Print

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

[mr.os]

Hello,
first post, first question that I can't answer myself with reading ....

I have a problem with the output format selection. Either I don't fully understand the concept or JRiver lacks the feature that I need.

I'm trying to implement a full digital loudspeaker system with FIR XOs and so on. Yesterday I tried my two two-way main speakers without problems. I created the filters and the convolution config and can use the "2 channels in 4" output container. Everything plays fine.
But now I want to add my subwoofers (with the sub routing functionality from JRiver) and need a "2.1 channel in 5.1 container". And when the system is completed I also have a center and some (2,3 or 4) surround speaker. So I need "5.1 channel in 10 container" or something like that. Can I configure something like that? I think at the moment it isn't possible, or miss I something?

And another question: can I configure the internal volume to be applied after the convolution? Now I have a delay of approximately 1s after i press the volume key. I can't use the soundcards volume control because it is channel independent.

Hopefully someone can help me.

Thanks and best regards
Andreas

[mattkhan]

You need a container that has enough channels for your output so if you have 2 way mains and a mono sub then you need a 5 channel (or more) container. You then need to write your convolver cfg file to populate those channels according to however you have it connected to your DAC, you can use whatever channel assignment makes sense for this. For example you might decide that 1=L1,2=L2,3=R1,4=R2,5=sub or use any random order you like, jriver doesn't (afaik) care.

[mr.os]

I understand that I need a container with enough output channels. Even that I can configure the channel routing in the convolver config.
But which container or format or whatever should I choose when I have 5.1 (subwoofer, mains, surrounds and center speakers) and I need 12 output channels for it? Or when I have 6.1 and I need 12 output channels.

(And my problem with the subwoofer was that I will use the subwoofer routing functionality from JRiver - so I need the 2.1 output format with (at the moment) 6 outputs (not available in any container))

[mattkhan]

this thread might be useful - http://yabb.jriver.com/interact/index.php?topic=81448.0

it's not 100% clear what your scenarios are, can you confirm? it reads like you're saying all your mains are active 2 way speakers where the crossover is implemented in convolver but you also have a sub that you always want to be active and you want it bass managed by jriver?

if this is correct then I think, but am not certain, you want to;

- assume normal channel positions for your mains (i.e. 1=L,2=R,3=C,4=LFE,5=SL,6=SR and so on)
- set output format to 16 channels
- turn room correction on to apply bass management to those input channels (and apply the +10dB to the LFE channel here) => this means you now have 1=high pass of the input L, 2 = high pass of the input R, 3 = high pass of the input C, 4 = LFE + low pass of all the main channels and so on
- position convolution after room correction and use your cfg to route input channels 1,2,3,5,6 etc to the relevant output channels (e.g. 1,9 for L;2,10 for R;3,11 for C; 5,12 for SL; 6,13 for SR)
- connect up your amps accordingly

i.e. you want to present a "conventional" set of channels to jriver for bass management and then break them apart for your active setup afterwards

[mwillems]

Hello,
first post, first question that I can't answer myself with reading ....

I have a problem with the output format selection. Either I don't fully understand the concept or JRiver lacks the feature that I need.

I'm trying to implement a full digital loudspeaker system with FIR XOs and so on. Yesterday I tried my two two-way main speakers without problems. I created the filters and the convolution config and can use the "2 channels in 4" output container. Everything plays fine.
But now I want to add my subwoofers (with the sub routing functionality from JRiver) and need a "2.1 channel in 5.1 container". And when the system is completed I also have a center and some (2,3 or 4) surround speaker. So I need "5.1 channel in 10 container" or something like that. Can I configure something like that? I think at the moment it isn't possible, or miss I something?

And another question: can I configure the internal volume to be applied after the convolution? Now I have a delay of approximately 1s after i press the volume key. I can't use the soundcards volume control because it is channel independent.

Hopefully someone can help me.

Thanks and best regards
Andreas

The 2.1 output setting actually already is "2.1 in a 5.1 container," and 5.1 is six channels.  It's just not labelled that way in the UI.  JRiver does it that way for compatibility because the sub is typically the fourth channel in a conventional channel ordering, not the third, and no receivers support 3-channel output.  So the default 2.1 setting outputs 6 channels, but the last three are blank.  

I personally use 2.1 for a Bi-amped system with a sub, so I can confirm it gives you six channels.  It's just labelled in an inconsistent way.

As for higher channel counts (once your system is complete), all channel counts higher than 7.1 use 7.1 as a mixing target.  So if you choose 10 channel output, what that really means is 7.1 in a 10 channel container.  If you choose 12 channels, it's really 7.1 in a 12 channel container.  Again, just a labelling issue.  If you want 5.1 in a 10 channel container, you'll need to do some custom downmixing in PEQ.

It would be nice to be able (in the future) to specify the mixing target and container size separately, but you can easily do what you need right now with the tools available.

[mr.os]

As for higher channle counts (once your system is complete), all channel counts higher than 7.1 use 7.1 as a mixing target.  So if you choose 10 channel output, what that really means is 7.1 in a 10 channel container.  If you choose 12 channels, it's really 7.1 in a 12 channel container.  Again, just a labelling issue.  If you want 5.1 in a 10 channel container, you'll need to do some custom downmixing in PEQ.

This helps me a lot. With this i can build the system.
The custom downmixing on the surround speaker shouldn't be a problem.

My complete plan is the following:
2x separate compensated subs
2x 2-way mains
1x 3-way center
2-3x one-way surrounds

So I have a 5.1 or 6.1 system and need 11 - 12 independent output channels.

It would be nice to be able (in the future) to specify the mixing target and container size separately, but you can easily do what you need right now with the tools available.

Of course that would be nice. Much less confusion.


But my 'problem' with the volume config stays. Can it be implemented after the convolution?

[mwillems]

But my 'problem' with the volume config stays. Can it be implemented after the convolution?

I've never experienced your issue when using convolution filters.  Changes to the internal volume happen for me more or less instantaneously.

Do you have unusually long filters, or maybe an underpowered CPU?  To my knowledge there's no way to change the order of operations with regard to internal volume.

[mr.os]

The CPU shouldn't be a problem (video and everything else works flawlessly at the same time). The filter are 'standard' ones from Acourate.
The delay is ~1s (maybe a bit less) - but significant and a little bit annoying.

Maybe I get the volume control from the soundcard working for all channels at the same time - but I don't think so.

[mattkhan]

The CPU shouldn't be a problem (video and everything else works flawlessly at the same time). The filter are 'standard' ones from Acourate.
The delay is ~1s (maybe a bit less) - but significant and a little bit annoying.

Maybe I get the volume control from the soundcard working for all channels at the same time - but I don't think so.

FWIW I use acourate, no problems with response time to internal volume changes.

[mwillems]

The CPU shouldn't be a problem (video and everything else works flawlessly at the same time). The filter are 'standard' ones from Acourate.
The delay is ~1s (maybe a bit less) - but significant and a little bit annoying.

Video is usually constrained by the GPU, not the CPU; in the interest of diagnostics and ruling things out, could you post your JRMark (found under help-->benchmark), and the length of your filters in taps?

[mr.os]

I have 2755 JRMarks. My filters have 65536 tapps.

But is a volume change so challenging that so much CPU power is needed?

[mwillems]

I didn't think that your computer was struggling with the volume change, I thought your computer might be struggling with the convolution so doing anything else was taking a minute.  Very long filters are hard on a CPU.  With that kind of JRMark though, that's probably not the issue.  The way to find out is to look at the convolver tab in DSP studio while music is playing.  At the bottom it will say "N x realtime" where N is how many times faster than realtime your computer is completing the filter.  I typically have between 30x and 80x on my filters on my main machine (depending on the filter length); if you have a very low number, that might suggest that your computer is struggling with the convolution.

A convolution filter of that length would introduce about 3/4 of a second of delay at 44KHz, so you may just be seeing the natural delay of having a filter that long.  That said, I've run filters that were 30k taps long without any lag, so I'm not sure what the story is.  What are your soundcard buffer settings like?  Is there any lag when adjusting internal volume without convolution running?

[mr.os]

I will have a look. But maybe it needs some time, because now I have to troubleshoot my amplifier at first.
But I know now that it is not possible to change the time where the internal volume is applied. So I will have a look at the buffers and so on. Maybe I find a solution.

Maybe it is the filter length - then it is so. Or I try shorter filters.

Many thanks to all. I will reply when I found a solution or the cause.

[mattkhan]

FWIW I get about 7-10x realtime with filters of the same length (haswell i3 system)

[mr.os]

I think I had 6-8x with an A10-6700.
Maybe I need a new CPU for the HTPC. Or shorter filters (I think this  is more realistic )