Simple audio compression in Audacity
In principle, a recorded voice can immediately be posted as a podcast. But it is better to sit for a while, figure out at least the basics and carry out simple sound processing. There is one minimal action, which is not at all difficult either in understanding or in implementation, and the structure of the sound will change dramatically for the better. This is a sound compression.
When I first heard about compression, I certainly thought about mp3 or ogg, or any other formats that allow compressing audio files in size. As it turned out, compression is not at all here :)
To understand what compression is, let's start with the practice right away and in the process we will get to the core of the matter. Here is a fragment of my speech without any processing (from some release of the first season of "Frames"):
')
Download file
And this is how this sound looks graphically in the Audacity editor:
At first glance, the sound is like sound. But everything is relative. This fragment has one drawback - the sound "jumps". According to the schedule, you can easily see how my voice becomes quieter, louder, while moving around the chair I changed the distance to the microphone during a conversation, and besides, it whispered, then actively proved something, raising my voice and tone. As a result, if you listen to this recording, the sound will then hit the ears, then it will be heard too quietly.
The task of compression is to “compress” the sound, average it, make it denser, and reduce the distance between the quietest moments and the loudest. That's all! That's all the compressor does!
He does this, of course, not as he pleases, but as we indicate to him. Here it seems to begin as complex terms and misunderstanding, but in reality everything is easier than ever.
Let's select the entire fragment (Ctrl + A, from English “a ll” - “all”) and go to Effects-> Compressor (Effects-> Compressor). In the screenshot below, I set all values ​​to maximum, just to show what this window looks like.
Here are three sliders.
The first is a “threshold” (threshold value, threshold). Anything above the threshold will sound quieter after compression. Roughly speaking, the threshold is the value at which the compressor is triggered, a certain boiling point.
The second is the "ratio". This value is set as a ratio of number to number. The maximum value in the screenshot is 10: 1. This means that as soon as the signal reaches the threshold, the volume will decrease 10 times. This is very simple - for every 10 decibels of the input signal, the compressor will miss only 1, as a result, the volume will decrease.
The third is the attack time. This is the time it takes for the compressor to respond to an increase in signal. In the example above, the value is 1.0 second. Here's how it works: the signal reaches the threshold value, the compressor thinks for 1 second, if the signal remains in this zone (does not decrease below the threshold), then the volume decreases. If a gun shot sounds in the recording, then at 1 second its volume will not even change - the shot is too fast, the compressor “will not have time to react.” In more advanced editors the compressor has more settings, at least there is also a release time - recovery time This is the same thing, only from the other end: how much time does the compressor take to stop turning down the volume.
The checkbox “Normalize to 0 dB after compression” speaks for itself. After compression, the overall level of the signal will usually decrease, the sound will become even denser, but quieter. With this check mark, the sound will become louder, reach a level of 0 dB, which is the standard for podcasting.
This is the whole mechanism of sound compression! Let's try to practice this knowledge. Here is what I do:
After a short process, we will see the following picture:
And the sound is such:
Download file
Feel the difference? Even if you say no, this is a hasty conclusion. Listen to the first 6 seconds of both recordings and you will understand that it is more pleasant to listen to the compressed sound, you do not need to listen to “what did he mumble about ??! ..”. And if you listen to the podcast for an hour and a half (O, Frames! ..), then your ears and brain are tired of poorly compressed or no compression at all. Good sound quality is not distracting, relaxing.
It is necessary to understand that the action described in the article: firstly, not the best . The same piece of sound can be compressed much better. Secondly, it is not universal for all records: in your particular case, other values ​​may be needed.
...
As you can see, it is not difficult at all, but the result affects a lot! At least out of respect for the audience, they should give a good sound. Pleasant.
I myself became acquainted with compression through the podcast " Theory and Practice of Sound Recording " from the notorious Umputun . What I advise you. He can also read small articles about compression:
Next time, let's take on other aspects of sound processing.
Original on Podkastim.ru
When I first heard about compression, I certainly thought about mp3 or ogg, or any other formats that allow compressing audio files in size. As it turned out, compression is not at all here :)
To understand what compression is, let's start with the practice right away and in the process we will get to the core of the matter. Here is a fragment of my speech without any processing (from some release of the first season of "Frames"):
')
Download file
And this is how this sound looks graphically in the Audacity editor:
At first glance, the sound is like sound. But everything is relative. This fragment has one drawback - the sound "jumps". According to the schedule, you can easily see how my voice becomes quieter, louder, while moving around the chair I changed the distance to the microphone during a conversation, and besides, it whispered, then actively proved something, raising my voice and tone. As a result, if you listen to this recording, the sound will then hit the ears, then it will be heard too quietly.
The task of compression is to “compress” the sound, average it, make it denser, and reduce the distance between the quietest moments and the loudest. That's all! That's all the compressor does!
He does this, of course, not as he pleases, but as we indicate to him. Here it seems to begin as complex terms and misunderstanding, but in reality everything is easier than ever.
Let's select the entire fragment (Ctrl + A, from English “a ll” - “all”) and go to Effects-> Compressor (Effects-> Compressor). In the screenshot below, I set all values ​​to maximum, just to show what this window looks like.
Here are three sliders.
The first is a “threshold” (threshold value, threshold). Anything above the threshold will sound quieter after compression. Roughly speaking, the threshold is the value at which the compressor is triggered, a certain boiling point.
The second is the "ratio". This value is set as a ratio of number to number. The maximum value in the screenshot is 10: 1. This means that as soon as the signal reaches the threshold, the volume will decrease 10 times. This is very simple - for every 10 decibels of the input signal, the compressor will miss only 1, as a result, the volume will decrease.
The third is the attack time. This is the time it takes for the compressor to respond to an increase in signal. In the example above, the value is 1.0 second. Here's how it works: the signal reaches the threshold value, the compressor thinks for 1 second, if the signal remains in this zone (does not decrease below the threshold), then the volume decreases. If a gun shot sounds in the recording, then at 1 second its volume will not even change - the shot is too fast, the compressor “will not have time to react.” In more advanced editors the compressor has more settings, at least there is also a release time - recovery time This is the same thing, only from the other end: how much time does the compressor take to stop turning down the volume.
The checkbox “Normalize to 0 dB after compression” speaks for itself. After compression, the overall level of the signal will usually decrease, the sound will become even denser, but quieter. With this check mark, the sound will become louder, reach a level of 0 dB, which is the standard for podcasting.
This is the whole mechanism of sound compression! Let's try to practice this knowledge. Here is what I do:
- Threshold value set to -25 dB
- The coefficient value is 6: 1
- The attack time is 0.1 seconds.
- I put a tick "Normalize to 0 dB after compression"
- ... and click OK :)
After a short process, we will see the following picture:
And the sound is such:
Download file
Feel the difference? Even if you say no, this is a hasty conclusion. Listen to the first 6 seconds of both recordings and you will understand that it is more pleasant to listen to the compressed sound, you do not need to listen to “what did he mumble about ??! ..”. And if you listen to the podcast for an hour and a half (O, Frames! ..), then your ears and brain are tired of poorly compressed or no compression at all. Good sound quality is not distracting, relaxing.
It is necessary to understand that the action described in the article: firstly, not the best . The same piece of sound can be compressed much better. Secondly, it is not universal for all records: in your particular case, other values ​​may be needed.
...
As you can see, it is not difficult at all, but the result affects a lot! At least out of respect for the audience, they should give a good sound. Pleasant.
I myself became acquainted with compression through the podcast " Theory and Practice of Sound Recording " from the notorious Umputun . What I advise you. He can also read small articles about compression:
Next time, let's take on other aspects of sound processing.
Original on Podkastim.ru
Source: https://habr.com/ru/post/26500/
All Articles