Automatic Line level Attenuation Circuit

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
System set-up ... CD-player ----> -10/+4 interface ----> digital processor ----> amp ----> speakers

Does a diy solution exist for resolving the following issue:

When playing cd's .. various era's have differing output levels (+/-3-4db). While the -10/+4 interface and the digital xo have input level controls ... it get's to be a PITA adjusting the levels for the various program materials. I also have people here that can't be bothered with monitoring levels. CD-player has no attenuation control.

I'd like to set the system gain and forget it. Putting an automatic device between the cd and the interface would accomplish that. Looking for something that I can set gain with and have it automatically attenuate higher cd output levels without user intervention. Not looking for fancy .. just simple.
 
Last edited:
I read up on that a little bit earlier today. Seems that what I'm after might not be that involved?

RDL has a device that sounds like what I'm after but I'm not too sure. It's a pretty small package .. but at $250 I was hoping that there might be a diy equivalent.
 
AGC were in all mic circuits used by the Ham radio operators.
They are also inside every car radio and all FM radios.
The AGC circuits were implemented with just a few transistors and other very cheap components. You could build a two channel version for a few $. Make sure the two channels track each other.
 
My understanding of an AGC is that it dynamically adjusts the level ... for this application that won't be optimal. I don't want the program material adjusted ... just the global level of it.

Thinking this through ... I don't believe what is needed can be fabricated. It should function as a shelving circuit when a given cd would redline the digital xo ... then release the shelving on a disc that won't redline <---- problem: without a cue on disc this action would be subjected to/by program transients. Unacceptable.

I see a switchable L-pad in my future.
 
System set-up ... CD-player ----> -10/+4 interface ----> digital processor ----> amp ----> speakers

Does a diy solution exist for resolving the following issue:

When playing cd's .. various era's have differing output levels (+/-3-4db). While the -10/+4 interface and the digital xo have input level controls ... it get's to be a PITA adjusting the levels for the various program materials. I also have people here that can't be bothered with monitoring levels. CD-player has no attenuation control.
<snip>

I tool my old low level CDs, ripped them, adjusted the audio levels and burned new discs. Adobe Audition can do wonders - almost miracles.

 
Puppet,
a gain block/ buffer at your pre-amp does what you want, but not automatically.

Set the gain to +0dB, or +6dB, or +12dB. That covers the range of inputs from CDP @ ~2Vac down to 500mVac.

You could add +18dB (250mVac) if you needed a really sensitive input for very low level gear.

That buffer gain block can be at your listening seat. The buffer can then drive long cables, 3m to 20m if that long is necessary, to reach the Power Amplifier/s attached to the back of your speaker/s.

You can set the steps as large or as fine as suits you and your gear. Steps of 2dB or 4dB or 10dB. Just select the resistors to suit the gain block topology.
 
Last edited:
AX tech editor
Joined 2002
Paid Member
My understanding of an AGC is that it dynamically adjusts the level ... for this application that won't be optimal. I don't want the program material adjusted ... just the global level of it. .

It might work if you gave it a very long time constant like several seconds. But yes, you don't want it to follow the program level envelope.

jan
 
Account Closed
Joined 2001
The procedure offered by stratus46 seems the only solution to this issue.

To "normalize" played back material on-the-fly in real-time would be impossible. You need to rip the music into a suitable editing program like Audition, Goldwave, Audacity, etc, etc and then analyze the tracks to determine the increase in levels to achieve 0dbFS on the loudest peaks. Then adjust and save it.

Even then, should a person normalize each track of an album this way (with possibly different normalized levels) or evaluate the entire album and increase each track only by the (same) amount it took to normalize the highest recorded level track? :)

Cheers,

Dave.
 
Last edited:
The procedure offered by stratus46 seems the only solution to this issue.

To "normalize" played back material on-the-fly in real-time would be impossible. You need to rip the music into a suitable editing program like Audition, Goldwave, Audacity, etc, etc and then analyze the tracks to determine the increase in levels to achieve 0dbFS on the loudest peaks. Then adjust and save it.

Even then, should a person normalize each track of an album this way (with possibly different normalized levels) or evaluate the entire album and increase each track only by the (same) amount it took to normalize the highest recorded level track? :)

Cheers,

Dave.
I've pretty much assumed that the engineer and the artist has covered the disc's content ... as far as dynamics go .. most of the time. Shouldn't need to go track by track adjusting levels. ... and I haven't run into a situation were that is required anyway. Ripping and normalizing everything is an option ... a daunting one :p

Thanks for all of your replies fellas ... much appreciated!
phil
 
I've pretty much assumed that the engineer and the artist has covered the disc's content ... as far as dynamics go .. most of the time. Shouldn't need to go track by track adjusting levels. ... and I haven't run into a situation were that is required anyway. Ripping and normalizing everything is an option ... a daunting one :p

Thanks for all of your replies fellas ... much appreciated!
phil

LP levels were all over the map. CD levels are less bad but I have some old discs where the PEAK level barely make -12dB FS - and this is a Telarc disc, generally regarded as good stuff. Same for some early Deutche Grammophon discs.

I do this for the MP3 files on a flash drive in my car routinely and it really doesn't take much time. It's far better than having to keep altering the volume. It only needs to be done once.

 
The procedure offered by stratus46 seems the only solution to this issue.

To "normalize" played back material on-the-fly in real-time would be impossible. You need to rip the music into a suitable editing program like Audition, Goldwave, Audacity, etc, etc and then analyze the tracks to determine the increase in levels to achieve 0dbFS on the loudest peaks. Then adjust and save it.

Even then, should a person normalize each track of an album this way (with possibly different normalized levels) or evaluate the entire album and increase each track only by the (same) amount it took to normalize the highest recorded level track? :)

Cheers,

Dave.
It's even worse - the CD maximum level is determined by the peak of the signal, but the audible volume is determined by the average. Recordings from different eras go through vastly different amounts of "hypercompression" and the volume difference between them will be great, even though both recordings may peak at the same level.

If you're going to use and audio file editor to change volume, you will be best off setting the volume to how it SOUNDS. This means REDUCING the volume of more recent, hypercompressed recordings.
 
Account Closed
Joined 2001
Yep, highly compressed audio material tends to (generally) "sound" pretty loud already and has peaks pushed to 0dbFS levels in many cases. So, yes, you could reduce levels of those tracks/recordings vice normalizing lower recorded level material.
I'm assuming a person doesn't want to "un-limit" or "un-compress" the audio, so those and doing nothing are the only three options.

It's a sad state of affairs that exists today.......we have playback equipment that will accomodate the dynamic range of recorded music better than ever before, yet we compress and limit the hell out of it. What an upside down industry. :(

Cheers,

Dave.
 
Yep, highly compressed audio material tends to (generally) "sound" pretty loud already and has peaks pushed to 0dbFS levels in many cases. So, yes, you could reduce levels of those tracks/recordings vice normalizing lower recorded level material.
I'm assuming a person doesn't want to "un-limit" or "un-compress" the audio, so those and doing nothing are the only three options.

It's a sad state of affairs that exists today.......we have playback equipment that will accomodate the dynamic range of recorded music better than ever before, yet we compress and limit the hell out of it. What an upside down industry. :(

Cheers,

Dave.
It's the same stupidity that runs FM radio. Loudest wins. The difference is the FCC requires they stay within their channel while on CD there are no cops - not that would make an iota of difference. If people continue ti buy crap, you'll get crap. As soon as they stop buying it, it will change but not until then.

BTW it did do an uncompress on a pair of discs though there was one track that had some actual audible clipping that became more pronounced after the un-limit. The original level on the disc was about 6 dB louder than FS and just simply clipped like mad though it wasn't particularly audible. In other words, it looked worse in Audition than is sounded. Since I wanted to reduce the average level to keep it in line with other material, it seemed worth an experiment but in fact dropping the level would have been OK.

 
Administrator
Joined 2004
Paid Member
What you need is called "Replay Gain". It's built into many software players and some portables- I can use it in JRiver Media Center, for example. I don't know of any CD players that do this - tho some BluRay might - nor have I seen a stand alone box.

What happens in Replay Gain is that the audio files is scanned and a level is determined. That level is written into a tag that tells the player how much to change the volume of that track to make it "Normal". No change is made to the actual audio - it's just an automatic gain control per song or per album, depending on which you choose at the time of playback. You would probably choose a "per album" setting. If you were mixing songs in a playlist, then a "per song" setting would work better.

To do this, you have to have your files on a computer. There may be some disc player out there that can do it, but I don't know how it would determine the levels without pre-scanning.
 
Pano ... that sounds like the only way this would work.


http://en.wikipedia.org/wiki/ReplayGain

hmmm ... I could see a standalone box designed for aftermarket cd applications. Wonder what the hang-up is on something like that?

Would be nice if the industry adopted a level "cue" fingerprint on pressing. Players could have built in attenuation circuitry activated by the cue.
 
Last edited:
Administrator
Joined 2004
Paid Member
Yes! It would be great to have the tag already in the file. Since there are multiple CD databases online, a player could look up the CD to find the usual tags like Title, Artist, Genre, etc AND replay gain settings. But so far, I've not seen that in the online CD databases.
 
What you need is called "Replay Gain". It's built into many software players and some portables- I can use it in JRiver Media Center, for example. I don't know of any CD players that do this - tho some BluRay might - nor have I seen a stand alone box.
Compressed file formats such as MP3 and Ogg Vorbis have a "playback level" data field for each block (which encodes perhaps 10 to 100mS of sound), and the playback device uses that to set the volume the block gets played back at. CD doesn't have any equivalent field.

But you could re-record the audio output of a ReplayGain modified file to .wav and burn that to a CD-R to get the same effect. Most player programs have a "write output to .wav" feature.

But the ReplayGain clearly manipulates this data field to change the playback volume.
What happens in Replay Gain is that the audio files is scanned and a level is determined.
This scanning is inherently imprecise - it's trying to recreate the levels the signal had before it was compressed, limited, clipped, and who knows what else.

The first video on this page is a description of such a process with audio samples (it's longish and he starts off with a lot of history, but the samples are about halfway through):
Loudness War: Peace is Almost Here!

With digital compression (as done in a DAW plugin) it's possible to generate data that could be used to regenerate the original dynamics. Even with an analog compressor it would be possible to digitize the gain control voltage and do this. But the likelihood of either of these actually being done and the data put on a consumer medium is nil. What's even worse is the intentional clipping (yes, you read that right). There is software to "undo" clipping, but it mainly mitigates the effects - it can't bring back the original waveform.

The second video at the link is an older one on the Loudness Wars, but it may also be instructive if you haven't seen it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.