Print

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

[whoareyou]

Hi.  I originally posted this in https://yabb.jriver.com/interact/index.php/topic,140273.msg973023.html#msg973023 but as it may be a different issue, I'm reposting under separate thread.

The attached convolution filter (it's a big one), has a channel for a 2nd subwoofer, so it's 7.2.4 configuration.   The SW is added last channel after the 7.1.4 channels of the standard layout. 
Edit: I couldn't attach as file seems to exceed maximum size.  Please let me know if you'd like the file to troubleshoot issue and I'll e-mail you link to download.  Thanks.

This convolution file works with JRiver convolution, but will not work with HLConvolver VST3.  Problem is that HLConvovler won't recognize the channel layout and ignores this convolution.   After speaking with developer of HLConvolver, he feels this is possibly JRiver issue requiring same change that was made to resolve the 4-way digital VST3 routing issue in the linked thread.

Thanks

[Matt]

For the next build:
Changed: Added speaker layouts for 13 and 14 channel VST3 output.

We were using Steinberg::Vst::SpeakerArr::kEmpty for that high a count before.

Thanks.

[whoareyou]

Thanks. 

[whoareyou]

With the new option selected, the channel processing is somehow not right. 

HLConvolver VST3 now accepts and processes configurations of 14 channel, 7.1.4 + 1 extra or 7.1.4 + 2 extra, but various channels are ignored and not being processed by the convolution.  Each configuration seems to result in a different set of missing channels.  I’m not sure how to best describe it, but there is something slightly off with this approach.     Thanks
 

[Matt]

Thanks for your help.

Coming next build:
Changed: Revised the layout of 12 channel VST3 to be the standard 7.1.4 and 13 channel to be 7.1.4 + extra LFE.

[Matt]

Please test and let us know:
https://yabb.jriver.com/interact/index.php/topic,140611.0.html

Thanks.

[whoareyou]

I think this is working, but need to speak with HlConvolver developer.  When playing test audio, HLConvolver's meters indicate silence on some of the channels, however the meters seem to be wrong as I do hear the proper test track.  I'll see what Mitch has to say.  Thanks.

[whoareyou]

I need to regroup and figure out how to best test all of this or if I can.  There is something just not right, but at moment I'm not sure of testing approach or what exactly is working and not working correctly.

In addition to the UI inconsistency for 7.2.4, even UI for 7.1.4 seems wrong when playing 5.1 source.  Instead of showing that the back rears are active, it indicates that the sides are active, and they are not.  This all displayed / worked properly in last official .63 release.  Weird thing is that the rears are active even though UI indicates the sides should be active.  So, I am very confused.

At this point (and especially since I seem to be only one requesting this) and unless anyone has different idea, perhaps changes should be backed out for next release?   I'll make due with one sub.  Not end of the world.  For now I've gone back to .63 and will do some more testing with .67.
 

[mattkhan]

The way I tend to test such things is to create a wav which has a sine wave on each channel centred on a different frequency then play it and record the output as a wav (use the disk writer audio output), compare input with output to see what ends up where. Alternatively use band limited pink noise if the input channels are less than output where the band targets a particularly crossover if it is an active system.

[Matt]

I can share the code that's being changed.

Code: [Select]

else if (m_afInput.GetChannels() == 12)
Arrangement = Steinberg::Vst::SpeakerArr::k71_4;
else if (m_afInput.GetChannels() == 13)
Arrangement = Steinberg::Vst::SpeakerArr::k71_4 | Steinberg::Vst::kSpeakerLfe2;

Before this change I had added like this:

Code: [Select]

else if (m_afInput.GetChannels() == 12)
Arrangement = Steinberg::Vst::SpeakerArr::k111;
else if (m_afInput.GetChannels() == 13)
Arrangement = Steinberg::Vst::SpeakerArr::k130;

And before that we simply returned Steinberg::Vst::SpeakerArr::kEmpty so I think it didn't work at all with that high a channel count.

I've been trying to avoid creating a dialog to allow a user to dial this in because it would be so dense.

Thanks.

[whoareyou]

After further correspondence with HLConvolver developer, there seems to be a defect with the UI visualization.  I can test by listening to what is played, but the UI is not correct.  This is what has been confusing me.  I'm going t circle back and solely listen to audio and ignore visualizations completely.

[whoareyou]

The way I tend to test such things is to create a wav which has a sine wave on each channel centred on a different frequency then play it and record the output as a wav (use the disk writer audio output), compare input with output to see what ends up where. Alternatively use band limited pink noise if the input channels are less than output where the band targets a particularly crossover if it is an active system.

I had been doing similar, but output was not matching visualization.  Since I never had issue with visualization in past, I was trusting that it was correct even though it didn't match output. 
Then when I tried to test outside of JRiver to confirm visualization I couldn't listen to output because my DAC'S WDM driver (Merging Hapi) seems to have a defect, and won't allow for more than 12 channels. 

[mattkhan]

I can share the code that's being changed.

Code: [Select]

else if (m_afInput.GetChannels() == 12)
Arrangement = Steinberg::Vst::SpeakerArr::k71_4;
else if (m_afInput.GetChannels() == 13)
Arrangement = Steinberg::Vst::SpeakerArr::k71_4 | Steinberg::Vst::kSpeakerLfe2;

Before this change I had added like this:

Code: [Select]

else if (m_afInput.GetChannels() == 12)
Arrangement = Steinberg::Vst::SpeakerArr::k111;
else if (m_afInput.GetChannels() == 13)
Arrangement = Steinberg::Vst::SpeakerArr::k130;

And before that we simply returned Steinberg::Vst::SpeakerArr::kEmpty so I think it didn't work at all with that high a channel count.

I've been trying to avoid creating a dialog to allow a user to dial this in because it would be so dense.

Thanks.

Seems you changed the mapping

L R C Lfe Ls Rs Sl Sr Tfl Tfr Trl Trr.

To

L R C Lfe Ls Rs Tc Tfl Tfc Tfr Trl Trr.

[Matt]

Yeah, using Steinberg::Vst::SpeakerArr::k71_4 seems right since that's what we call it in MC now.

[mattkhan]

It would be useful to post all of those mappings on the wiki

But still if the channel layout changed then presumably it explains what the op is hearing now

[whoareyou]

Since all of this, HLConvolver's developer has confirmed there is some kind of defect with the visualization.
Based on that info, I'm guessing the newly added Steinberg parameters exposed a defect in the plugin's visualization. Anyway, this visualization issue threw me for a loop.
I tried to confirm what was going on by hooking stuff up external to JRiver but ran into another defect where windows wouldn't recognize more than 12 channels with my DAC's WDM driver. 

I need to retest, confirm channel output and ignore the visualization. 

I think the approach of a dialog sounds good as there will always be some new configuration.  Just a matter of time before someone adds a 3rd sub. 

I'll retest paying attention ignoring the visualizations and confirm channel output for 4.0, 5.1, 5.0, 7.2.4, and 7.1.4.  Might take a bit of time, but now that I understand the issues I think it should go better.  Thanks.

[whoareyou]

Hopefully this will be last post in a while on this subject.  I tested with 4.0, 5.1, 5.1.4, 7.1.4.  Audio plays via correct channels. 

Per following post when Matt first introduced the Atmos mappings: https://yabb.jriver.com/interact/index.php/topic,139490.msg968285.html#msg968285

This seems to be a VST UI issue.

PS: Some type of user configurable dialog would probably be good for all of this.  Let me know if you need any help with it and I could possibly help out with testing.