I'm looking to build some kind of voltage controlled amplifier that will make the peak voltage of my audio always the same.
For example if somebody plugs in an MP3 player to my system I'd like the speakers to be the same 'loudness' regardless of what the volume is set to on their MP3 player.
However, I'd like it to sample over a minute or something so that if a song is just quiet during a portion it doesn't boost the volume to create a bunch of static.
For example if somebody plugs in an MP3 player to my system I'd like the speakers to be the same 'loudness' regardless of what the volume is set to on their MP3 player.
However, I'd like it to sample over a minute or something so that if a song is just quiet during a portion it doesn't boost the volume to create a bunch of static.
Compression has been with us for a lifetime in electronic instruments, mixing and studio production but taken to the extreme of levelling out all and any valid source/programme signal will also cut the dynamic range of the music to a dull, monotonous "one level" and adds some distortion in the process, assuming you have a wide range of signal levels to deal with. DSP can be used to engineer this better but the compression issue remains.
If you want auto level setting rather than compression, you do have to sample the signal level over the whole piece, track, movement etc. to enable an appropriate single level setting. I'd say that would be impractical as an add-on for using personal devices like MP3 players.
If you want auto level setting rather than compression, you do have to sample the signal level over the whole piece, track, movement etc. to enable an appropriate single level setting. I'd say that would be impractical as an add-on for using personal devices like MP3 players.
Last edited:
I remember when I used to burn my old LPs to CDR there was an option to 'normalise' the A-D conversion so that the highest peak in the track didn't digitally 'clip'. This had the side effect of making music loudness fairly even between LPs without squashing the dynamics at all. So it must be possible in the digital domain but it's fraught with problems in the purely analogue domain - to hear compression at work just listen to most pop radio or internet radio to hear instruments going up and down in volume as others join in.
To sample over a minute and still be ready for peaks means that you need to delay the audio for a minute. Doing this and maintaining some semblance of audio quality means digitising, buffering, then converting back to analogue. You also have to decide whether to use an exponential law for your AGC or a 'hang' type.
The quick answer is that it can't be easily done.
The quick answer is that it can't be easily done.
You could do part of a ReplayGain scanner in analog - the weighting and RMS bits at least. The rest, not so much.
Solving the OP's problem is not doable with analog electronics IMO - some AI would be handy, too.
Solving the OP's problem is not doable with analog electronics IMO - some AI would be handy, too.
Just spent 10 mins trying to find the DSP based limiter I stumbled upon before with no luck.
It cost about the same as the "Behringer Composer Pro-XL MDX2600 Processor" which seems to be an analogue solution.
You might want to look at Fast Audio Peak Limiter By Phil Allison (Edited By Rod Elliott - ESP).
Also think about making this "anything" input balanced. It may take single ended input but you wont have the nasty humm caused if soem one touches the input connectors.
I agree with others this is better done in the digital domain especially for Audio where you can put delay in the audio chain sadly I could not find the DSP solution I stumbled upon before.
It cost about the same as the "Behringer Composer Pro-XL MDX2600 Processor" which seems to be an analogue solution.
You might want to look at Fast Audio Peak Limiter By Phil Allison (Edited By Rod Elliott - ESP).
Also think about making this "anything" input balanced. It may take single ended input but you wont have the nasty humm caused if soem one touches the input connectors.
I agree with others this is better done in the digital domain especially for Audio where you can put delay in the audio chain sadly I could not find the DSP solution I stumbled upon before.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.