Topic: RockBoxed h120's AGC

[travelinbeat]

I've found that this unit's AGC sets the levels too high anyways.  I run CA Cards > CA-9100 Pre > h120 and generally tape loud rock shows.  I've experimented twice with setting the Preamp fairly high (roughly + 15dB), then tossing about +10dB on the iRiver and turning on the AGC.  In both instances I have walked away with hot recordings.  A Couple of days ago I taped another show, using the exact same method, however once the iRiver had 'turned down it's volume' so to speak, I twisted the knob of the CA pre to turn it down even more.  This worked like a champ since the iRiver AGC is smart enough to not increase volume, but only decrease it.  Thus I taped a whole show well below peak and bumped the volume 160% in Soundforge later.

I'm writing not only as a warning to people who may want to engage in "set it and forget it" recording, but also I am wondering if anyone else has had similar issues or has devised a better way of solving the issue?
Thanks for any input!

[som]

I think I may have encountered the exact same thing....

http://taperssection.com/index.php?topic=90301.0

[travelinbeat]

I think so-- thanks for the heads up =)

[SmokinJoe]

I've used it a couple of times without a problem... maybe I just got lucky.  My shows probably weren't "really loud" though.  I had CA cards > ST9100 at full gain.  My recollection is that I set the gain to about +15, and that it works it way down to about +6.

There is an adjustable parameter as to how long the peak has to be before it ratchets down.  I have my "AGC clip time" set to 400 ms, which is the default I believe, with the AGC set to "Safety".  I happen to have an old (maybe a year old) version of Rockbox.

[petur]

I have my agc-safety set to 200ms and it always responds to peaks. 99% of the shows I tape are loud rock concerts, so I'm fairly confident about the iriver/rockbox side of things. I have an h380 but the adc is the same as in the h1x0 series.

There used to be issues in the peakmeter (possibly missing high peaks), but these were fixed ages ago (at the time AGC was added). I can't think of anything else here....

Is the overloading constant or just very short peaks?

If you (or anybody else) find a way to reproduce this, do let me know (PM me to make sure I see it) and I'll have a look.