Topic: Edirol R-09 AGC issue and field report

[pgoelz]

Guys,

I'm posting this here so it doesn't get lost in the other big thread.

I just got back from a weekend music festival where I used my new R9 extensively.  With one exception, it performed flawlessly and the battery life was amazing. 

However, there is a serious issue with the AGC function.  I have posted here about it and I also have contacted Roland, but I wanted to post here again to warn people before they ruin a recording. 

The issue is that when the gain is changed in record, there is a click for each gain step.  This shows up as zipper noise when you ramp the gain up and down and can also be heard as a buzz as the AGC releases after a loud sound.  If you record something that is sufficiently loud to cause the AGC to work constantly, you will find clicks and buzzes throughout your recording.  I ruined one that way before I realized how bad it was and disabled the AGC. 

A secondary issue is that the AGC works on the two channels independantly.  This causes the balance to shift with program material. 

The only time I was able to use AGC without spoiling a recording was when I recorded a concert where the average level was below the AGC threshold.  The AGC reduced the applause OK and the applause masked the clicks. 

Just a heads up.  Other than that, my R9 is GREAT.  Anyone interested in a nice mint condition R1?  All accessories and box.....

Paul

[Brian Skalinder]

Just curious - why use the AGC feature at all?  It basically provides dynamic gain adjustments, something that to me seems undesirable for live music recordings.

[sleepypedro]

However, there is a serious issue with the AGC function. 

not to be glib, but i fixed that issue myself -- by locking AGC in the 'off' position and covering it with electrical tape.

AGC is evil.  normalize in post, no two ways about it. 

[BayTaynt3d]

Please don't use AGC... EVER! Thank you.

[pgoelz]

Why use AGC?  Good question.  I agree that in most cases, AGC is undesireable.... at least in its current incarnation.

I was using it experimentally.  In one case, the idea was to deal with dynamically changing levels in a situation where it was not convenient to monitor levels.  In the other case, I was trying to keep the record level reasonable while avoiding clipping on applause. 

I agree.... fix it in post is the best approach.  But it would be handy to have it available when you really need it.  And the way it is implimented, it is simply not workable due to the zipper noise. 

I still say that properly done, AGC can be very useful.  But it needs a variable release time including infinity....

Paul

[poorlyconditioned]

Why use AGC?  Good question.  I agree that in most cases, AGC is undesireable.... at least in its current incarnation.

I was using it experimentally.  In one case, the idea was to deal with dynamically changing levels in a situation where it was not convenient to monitor levels.  In the other case, I was trying to keep the record level reasonable while avoiding clipping on applause. 

I agree.... fix it in post is the best approach.  But it would be handy to have it available when you really need it.  And the way it is implimented, it is simply not workable due to the zipper noise. 

I still say that properly done, AGC can be very useful.  But it needs a variable release time including infinity....

Paul

I find AGC useful for voice recording, like family events, birthday parties, etc.  But I use minidisc, which I think has a good implementation.

  Richard

[BayTaynt3d]

I personally think AGC sucks beyond all belief. Now maybe an on-the-fly compressor or something I could potentially be talking into, but AGC is usually WAY TOO heavy-handed to trust and rely on. It is widely implemented on all consumer camcorders, and the destruction it wields is outrageous -- and is one of the reasons you should always buy a consumer camcorder that at least has a manual override on the audio levels. Anyway, just one man's opinion -- I'd pretty much NEVER use it for music even if it did work correctly (but I might use a limiter/compressor at times if it had one of those that was configurable on the knee/ratio).

[pgoelz]

(but I might use a limiter/compressor at times if it had one of those that was configurable on the knee/ratio).

That is exactly whay I have been trying to get Roland interested in adding variable release times for the AGC.  Get rid of the zipper noise and allow the user to set the desired release time and you suddenly have a VERY useful AGC system.  And the best implementation would be to have a compressor preceding a slow attack / slow release AGC.  That way the occasional loud sound would merely get compressed but longer term loud sounds would cause the ACG to gently reduce the gain.  Seems to me it would not be that hard to do in firmware. 

Paul

[udovdh]

The slow compressor etc is interesting.
The Microtrack has an ADC that can do similar stuff but it is not implemented!