If the remaster sounds louder, is it too compressed?

Hi Davey,

I believe what you're saying is true IF the compressor has an auto gain compensation feature and it is engaged (or gain is applied manually).

When you set a compression ratio and a threshold on a compressor, you are telling it what volume above which you want it to act (threshold) and how much you want it to act (ratio). As such, it will reduce the level of sounds exceeding the threshold. Gain "compensation", used to "make up" the reduction, is what raises the level of the quieter parts. I am not aware of anyone who has used compression without gain compensation of some sort. Perhaps that is why compression by itself is often assumed to raise level. It is used as part of a process (with gain) to raise level.

Barry
www.barrydiamentaudio.com
www.soundkeeperrecordings.com

 

Technically, gain compensation is part and parcel of modern compression. That's the distinction between compression and limiting. Compression will always drop the gain as the sound coming in gets louder. Limiting sets a threshold for peaks only, leaving the low bits (and the gain) alone. Compression engages both ends.
This is an extremely generalized explanation, though. Here's an article I like that seeks to make the differences clear. Even though it deals with keyboard sounds, the explanation of compression vs. limiting applies across the board (so to speak):
http://members.aol.com/uniquenyc/key10.htm

 

Hi Chris,

"Part and parcel of modern compression" agreed but on all the units I'm familiar with the engineer must engage the compensation (unless he/she is using presets of some sort).

Speaking only for myself, I would avoid any device (or software) that does anything to the sound "automatically".

Best regards,
Barry
www.barrydiamentaudio.com
www.soundkeeperrecordings.com

 

Good point; I guess I meant, "what people mean when they say compression has been applied."

 

Barry,

Thank you so much. That is EXACTLY what I was getting at with my earlier post (which you kindly replied to).

Without the GAIN, all that happens it that the peaks will be lower and (overall) the recording will be quieter.

I realised it was inaccurate of me to imply that "the WHOLE thing will be quieter", because any passages below the threshold would obviously be unaffected.

Would I be correct to state that the (compensating) GAIN control affects the WHOLE of the recording? If this is set too HIGH, the end result will be "clipped / distorted."


Thank you once again,
Stephen

 

That is why brickwall limiters are used. They prevent clipping.

 

Yes, gain compensation controls the whole of the recording and is not threshold based. If set too high the result *will* be clipped and distorted.

Grant rightly points out that this is why many folks use brickwall limiters. I'm not sure though, that I'd always agree with the statement that "they prevent clipping". At a certain point they'll "soften the edges". Beyond that point (where all too many records go nowadays) to my ears, the signal is clipped and distorted, no matter how much "softening" is applied.

I'm all for pumping it up when the music and mood call for it. I just prefer to do it with my volume control. ;-}

Barry
www.soundkeeeperrecordings.com
www.barrydiamentaudio.com

 

Well, they are still going right up to FS "0" with the limiters before the dithering to 16-bits. There's no headroom.

I'm all for pumping it up when the music and mood call for it. I just prefer to do it with my volume control. ;-}
[/QUOTE]

 

Thanks for the clarification, Barry. All of the software-based compressors I've used boost the average level, so that's what led to my post. But distinction duly noted, nice to have folks around here like you, with all your knowlege and experience. And a nice guy as well

 

I am a more visually-oriented person so this site, full of colorful graphs, helped me a lot to understand overcompression:

The Death of Dynamic Range: A Chronology of the Compact Disc Loudness Wars

BTW, does this sound correct?:

As far as actual loudness - and I'm thinking of the loudness most people think of when they see the word "volume" on a home receiver or car stereo - if a CD is a little louder than another, just reach over & give the knob a slight twist to the left - no biggy.

As far as longer vs. shorter sample word lengths: it is my understanding that using all the bits in a particular word is better because then the DAC has more voltage level "choices" when building up the soundwave.*

using the binary numbering system.....

2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2 = 65,536 individual voltage levels to choose from


So if an engineer sets up his recording gear so that only a maximum of 14 bits are used (so the CD is lower in overall volume than another version that uses all 16 bits) this is what happens:

2x2x2x2x2x2x2x2x2x2x2x2x2 = 16,384 individual levels


Now there's a lot less levels to choose from but the real problem is that these levels are now much more "chunky" than the levels obtainable from a 16 bit word. So the DAC in the music fan's CD player cannot build the soundwave as accurately and very low-level sounds in particular aren't as detailed (a common complaint of many classical fans when discussing the CD format vs. dvd-audio and sacd)

For some more voltage level perspective, here's a 24 bit word:

2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2x2 = 16,777,216 individual levels (MUCH more finely divided; as far as higher fidelity is concerned, this is why many people put more faith in this concept than they do in a higher sampling rate).

* obviously that also means the analog-to-digital convertor back in the studio also has more choices when looking at the soundwave signal from the microphones

 

All but one software-based compressors that I have here all have means to adjust the gain.

 

Hi Taurus,

You're writing about two different things here:
Word length is independent of level, which is what you'd be referring to when you say "using all the bits".

A "bit", whether it is part of a 16 bit word or a 24 bit word, etc. takes up 6 dB of the dynamic spectrum so if the peaks are in the top 6 dB (i.e. peaks at -6 or higher), you're using all the bits.

While this is a good idea, taken to the logical extreme, it can be a problem. For example, if the musical performance has a dynamic range of greater than 6 dB as (one would hope) most does, the lower level information is not going to be encoded using all the bits. The only way to change this is to compress the signal dynamically before encoding it. Squeezing it into the top 6 dB will ensure maximum bit usage.

So maybe the major labels are trying to do us all a favor and make sure every single bit is being filled? ;-}

Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

OK, thanks for the clarification to you too, Grant. I guess the ones I've used are like that one of yours. I've only used the compressors that come with audio software packages, and they only offer compression level, threshold and attack characteristics. Must have built in gain compensation based on compression level selected. Sorry for keeping this thread alive long past it's expiration date. But I think most of us can agree that remasters shouldn't always get a new dose of added compression

 

Thanks for the clarification (I had to rewrite my post twice about this brain-melting subject before sending it).