of the sound wave visualization, and then delete one of the resulting tracks by clicking the 'x' just above that dropdown.
• You can insert ‘splits’ by hitting CTRL+i, they show up as back lines and get snapped-to when selecting. Very useful...

[insufficient]

2.4 - A BASIC AUDIO-EDITING WORKFLOW

1. Open your source audio file in Audacity
2. Zoom out until the whole track is visible
3. Turn on 'show clips' (Check 'show clipping' in the view menu). If you see some red lines appear, then continue to step 4. If not, skip foward to step 9.
4. The red lines indicate areas of 'clipping'. In simple terms, this means that the audio recorded was of too high a volume to be picked up by the microphone. In technical terms, the amplitude of the sound wave exceeded the threshold of the microphone. You can see this by zooming in on one of the red lines until the curve of the sound wave is clearly visible. You will see a smooth wave rising up until it hits the top of the track and this is 'clipping'. Thankfully, it is usually easy to fix -
5. Amplify (Effects > Amplify) the whole track by -10dB to provide room for the reconstruction
6. Now select some (no more than a few minutes) of the previously clipped audio - if the clipping is quite prominent, you will have to do this in sections.
7. Clip fix (Effects > Clip Fix) with the default settings. Zoom in onto the section of the wave that was clipped. If it now approximates a smooth curve like you would expect, then move on to the next section. If not, then edit > undo and re-do the 'clip fix' this time lowering the 'threshold' percentage. Rinse and repeat until you hit the right balance.
8. Normalize the track so expand it to the maximum amplitude range. Do not worry about any red lies that re-appear; they are (not quite) clips.
9. Now, we need to remove subsonic noise. This is audio that humans can't hear because it is of too low a frequency, and it can wreak havoc with our editing. It is easily removed. Use a High-Pass Filter (From the 'effects' menu) with a frequency of 20Hz and attenuation of 6dB.
10. Next we will remove background noise. Search through the track to find a place where there is a good few seconds of it. Select a few seconds of it (preferably about 5s although less will do) and play it back to make sure it is only background noise and that you have not inadvertently selected someone moving chairs, etc.
11. Go to Effects > Noise Removal. Click 'get noise profile'. You will be dropped back into the main window. Now select a few seconds of speech and go again into Effect > Noise Removal. Click 'preview' and listen to the audio. If there is too much noise, increase the 'noise reduction' value and listen again. Keep fine-tuning until you find the sweet-spot. too much noise removal will cause voices to sound distorted. It is always better to leave in a small amount of noise to help the recording sound natural. Once you are done, cancel the dialogue, select all of your audio and then re-enter the Noise Removal dialogue. Re-enter the number you settled on and click 'OK' to process.
12. Now, we will increase the clarity of our speaker's voice by attenuating non-speech frequencies of sound. So run a high-pass filter with a frequency of 300Hz followed by a low-pass filter with a frequency of 3000Hz, leaving all other settings at their defaults.
13. Now we can manually remove any remaining noise. Simply select the offending segments of audio, and hit the 'silence' button the toolbar to eliminate them. Make sure they don't coincide with the speech; if they do, leave them be, there is little you can do, and that little will be explained in the next tutorial inshaAllah.
14. You can also delete silence from the beginning and end of the track by selecting it and hitting the 'delete' key on your keyboard.
15. Normalize (Effects > Normalize) to increase the volume of your audio to max. 0dB
16. Now we are ready to export. Go to file > export. In the dropdown box select 'mp3 files', and then click 'options'. Select a variable bit-rate mode at Quality 9, 45-85 kbps and variable speed fast. Click 'OK' and then 'save'. The next dialog will allow you to enter metadata, if you want.

[insufficient]

Part 3 - GOING FURTHER

3.1 - INCREASING 'VOLUME'

The best way to increase volume is to normalize. This will automatically increase the max. amplitude (ie. volume) to 0dB. If that doesn't seem to work, then do it in sections - ie. split* the track into 3, normalize each section. You will notice which general areas are quieter; split off and normalize those. Keep going until you are happy. This will ensure nice even volume throughout the duration of the track and perfectly normalized* (which basically means 'amplified to maximum possible volume') audio.

Another way to increase audio volume while avoiding clipping is to use the 'Compress Dynamics' filter mentioned in the prerequisites - http://pdf23ds.net/software/dynamic-compressor/

This basically automates the splitting+normalizing process explained above, but on a much more detailed level. It takes a long time to process, so use it on a small sample of audio until you are happy with the results and then edit>undo and apply it to the whole track. I find a compression value somewhere between 0.7 and 0.9 works best, along with attack/decay values of between 250ms and 400ms. Most the other options are usually best left at default. Of course you will also have to fine-tune the 'noise threshold' to match your recording - this specifies a volume below which sound will not be amplified. Do not enable the 'noise gate falloff', either, as it will completely 'deaden' silent parts of the recording and make the speech sound really unnatural (because that will still have some background noise).

Bear in mind that the filter only judges background noise by sound level, so on some recordings you will hear much more background noise in the sections that have been amplified by the filter, so I recommend using the noise removal filter after compressing. You only need ~1.5s of audio to use as a noise sample for noise removal.

This filter has a tendency to make the sound 'too' even and unnatural so I have stopped using it, in favour of the manual method I described earlier. But it does work well when used in moderation.

[jav12345]

Part 3 - GOING FURTHER

3.1 - INCREASING 'VOLUME'

The best way to increase volume is to normalize. This will automatically increase the max. amplitude (ie. volume) to 0dB. If that doesn't seem to work, then do it in sections - ie. split* the track into 3, normalize each section. You will notice which general areas are quieter; split off and normalize those. Keep going until you are happy. This will ensure nice even volume throughout the duration of the track and perfectly normalized* (which basically means 'amplified to maximum possible volume') audio.

Another way to increase audio volume while avoiding clipping is to use the 'Compress Dynamics' filter mentioned in the prerequisites - http://pdf23ds.net/software/dynamic-compressor/

This basically automates the splitting+normalizing process explained above, but on a much more detailed level. It takes a long time to process, so use it on a small sample of audio until you are happy with the results and then edit>undo and apply it to the whole track. I find a compression value somewhere between 0.7 and 0.9 works best, along with attack/decay values of between 250ms and 400ms. Most the other options are usually best left at default. Of course you will also have to fine-tune the 'noise threshold' to match your recording - this specifies a volume below which sound will not be amplified. Do not enable the 'noise gate falloff', either, as it will completely 'deaden' silent parts of the recording and make the speech sound really unnatural (because that will still have some background noise).

Bear in mind that the filter only judges background noise by sound level, so on some recordings you will hear much more background noise in the sections that have been amplified by the filter, so I recommend using the noise removal filter after compressing. You only need ~1.5s of audio to use as a noise sample for noise removal.

This filter has a tendency to make the sound 'too' even and unnatural so I have stopped using it, in favour of the manual method I described earlier. But it does work well when used in moderation.

Assalamu alaykum,

I was trying to correct the clipping in an audio file I need to edit. I have downloaded Audacity but cannot find the option for viewing clipping under View in the toolbar, nor is there an option for fixing clipping under Effects. Could you help me out please?

[insufficient]

Assalamu alaykum,

I was trying to correct the clipping in an audio file I need to edit. I have downloaded Audacity but cannot find the option for viewing clipping under View in the toolbar, nor is there an option for fixing clipping under Effects. Could you help me out please?

Mufti saheb,

Did you download the Beta (1.3.12) version? Please go to Help > About Audacity and if it does not read 'Audacity 1.3.12 ...', then uninstall through Control Panel and install from the link above.

If you already have version 1.3.12 then please reply and I shall investigate further.

Do not hesitate to ask if I can be of further help.