Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[etc6849]

LFE levels to calibrate to when measuring: +10dB above other channels or not?

I am using several external pro audio active DSPs/speaker processors (Xilica XD4080, 4 channels in (via AES), 8 analog out to amps, but I use three since I'm running 5 subs and tri-amping 5 channels) connected to my PC running JRiver via RME AES outputs.

My understanding is movies are distributed with the LFE channel being -10dB lower (so as to gain 10dB of headroom during playback).

Now that I don't use an AVR for bass management or LFE, I need to know where to set levels for the LFE channel using RoomEQ Wizard (REW).  I am assuming when I plot SPL in REW, that the LFE should measure 10dB acoustically higher than the other channels?

Also, what impact does using JRiver have on LFE levels?  I plan to keep using my Xilica processors for bass management, so I have this set to silient in JRiver.  Anything else to be aware of?

[mattkhan]

LFE levels to calibrate to when measuring: +10dB above other channels or not?

Bear in mind that LFE is an input channel whereas your sub(s) receive the SW output channel which typically plays back the LFE + the low pass of each main channel. Thus level setting your sub & managing the LFE are 2 different things.

My understanding is movies are distributed with the LFE channel being -10dB lower (so as to gain 10dB of headroom during playback).

no, 0dBFS is 0dBFS irrespective of which channel it is on. The difference is that the LFE requires +10dB in order to bring it to the correct level relative to the mains. This must be implemented before bass management so the typical scheme when doing this in jriver

-15dB (technically -18dB is required for 7.1 source in order to guarantee no clipping during bass management) applied to all main channels
-5dB applied to LFE
low pass the mains
sum them all
send the result to the SW output
add 15dB back to the mains and send them on to the DAC to head to the respective speaker

what to do in jriver depends entirely on how you've implemented bass management in the xilica

[etc6849]

With respect, I'm not asking about bass management.

My sub does receive the LFE channel if I measure AES channel 4 in my measurement software(LFE)   So the question is if I measure AES channel 4 in REW, should the level be 10dB higher (as acoustically measured at the MLP) than measuring AES Channel 1 (FL) or FR, C, etc...

It would seem a simple AVR would measure 10dB high if you measure PCM over HDMI for HDMI4 (LFE) versus HDMI1, 2, 3, 5, 6, 7 and 8. 

Let me address your concern though, but please lets ignore bass management going forward, as that is separate from my question...  As you say, I use the three Xilicas (each Xilica contains a DSP and DAC channels) to filter and sum these, then output each separately to 5 subs (so as to time align my 5 subs at the MLP).  I totally understand your concerns about having headroom as I have the subs set low in the Xilicas (which also have internal DACs) and gain matching is perfect.

Just asking about LFE level calibration (relative to the other channels).  I can correctly call AES channel 4 the LFE channel too as I am not using JRiver to do any bass management (have this option set to "silient.  redirect bass using room correction"); e.g. AES 4 contains no bass from bass management, only the LFE from 5.1 and 7.1 mixes.

[mattkhan]

it's hard to answer questions like this with partial information if all you're asking is whether a signal sent down channel 4 in REW (to an input before the LFE is subject to a 10dB gain increase) should come out 10dB higher then a signal sent down a.n.other channel then the answer is yes

[etc6849]

Yes, this is what the title and OP was asking (and what I thought/did too).  Thank you for confirming this as it is the most complex setup I've done and I wanted to be sure what I did made sense.

My setup has 20 unique channels of amplification: 10 Benchmark AHB2 amps (4 in bridge mode), and 5 JL subs with internal amps (2 F212 subs and 3 F113 subs).  The THD at the MLP is .1%-.2% for 98% of the audio band (and measured close to 90dBC).  Many things got me to incredible distortion levels such as: the quality DACs and FIR filters of the Xilica XD4080 processors, room treatments, top notch world class amps (best measurements of any amp), multiple subs and actively tri-amping + crossing to multiple subs.  The subs essentially receive the same signal except for it is time delayed for each sub separately.

This is all for reproduction of only 5.1 channels.  I will eventually post some diagrams and put a link in my signature.  Suffice it to say, it is the best/cleanest system I have heard and the culmination of years of study.  Just making the cables, rack mounting and setup was probably 100 hours.

it's hard to answer questions like this with partial information if all you're asking is whether a signal sent down channel 4 in REW (to an input before the LFE is subject to a 10dB gain increase) should come out 10dB higher then a signal sent down a.n.other channel then the answer is yes

[blgentry]

and 5 JL subs with internal amps (2 F212 subs and 3 F113 subs).

You had me at 5 JL subs! Wow.  If done correctly that could be amazing!  Have you consulted the Sound Doctor? He's JL's in house tech support for JL subs and is the most knowledgeable person I've ever talked to about subwoofers.  I've spent a lot of time studying subs.  I've built dozens.  Sound Doctor knows WAY more than I do.

Brian.

[etc6849]

Some measurements after running Dirac.

https://drive.google.com/drive/folders/0B1J0a4OV_WGLSzN1RFpma1RaM1E?usp=sharing


The waterfall plot came out way better versus using two subs:

You had me at 5 JL subs! Wow.  If done correctly that could be amazing!  Have you consulted the Sound Doctor? He's JL's in house tech support for JL subs and is the most knowledgeable person I've ever talked to about subwoofers.  I've spent a lot of time studying subs.  I've built dozens.  Sound Doctor knows WAY more than I do.

Brian.