The latest version of this document is available at: http://www.radiovague.com/howtos/broadcast/studio.html
This document attempts to provide hints and tips on handling analogue audio up to the point at which it enters the encoder (internal or external soundcard) for streaming to Radio Vague servers.
It is useful in audio to think of the analogy of the weakest link in the chain. One duff component can break the chain. Audio is fragile unless handled correctly, and damage done at any point in the signal chain is unlikely to be corrected by the further inclusion of more equipment. The better you treat your audio up to the point at which it is encoded the better the resulting stream.
Remember that humans are blessed with incredibly sensitive ears capable of phase measurements down to several millionths of a second. After applying the commonsense technical tips (balancing audio, minimising noise, keeping levels around 0dB, avoiding feedback and distortion etc.) use your ears.... listen to the stream coming back in when setting up if you can, and take (human) feedback on the chin willingly, it will help.
Live audio is an art, this document does not confess to being anywhere near the whole truth, just a helping hand. Please direct feedback to someone@radiovague.com
Cable interconnects between equipment are crucial to good audio. I strongly recommend getting good quality cables with good connectors.
When choosing / using cables there are a few points to consider:
In the Radiovague mobile studios we use Van Damme OFC microphone cable and Neutrik connectors and solder them up ourselves, although I know that VDC will make custom cables using the same components. If you are doing a regular show on Radio Vague then contact us for cable help: someone@radiovague.com.
There are two main places where it seems advantageous to apply dynamics processing to live stream broadcasts.
Most dynamics processing units have settable response times: attack and release in milliseconds. Attack being how quickly the processing comes into effect, and release being how long it compresses after the louder transient has passed. These settings depend on the specific use and are covered separately.
Compressors generally have two important settings:
The way it works is this: The threshold is the audio level above which the audio starts to get compressed; compress meaning gently lessening the effects of any increase in the input volume. The ratio is the increase of dB it requires in audio input, above the threshold, to cause 1 dB of increase in the output audio. For example 4:1 means that you need 4dB more in to get 1dB more out, once above the threshold.
Say you want your stereo encoder input to be as close to 0dB as possible. So if you set your threshold to -4dB and your ratio to 3 to 1 (3:1), it will take 3x4dB = 12dB over the threshold to reach 0dB. This gives you much larger room for volume fluctuations in your broadcast content (12dB instead of 4dB).
We have had a lot of good human feedback using around 2.5:1 ratio and then setting the threshold to suit the encoders input level whilst varying the input audio to be broadcast wildly to setup.
Attack setting should not need to be too fast as limiting has hopefully been done further up the line, and release doesn't want to be too short, it's a smooth volume that seems to work best for streaming.
Do not forget that many people will listen to the show on laptop or tinny little "multimedia" speakers. It is a good idea to run test streams prior to broadcasting and listen to it with as many different systems as possible.
If you wish to use a compressor, or a limiter, as a limiter then you need to set the threshold just below the audio headroom you wish (sticking to our previous example about -1dB) and then turn the ratio up beyond 20:1 even close to infinity:1. Thus the resulting audio never quite reaches the maximum headroom of 0dB. Use your ears to see what ratio suits the setup.
Attack and release settings want to be fairly fast. Play with it and use your ears.
Noise gates can be very useful in conjunction with microphones. Using a gate means that when a very low level signal is going through a microphone (nothing but background studio noise for example) the output audio is compressed out of existence, meaning that the background noise is not mixed in with the main audio out.
When a louder voice goes through the microphone the gate turns back off and the ouput volume increases normally.
It is sometimes possible to use a compressor as a gate.
In the Radio Vague mobile studios we are currently using a handful of dBx 266L compressor / limiter / gate / expanders.
This is a big chapter - nothing to see apart from notes:
There will be a whole section about signal levels here.
There will be a section about power, keeping all audio kit all on the same power supply, keeping impedances low, and checking 4 ways etc.
A decent microphone is a great investment for making spoken broadcast sound "present" and intelligible.
There are many different types of microphone. The main types we deal with are:
When choosing / using any microphone:
Close vocal mics have forward gain so it is important that the microphone head points straight toward the mouth talking / singing.
Use a windshield / pop guard. The human voice has a large dynamic range and considerably variable tonal qualities. It is most sybillant on the letter S, and it can pop loudly on the consanants B and P. It is possible to buy windshields and pop shields but if you are strapped for cash a piece of fun fur or folded foam covering the microphone head will do, just make sure it is securely attached and doesn't squeak.
Practise mic technique. The closer you go to the microphone the quieter you should talk. If you want to shout pull away. This way you get your full range of expression without being unheard or exceeding the audio headroom of your mixer / preamp/ mic-in causing distortion.
Ambient mics pick up the surrounding sounds in much more of a "human ear way" than the close audio mics. They do not seem to be particularly sensitive to their orientation. They are great mics for catching the "vibe" of an event whilst amongst the crowd, but do not seem to respond well to being in front of loud sound systems.
It is important to mount or carry ambient mics well. Any rustling, chatting or noise near them will be picked up loudly.
Generally we would not use effects over music whilst streaming, but if a close vocal microphone is being used then a small amount of ambient reverb stops the resultant voice being "cold".
Reverb is simple to use. There is a delay value (which would represent roughly the time it takes the music to reach the nearest wall and back) and an amount of feedback (which represents the rate at which the miniuature echos die away).
Your reverb unit may also have a dry/wet mix setting. This is the amount of reverb you want mixed with the original audio; dry being no reverb, wet being nothing but reverb. How you use this depends on how you are set up. Use your ears on this one.
Subtle use is recommended for reverb in stream broadcasting. Use a fairly short delay, and not much feedback, although try not to make it sound like you are sitting in a toilet cubicle! (If it does, add a little more delay and feedback).
You may find that the reverb unit has presets on it rather than numbers, so choose an ambience setting or a reasonable size room, rather than a concert hall or underground car park.
Some reverbs are funkier than others, featuring multi-tap, release and decay settings, vintage reverb modelling etc. Again use your ears..
You may find that different radio presenters feel more comfortable with different settings so try to keep a record to help them feel at home if possible. This is a silly Radio Vague type comment which makes it sound like we have corridors of technicians running around sorting things out. haha.
In the Radio Vague mobile studios we are currently using Lexicon MPX110 Dual Channel Processors on close vocal microphones.
Balanced audio and balanced cables use a very cunning method to eliminate induced noise in cable runs.
Unbalanced cables have one positive core and one shielding connection within the cable. Any noise which reaches the core cannot later be removed. Phono (domestic HiFi) and guitar leads are a good example.
Balanced cables have a pair of inner cores (hot+ and cold-) and one shielding connection. The two inner cores run in anti-phase (when one audio signal goes up the other goes down). When the audio reaches the other end the signal in one core is SUBTRACTED from the other. This means the audio signal, being 180 degrees out of phase, doubles in strength and the noise, being common to both inner cores, is subtracted from itself. A massive improvement in signal to noise ratio is caused. XLR and Balanced stereo jack leads are a good example.
Always interconnect equipment which caters for balanced audio with the correct balanced leads.
Equipment is usually marked with the connections for shield, hot+ and cold- pins somwhere near the socket. Sometimes conventions change for XLRs and stereo jack connectors from one piece of equipment to another. It the equipment is not marked refer to the manual.
Article copyright (c) 2005 Radio Vague. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.1 or any later version published by the Free Software Foundation; with no Invariant Sections, with no Front-Cover Texts, and with no Back-Cover Texts. A copy of the license is included in the section entitled "GNU Free Documentation License".