Unoffical empeg BBS

Quick Links: Empeg FAQ | RioCar.Org | Hijack | BigDisk Builder | jEmplode | emphatic
Repairs: Repairs

Topic Options
#9004 - 05/06/2000 11:41 dynamic range compression
eternalsun
Pooh-Bah

Registered: 09/09/1999
Posts: 1721
Loc: San Jose, CA
The empeg tech specs indicate there is an "other audio feature" called dynamic range compression. What is that?

Calvin


Top
#9005 - 05/06/2000 13:30 Re: dynamic range compression [Re: eternalsun]
tfabris
carpal tunnel

Registered: 20/12/1999
Posts: 31597
Loc: Seattle, WA
The empeg tech specs indicate there is an "other audio feature" called dynamic range compression. What is that?

Heh, I didn't notice that tech spec.

It's the whole thing that we've been having long discussions about elsewhere on the BBS. Dynamic range compression is the ability to balance and control the difference between the loud and soft parts within a song, or from song to song in your collection.

Until recently, we were mostly under the impression that we were talking about normalization. But most CDs are already more or less normalized when you buy 'em, so all discussions of normalization end up deteriorating into a discussion about dynamic compression.

So where did you see this tech spec? On the new web site? If so... why haven't the guys@empeg chimed in here and told us about it yet? C'mon, spill it, guys. Were you hoping to surprise me with it?

___________
Tony Fabris
_________________________
Tony Fabris

Top
#9006 - 05/06/2000 14:31 Re: dynamic range compression [Re: tfabris]
altman
carpal tunnel

Registered: 19/05/1999
Posts: 3457
Loc: Palo Alto, CA
The DSP has a dynamic range compression option, though it's not brilliant (which is why we didn't bother with it). It works, but the effect is maybe a bit too slight... it also is very fixed in how it operates.

Hugo



Top
#9007 - 05/06/2000 15:09 Re: dynamic range compression [Re: altman]
tfabris
carpal tunnel

Registered: 20/12/1999
Posts: 31597
Loc: Seattle, WA
The DSP has a dynamic range compression option, though it's not brilliant (which is why we didn't bother with it). It works, but the effect is maybe a bit too slight... it also is very fixed in how it operates.

Aw, shucks.

Okay, then, have you had a look at Rjlov's code? I haven't been able to try it out yet, although I keep wanting to. I know you're busy this week trying to actually, you know, ship an actual product and everything. But I was just wondering what you thought of his approach and if you'd consider implementing it or something like it down the road.

In fact, while I'm on the subject, I'm in a bit of a quandary with versions. I'm at 10a (pre-buttonfix) now. I want to try the version 11 stuff that I saw in San Jose, but I also want to try rjlov's volume compressor. Rjlov's stuff is at beta 10 (not 10a), but the version 11 notes says I should have the 10a-buttonfix kernel installed before installing the version 11 player. Is there any way I can run Rjlov's v10 kernel and the version 11 player together, or does he need to compile a v11 kernel with his stuff?

___________
Tony Fabris
_________________________
Tony Fabris

Top
#9008 - 05/06/2000 15:57 Re: dynamic range compression [Re: tfabris]
rjlov
member

Registered: 16/12/1999
Posts: 188
Loc: Melbourne, Australia

Hiya.

I've built a beta10a kernel, but I haven't tested it, 'cause I'm still running beta10 on my empeg. Sometime I'll get around to installing beta10a, then beta11pre, and then I can see what happens. I'll make the beta10a kernel available this weekend in the same place as the other one, but I don't know if it will work!

By the way, I'm really happy with it so far. I definitely would like more lookahead, and it would be nice if the volume correction was reset at the beginning of each track, but these things aren't easy to do from the kernel (at least, not for me).

Pros: I can listen to 2112 and actually hear him tuning his guitar. I can listen to The Planets, and actually hear Neptune!! All this travelling in the car without changing the volume.

Cons:
- When you have a long quiet bit, and then a very fast gradual (not immediate) loud attack (e.g. string entry), you can get an audible artifact. I'm pretty sure that a longer lookahead would significantly alleviate this problem, but I'm also looking into some slight changes in the behaviour when responding to potential clipping.
- I'm not 100% happy with the silence treatment. Currently, if it detects silence, it just leaves the output multiplier whatever it last was. I'm thinking it might be better to have it gradually decay back to 1.
- It might be better to base the multiplier calculation on the power, rather
than the peak value.

Richard.


Top