Editing Audio: Difference between revisions

From Librivox wiki
Jump to navigationJump to search
Content deleted Content added
Jc (talk | contribs)
LordOider (talk | contribs)
→‎Tips & Tricks: added cross references
 
(6 intermediate revisions by 5 users not shown)
Line 1: Line 1:
How to, why to, tips & tricks

== Why To ==
See [[Audio Processing Concepts]]

== How To ==
You will make mistakes in your reading, there is no doubt about that. Don't worry about the mistakes too much as you're recording, because they can be edited out. You can also make adjustments to volume, and remove small amounts of background noise in the editing process.
You will make mistakes in your reading, there is no doubt about that. Don't worry about the mistakes too much as you're recording, because they can be edited out. You can also make adjustments to volume, and remove small amounts of background noise in the editing process.


Line 10: Line 4:
There are a few things that you can do while you're recording to make the editing process easier later on.
There are a few things that you can do while you're recording to make the editing process easier later on.


# As soon as you make a mistake tap the mic three times, or say a really loud "MISTAKE!". Then repeat the sentence that you goofed up. The point is to create a spike in the sound waves of your editing software. This allows you to more easily find your mistakes in the editing process.
# As soon as you make a mistake clap your hands three times, or say a really loud "MISTAKE!". Then repeat the sentence that you goofed up. The point is to create a spike in the sound waves of your editing software. This allows you to more easily find your mistakes in the editing process. [Some people with delicate mics, just click their tongue twice; this also gives nice easily seen spikes]

# Make a test recording to be sure your mic and editing software is set at the correct volume before recording the text. It is possible to alter the volume later, but raising the volume too much will also amplify the background noise.
# Make a test recording to be sure your mic and editing software is set at the correct volume before recording the text. It is possible to alter the volume later, but raising the volume too much will also amplify the background noise.

# Don't stop the recording for anything. If you need to get a drink of water, let it record through that. Need to clear your throat? Keep recording so you end up with one long track. This will prevent inconsistencies in sound or volume. If you stop recording and return later, it will probably be obvious to the listener that you stopped and started again later.
# Don't stop the recording for anything. If you need to get a drink of water, let it record through that. Need to clear your throat? Keep recording so you end up with one long track. This will prevent inconsistencies in sound or volume. If you stop recording and return later, it will probably be obvious to the listener that you stopped and started again later.


Line 22: Line 14:


==== If you created a sound spike after your mistakes. ====
==== If you created a sound spike after your mistakes. ====
:<i>Main Page: [[Deleting errors in Audacity]] and [[Alternative Editing Method]]</i>
:<i>Main Page: [[Deleting Errors in Audacity]] and [[Alternative Editing Method]]</i>
It is possible to edit your file without listening to the whole thing to find the mistakes; you can just search for the editing spikes. When you find a spike, listen through to the recording around it, and delete the mistake including the spike. '''Make sure you find someone to proof-listen your recording if you don't listen through it yourself during editing'''
It is possible to edit your file without listening to the whole thing to find the mistakes; you can just search for the editing spikes. When you find a spike, listen through to the recording around it, and delete the mistake including the spike. '''Make sure you find someone to proof-listen your recording if you don't listen through it yourself during editing'''


Line 34: Line 26:
=== Cleaning out noises during silences ===
=== Cleaning out noises during silences ===
Between sentences and paragraph, it is better if the silence is total. However, there are often small noises like mouth/lip noises, breathing or blowing into the microphone, chair creaking etc. all the more so as the reader takes the ooportunity of this silence to sit back, relax, breathe, or whatever. The basic command in audacity is '''Ctrl L''': select the silence, then hit '''Ctrl L''' to turn it into true silence. You can leave some noise, like the short breath right before the new sentence starts, so that we know this is no machine-generated audio-book.
Between sentences and paragraph, it is better if the silence is total. However, there are often small noises like mouth/lip noises, breathing or blowing into the microphone, chair creaking etc. all the more so as the reader takes the ooportunity of this silence to sit back, relax, breathe, or whatever. The basic command in audacity is '''Ctrl L''': select the silence, then hit '''Ctrl L''' to turn it into true silence. You can leave some noise, like the short breath right before the new sentence starts, so that we know this is no machine-generated audio-book.

For removing general background noise, see: [[Noisecleaning With Pics]]


=== Trimming silences to a consistent value ===
=== Trimming silences to a consistent value ===
You may want to make sure silences within paragraphs all have roughly the same duration (e.g. all between 0.5s and 1s), and all silences between paragraphs another duration (e.g. all between 1.5s and 3s). Apparently, Audacity does not have an automated command to trim all silences at the same length. The basic solution is to select the amount of silence you want removed, and cut it '''Ctrl X'''. Conversely, to increase the duration of a silence, select the amount of silence you need to add, hit '''Ctrl C''' (copy), then '''Ctrl V''' (paste) twice (one in place, then duplicate).
You may want to make sure silences within paragraphs all have roughly the same duration (e.g. all between 0.5s and 1s), and all silences between paragraphs another duration (e.g. all between 1.5s and 3s). Apparently, Audacity does not have an automated command to trim all silences at the same length. The basic solution is to select the amount of silence you want removed, and cut it '''Ctrl X'''. Conversely, to increase the duration of a silence, select the amount of silence you need to add, hit '''Ctrl C''' (copy), position the cursor where you want to add silence, then hit '''Ctrl V''' (paste).


Anyone has something quicker ?
Anyone has something quicker ?
Line 46: Line 40:


For the perfectionist: the selection should start and end with a zero value and also with a rising signal, and the cursor should also be placed on a zero with a rising signal. With this technique, you can now edit out (or in) not only sentences or words, but individual syllables. Try shortening an 'aaaah', or stretching a 'oh'.
For the perfectionist: the selection should start and end with a zero value and also with a rising signal, and the cursor should also be placed on a zero with a rising signal. With this technique, you can now edit out (or in) not only sentences or words, but individual syllables. Try shortening an 'aaaah', or stretching a 'oh'.

See also: [[Removing Peaks]]


=== Joining several MP3 files into one ===
=== Joining several MP3 files into one ===
If you have several MP3 files that you wish to convert into one, long file (eg if you wish to combine a set of chapter files into a single book file), there are a couple of ways that this can be done.
If you have several MP3 files that you wish to convert into one, long file (eg if you wish to combine a set of chapter files into a single book file), there are a couple of ways that this can be done.


If you are using Linux or a Mac with OS X, you can use the {{{cat}}} command from a terminal. To join the contents of files {{{1.mp3}}}, {{{2.mp3}}}, and {{{3.mp3}}} into one file named {{{out.mp3}}}:
If you are using Linux or a Mac with OS X, you can use the <b>cat</b> command from a terminal. To join the contents of files <b>1.mp3</b>, <b>2.mp3</b>, and <b>3.mp3</b> into one file named <b>out.mp3</b>:


cat 1.mp3 2.mp3 3.mp3 > out.mp3
{{{

cat 1.mp3 2.mp3 3.mp3 > out.mp3
Since OS X is built on top of a Unix kernel, the ''cat'' command certainly is available from the OS X command line as well.
}}}


Alternatively, you could use the [http://mp3wrap.sourceforge.net/ MP3Wrap] program. It's released under the GPL and is free to download. Versions are available for Linux and Windows.
Alternatively, you could use the [http://mp3wrap.sourceforge.net/ MP3Wrap] program. It's released under the GPL and is free to download. Versions are available for Linux and Windows.


To overlay tracks, see: [[Merging Tracks]]
''Does anyone know of an equivalent for the Mac?''

Since OS X is built on top of a Unix kernel, the ''cat'' command certainly is available from the OS X command line as well.


=== Splitting an MP3 file into shorter files ===
=== Splitting an MP3 file into shorter files ===
:''Main page: [[How To Split With Mp3Splt]]''
If you need to split MP3 files into shorter recordings, the [http://mp3splt.sourceforge.net/ Mp3splt] program should work. It's released under the GPL and is free to download. Versions are available for Linux, Mac and Windows, and there are command-line and GUI (graphical user interface - point & click) versions available.
If you need to split MP3 files into shorter recordings, the [http://mp3splt.sourceforge.net/ Mp3splt] program should work. It's released under the GPL and is free to download. Versions are available for Linux, Mac and Windows, and there are command-line and GUI (graphical user interface - point & click) versions available.



Latest revision as of 23:45, 30 October 2013

You will make mistakes in your reading, there is no doubt about that. Don't worry about the mistakes too much as you're recording, because they can be edited out. You can also make adjustments to volume, and remove small amounts of background noise in the editing process.

While Recording

There are a few things that you can do while you're recording to make the editing process easier later on.

  1. As soon as you make a mistake clap your hands three times, or say a really loud "MISTAKE!". Then repeat the sentence that you goofed up. The point is to create a spike in the sound waves of your editing software. This allows you to more easily find your mistakes in the editing process. [Some people with delicate mics, just click their tongue twice; this also gives nice easily seen spikes]
  2. Make a test recording to be sure your mic and editing software is set at the correct volume before recording the text. It is possible to alter the volume later, but raising the volume too much will also amplify the background noise.
  3. Don't stop the recording for anything. If you need to get a drink of water, let it record through that. Need to clear your throat? Keep recording so you end up with one long track. This will prevent inconsistencies in sound or volume. If you stop recording and return later, it will probably be obvious to the listener that you stopped and started again later.

Editing

The process of editing simply involves removing the pieces of your recording that shouldn't be there. Read the instructions for the recording software that you're using. It should tell you how to remove sections of a track, and edit out pieces.

Try to avoid too much processing through filters on your recording. Excessive alterations to the sound can sometimes cause additions like a metallic quality to your recording. A little background noise is OK. If you're unsure how much is too much, ask a fellow volunteer to lend their ear. See [:ImproveYourRecording:How to Improve Your Recording] for more tips.

If you created a sound spike after your mistakes.

Main Page: Deleting Errors in Audacity and Alternative Editing Method

It is possible to edit your file without listening to the whole thing to find the mistakes; you can just search for the editing spikes. When you find a spike, listen through to the recording around it, and delete the mistake including the spike. Make sure you find someone to proof-listen your recording if you don't listen through it yourself during editing

Common Keyboard Shortcuts For Audacity

Ctrl S = save Ctrl C = copy Ctrl V = paste Ctrl Z = undo Ctrl N = new Ctrl A = select all (selects the whole file) Ctrl O = open

Ctrl W = close Space bar = play and pause

Tips & Tricks

Cleaning out noises during silences

Between sentences and paragraph, it is better if the silence is total. However, there are often small noises like mouth/lip noises, breathing or blowing into the microphone, chair creaking etc. all the more so as the reader takes the ooportunity of this silence to sit back, relax, breathe, or whatever. The basic command in audacity is Ctrl L: select the silence, then hit Ctrl L to turn it into true silence. You can leave some noise, like the short breath right before the new sentence starts, so that we know this is no machine-generated audio-book.

For removing general background noise, see: Noisecleaning With Pics

Trimming silences to a consistent value

You may want to make sure silences within paragraphs all have roughly the same duration (e.g. all between 0.5s and 1s), and all silences between paragraphs another duration (e.g. all between 1.5s and 3s). Apparently, Audacity does not have an automated command to trim all silences at the same length. The basic solution is to select the amount of silence you want removed, and cut it Ctrl X. Conversely, to increase the duration of a silence, select the amount of silence you need to add, hit Ctrl C (copy), position the cursor where you want to add silence, then hit Ctrl V (paste).

Anyone has something quicker ?

Making sure the edits are transparent

Sometimes, the 'edges' of an edited portion make a distinctive 'clicking' sound. When you zoom in, you see that there is a discontinuity in the signal (a sharp edge). To remove this, you need to cut (or add) signal so that the transition is smooth. The easiest way to do it is to start and end all your selections (when you want to cut or when you want to paste) at samples where the signal has a zero value, and place the cursor (in case you want to paste) at another sample having a zero value. That way, you ensure that the signal will have a zero value on both sides, thus ensuring continuity. Zero values are not hard to find: there are several hundreds of them each second.

In Audacity, Edit > Find Zero Crossings, or the Z key, moves the edges of the selection slightly until they are close to a zero crossing.

For the perfectionist: the selection should start and end with a zero value and also with a rising signal, and the cursor should also be placed on a zero with a rising signal. With this technique, you can now edit out (or in) not only sentences or words, but individual syllables. Try shortening an 'aaaah', or stretching a 'oh'.

See also: Removing Peaks

Joining several MP3 files into one

If you have several MP3 files that you wish to convert into one, long file (eg if you wish to combine a set of chapter files into a single book file), there are a couple of ways that this can be done.

If you are using Linux or a Mac with OS X, you can use the cat command from a terminal. To join the contents of files 1.mp3, 2.mp3, and 3.mp3 into one file named out.mp3:

cat 1.mp3 2.mp3 3.mp3 > out.mp3

Since OS X is built on top of a Unix kernel, the cat command certainly is available from the OS X command line as well.

Alternatively, you could use the MP3Wrap program. It's released under the GPL and is free to download. Versions are available for Linux and Windows.

To overlay tracks, see: Merging Tracks

Splitting an MP3 file into shorter files

Main page: How To Split With Mp3Splt

If you need to split MP3 files into shorter recordings, the Mp3splt program should work. It's released under the GPL and is free to download. Versions are available for Linux, Mac and Windows, and there are command-line and GUI (graphical user interface - point & click) versions available.

Using a Drawing Tablet

Using a mouse for extended periods of time to edit can be painful and uncomfortable. If you have a drawing tablet, consider using it for editing. Set any function buttons to the keyboard shortcuts for cutting sections of audio, and saving. This saves time, and discomfort.

Example: Kristin uses Audacity and a Wacom drawing tablet. There are two buttons on the pen that she has set to CTRL+X (cut in Audacity), and SPACE (play/pause in Audacity). One of the function buttons on her drawing tablet is set to CTRL+S (save). Deleting a piece of audio is a simple drag of the pen to highlight, and click of a button on the pen. She then hits the second button to resume play, listening for more mistakes.