More > JRiver Media Center 21 for Windows

NEW: Volume Leveling uses the additional Headroom provided by Internal Volume

<< < (2/10) > >>

mwillems:

--- Quote from: lisbethfox on August 13, 2015, 12:53:34 pm ---At the moment I use a receiver thats confirmed in its pure direct mode to do nothing to process the sound to create bi amped front outputs for a 5.1 rig. In the future I'm moving to a studio interface and many many more channels. However that would mean needing to control volume in software and not in hardware passively (my ideal goal but I don't want to build a 16ch passive volume control). Mainly what I'm wondering is what exactly internal volume / gain does when it comes to 'room correction'. I currently do nothing in room correction other than generation of sub signals since I have passive gain controls on my array of amps, however when its enabled i still see 'overflow' on some films (which I can confirm disappear if it or any sort of dsp is disabled, including my simple copy l to rl and r to rr to create a bi amp) show up if I have "flatline overflows" on and with clipping protection it can stick at 100 on alot of tracks which really should not be anywhere near that level. Do you know why this happens? Im sure its not worrying as it often doesn't sound bad. Just trying to better understand it for when I move to internal volume.

Thanks

--- End quote ---

Could you post some screencaps of your relevant DSP studio settings?  Most audio shouldn't clip regardless of JRiver's volume settings, but if adding DSP is causing it to clip, then something is adding gain somewhere and it would be good to find out where.  To be clear, I don't see anything like that when using internal volume and lots of DSP/Room Correction (so long as I manually offset any boost).  With clip protection on, it won't sound bad at all because MC just turns things down when clips would happen.  But it's not ideal as tracks won't be as dynamic as they would be if the clipping were resolved.

Lots of audio does have momentary clipping in it (peaks slightly above 0dB), but that doesn't sound like what you're describing.

lisbethfox:
I never boost. I'm out of the room at the moment (gimme 20 mins to run across the house) but pretty much for the 2.0 zone its 7.1 channels with no up mixing then the only DSP is Parametric EQ which has the following

Copy L to RL (I think its whatever the rear 7.1 is)
Copy R to RR (same deal)
Add Left to Sub -3db
Add Right to Sub -3db (because apparently when you mix channels you need to take each down 3db)
Low pass at 50 on the subs (I don't mind a tiny bit of overlap occasionally, if it bugs me for certain tracks I put it back to 40)
High pass at 40 on the rears which make up the low end of the bi amp'd fronts.
Thats it.

mwillems:

--- Quote from: lisbethfox on August 13, 2015, 01:03:11 pm ---Add Left to Sub -3db
Add Right to Sub -3db (because apparently when you mix channels you need to take each down 3db)

--- End quote ---

This is probably part of your problem.  A -3dB reduction is not enough to avoid clipping when summing channels in all (or most) cases.  It's true that combining two perfectly uncorrelated signals sums +3dB; however, summing two perfectly correlated signals sums +6dB (which is why a linkwitz-riley crossover sums flat and a butterworth doesn't).  

Most real world signals are somewhere between those two poles (not perfectly correlated, but not perfectly uncorrelated either).  But you will always need a minimum of 3dB of attenuation to avoid clipping when summing channels, and except when summing uncorrelated noise you will need more.  4.5 dB is a compromise position that won't guarantee no clipping, but will do a much better job than 3dB.  If you want to be perfectly safe, 6dB is the answer.

Additionally, I think the "add" reduction may only reduce the volume of the channel being added, not the destination channel, but I'd need to check on that to be sure (I'm not in front of JRiver right now).  It's worth investigating to make sure though, because if that were so you'd have lots of clipping even with uncorrelated noise.

You mentioned room correction above, do you have any settings in that DSP module?

All that said, here's the good news: none of that will matter if you use internal volume unless you have the volume maxed out.  JRiver's DSP stack uses any headroom created by the internal volume setting, which allows signals that would clip at full volume to be represented without clipping or additional attenuation.  So if you have internal volume at 100%, and clip protection engages and cuts the volume by 3dB, if you have the internal volume set to -3dB or below, clip protection will never engage because the necessary 3dB of headroom is already there.  Also if you need to add boost to the chain for some reason, you can just make sure to set your maximum internal volume so you don't compromise your headroom "envelope."

If you use other volume control methods, clip protection will engage and turn down the volume anytime a clip could happen.  That's probably what's currently happening in your setup when you see clipping, and why it doesn't sound like anything.  Watch the "audio path" while it's happening and see if clip protection shows up.

lisbethfox:
So if I do 6 db each for the subs I should be in a possible non clip situation? I guess I didnt understand the logic there before but that makes more sense. I have enough gain on the analogue side to easily turn it up so thats no issue, its only about making the initial digital signal as clean and hopefully unclipped as possible.

I would love if you could check on that and tell me but I definitely am not hearing clipping on the lows. For reference my low woofers are comfortable playing 100 dbs with massive excursion so I feel as though I would be hearing distortion or worse seeing it.

Room correction is only engaged on my 5.1 channel zone which selects when I play anything with more than 2 channels. The PEQ for it is only copy l to rl and r to rr. Room correction does absolutely nothing but creating a sub woofer channel from all 5 in play (For 7.1 sources I have JRSS mixing enabled which I assume is processed long before it reaches room correction)
.

lisbethfox:
Very interesting.

Using the previously described 2 channel preset
Disabling both  L to RL and R to RR change an overflowing track (a max of 106% more than it should be at any point) to 84% consistent. The subs are still at -3 with no negative impact. Worried my copying is causing that extra % of gain??

Navigation

[0] Message Index

[#] Next page

[*] Previous page