How I conduct live broadcasts on Internet radio stations
In order to conduct live broadcast it seems not so much and it is necessary: adjust the sound so that the music simultaneously plays and recording from the microphone is good, if Skype works well. In practice, it is not so simple. In the article I will talk about the settings of my workstation and the software used during the broadcast on the Internet radio.
The described setup is quite suitable for conducting musical ethers, moreover, for them it was originally made.
')
The overall architecture is as follows: even while conducting paired airs, all the finished sound is formed on the side of one of the hosts, then this stream is transmitted to the server, which distributes it to listeners (icecast2 is involved in this all). In parallel, by the way, inclusions are recorded for subsequent automated processing, but first and foremost.
I use a Samson C01U USB microphone, this one is right here:
It gives quite acceptable sound for such cases, although I think any USB microphone or headset will work. The main problem is that I need to hear the player and what the second presenter says and I should not hear myself. USB microphones work with a delay, it is very knocking and only interferes. Specifically, this microphone does not have an output for monitoring, but even if it were, it would not change anything: I need to hear the music / co-host.
The minimum requirements are as follows:
1) hear everything except yourself in the headphones;
2) give the full stream (including yourself) on the air.
That is, it is necessary to form two different sound streams, and this will require two sound cards (hardware or virtual). I just had two sound cards, let's call the cards snd1 and snd2. Both have two outputs, front and back, which is just as important. On one sound card (snd1) I form that sound that I want to hear myself and from one output (snd1 / front) I shoot it into the headphones, and from the second (snd1 / back) I send it to the line input of the second card.
The sound transmission from snd1 to snd2 is made with a short cable. Since the amplified linear signal is transmitted, this does not affect the sound quality too much. On the second card (snd2), monitoring of the line input is enabled, plus I add to it what I say into the microphone, and, thus, a ready sound is obtained for the broadcast. In order to capture sound from snd2 for recording or transmitting digitally, it must have the “what you hear” recording mode selected on it. Previously, such a recording mode was rarely met, now, it seems to be easier with this ... in general, snd2 has such a mode, otherwise it will not work.
Thus, the sound that I send to the server, that is, a ready-made stream containing: me, the second presenter, and jingles / interviews / music, I take from the exit called “what you hear” of the snd2 card.
So at my workstation, I use:
- foobar2000: music player;
- Skype: connection with the second presenter, guests, phone calls;
- VST Host: realtime effects to the microphone;
- Audacity: record and monitor stream;
- Edcast Standalone: translation of the generated stream to the server.
By the way, if you use not hardware sound cards, but virtual ones - a set of software and in general the picture will not change much.
Then a little more about each of the programs used.
foobar2000
The choice is justified only by the fact that it is familiar to me easily customizable player. During the broadcast, it is important that it is convenient to start / brake the music, control the volume. In fact, I launch two copies of FUBAR, two players are needed to make smooth transitions between songs or play some jingles against the background of sounding music. I know that there is some kind of specialized DJ software for this whole thing, but I am doing quite well just with two fubars.
Skype
Well, everything is clear, except for one thing: new versions of Skype are played only on the front output of the sound card. It's just a little bit how they came up with a great idea, only because of this you have to use the old versions. At one time I used the hardware fork, but this method has its drawbacks. In short, it turns out too quiet and there is no possibility to separately control the volume of the sound going to the headphones and on the snd2. And it’s very convenient: everyone has it, you can hold conferences, invite guests on the air, call reporters on assignments on mobile phones, it has happened in our practice and such.
VST Host
In the chain appeared almost by accident. The fact is that this particular microphone model has some strange approach to the sound: it sounds only in the right channel. In order to fix it on the fly and a vst host was added, such things are done in it quite easily. In addition, you can turn on, for example, a compressor or some other effects. At the moment I’m just fixing them for the lack of my microphone. Another function that it performs is the transmission of sound from a microphone to snd2. This could have been done in some other way, but since I use this host, why not in it. Accordingly, in it I turn off the microphone when it is required (usually in pauses, while the music is playing).
Audacity
In addition to recording, I also look at the wave being formed to monitor the ratio of the volume levels of myself / music / guests. The record I keep in my home is just a backup, the master record is kept by the radio station server itself. If the sound after the ether is processed by me, then this is done just in Audacity: normalization, compression, noise removal.
Edcast standalone
Transmits the generated audio stream to the server, where it is received by icecast2, which is running on our server side. SB Live! - this is the same snd2.
That's probably all on the part of the workstation. This option is not a dogma at all, rather an illustration of how the problem can be solved with the available means. Now I think how to simplify the circuit using a normal hardware mixer, or vice versa, switching to virtual sound cards.
Bonus - a screenshot at the time of the broadcast. On the next desktop, shunots and chat are usually running to communicate with listeners.
Clearly you can ask questions, suddenly forgot to write about something.
The described setup is quite suitable for conducting musical ethers, moreover, for them it was originally made.
')
The overall architecture is as follows: even while conducting paired airs, all the finished sound is formed on the side of one of the hosts, then this stream is transmitted to the server, which distributes it to listeners (icecast2 is involved in this all). In parallel, by the way, inclusions are recorded for subsequent automated processing, but first and foremost.
I use a Samson C01U USB microphone, this one is right here:
It gives quite acceptable sound for such cases, although I think any USB microphone or headset will work. The main problem is that I need to hear the player and what the second presenter says and I should not hear myself. USB microphones work with a delay, it is very knocking and only interferes. Specifically, this microphone does not have an output for monitoring, but even if it were, it would not change anything: I need to hear the music / co-host.
The minimum requirements are as follows:
1) hear everything except yourself in the headphones;
2) give the full stream (including yourself) on the air.
That is, it is necessary to form two different sound streams, and this will require two sound cards (hardware or virtual). I just had two sound cards, let's call the cards snd1 and snd2. Both have two outputs, front and back, which is just as important. On one sound card (snd1) I form that sound that I want to hear myself and from one output (snd1 / front) I shoot it into the headphones, and from the second (snd1 / back) I send it to the line input of the second card.
The sound transmission from snd1 to snd2 is made with a short cable. Since the amplified linear signal is transmitted, this does not affect the sound quality too much. On the second card (snd2), monitoring of the line input is enabled, plus I add to it what I say into the microphone, and, thus, a ready sound is obtained for the broadcast. In order to capture sound from snd2 for recording or transmitting digitally, it must have the “what you hear” recording mode selected on it. Previously, such a recording mode was rarely met, now, it seems to be easier with this ... in general, snd2 has such a mode, otherwise it will not work.
Thus, the sound that I send to the server, that is, a ready-made stream containing: me, the second presenter, and jingles / interviews / music, I take from the exit called “what you hear” of the snd2 card.
So at my workstation, I use:
- foobar2000: music player;
- Skype: connection with the second presenter, guests, phone calls;
- VST Host: realtime effects to the microphone;
- Audacity: record and monitor stream;
- Edcast Standalone: translation of the generated stream to the server.
By the way, if you use not hardware sound cards, but virtual ones - a set of software and in general the picture will not change much.
Then a little more about each of the programs used.
foobar2000
The choice is justified only by the fact that it is familiar to me easily customizable player. During the broadcast, it is important that it is convenient to start / brake the music, control the volume. In fact, I launch two copies of FUBAR, two players are needed to make smooth transitions between songs or play some jingles against the background of sounding music. I know that there is some kind of specialized DJ software for this whole thing, but I am doing quite well just with two fubars.
Skype
Well, everything is clear, except for one thing: new versions of Skype are played only on the front output of the sound card. It's just a little bit how they came up with a great idea, only because of this you have to use the old versions. At one time I used the hardware fork, but this method has its drawbacks. In short, it turns out too quiet and there is no possibility to separately control the volume of the sound going to the headphones and on the snd2. And it’s very convenient: everyone has it, you can hold conferences, invite guests on the air, call reporters on assignments on mobile phones, it has happened in our practice and such.
VST Host
In the chain appeared almost by accident. The fact is that this particular microphone model has some strange approach to the sound: it sounds only in the right channel. In order to fix it on the fly and a vst host was added, such things are done in it quite easily. In addition, you can turn on, for example, a compressor or some other effects. At the moment I’m just fixing them for the lack of my microphone. Another function that it performs is the transmission of sound from a microphone to snd2. This could have been done in some other way, but since I use this host, why not in it. Accordingly, in it I turn off the microphone when it is required (usually in pauses, while the music is playing).
Audacity
In addition to recording, I also look at the wave being formed to monitor the ratio of the volume levels of myself / music / guests. The record I keep in my home is just a backup, the master record is kept by the radio station server itself. If the sound after the ether is processed by me, then this is done just in Audacity: normalization, compression, noise removal.
Edcast standalone
Transmits the generated audio stream to the server, where it is received by icecast2, which is running on our server side. SB Live! - this is the same snd2.
That's probably all on the part of the workstation. This option is not a dogma at all, rather an illustration of how the problem can be solved with the available means. Now I think how to simplify the circuit using a normal hardware mixer, or vice versa, switching to virtual sound cards.
Bonus - a screenshot at the time of the broadcast. On the next desktop, shunots and chat are usually running to communicate with listeners.
Clearly you can ask questions, suddenly forgot to write about something.
Source: https://habr.com/ru/post/118667/
All Articles