Dead in the water with JRiver

[Bob Katz]

Perhaps you all can help. There are so many permutations of patches that I don't know where to turn. Yes, it's a multi-vendor environment and I could put this in the JRiver forum, but maybe I should start here with friends  :-).

For starters, I'm doing 2.0 with a pair of small speakers crossed over to a pair of subs. I know, this is what Bernt calls "2.2"  :-).  Then later I'll conquer 5.1 once I see how this EQ, digital XO and convolution thing is performing.

1) My first discovery is that the .cfg file which was generated by Audiolense is simply a short wordpad file with pointers to the real .cfg file which is in the corrections folder right next to it. And even that is just a short wordpad file. I expected it to look like a Matlab set of multitap filter coefficients, multipliers and delay indicators. So I don't know how JRiver gets its coefficients anymore. I thought it was frowned upon to use a WAV file

2) I just started with JRiver today and the best I can get is high pass information out of the small speakers and nothing out of the subs. 

3) The best I can get from JRiver is internal playback using music files, but I haven't the slightest idea how to tell JRiver to play an external source through my ASIO card, through the convolution and then out the ASIO card. Yes, I DO know how to digitally patch signal into the ASIO card and I DO know how to use the ASIO card's mixer if I need to, but I didn't think I would have to, and I still couldn't figure out how JRiver would insert the convolution anyway. 

4) I can see the status message in the JRiver setup window for convolution when audio is playing. It does show the correct address of the .cfg file that Audiolense provided. Doesn't tell much else for debugging purposes though. 

5) JRiver doesn't apparently have any way to reassign speakers to channels except with a primitive "rotation" function. Where you can increment from 0 to 2 to 4, etc. until you get sound. It doesn't seem to have a configuration to allow true stereo subs for bass management, either. Is this really the case? I could repatch Audiolense to feed, for example, channels 1&2 for the small front speakers, skip channels 3&4 (normally used for C/LFE), skip channels 5&6 (normally used for SL/SR), and feed the low pass information to 7&8, but I'll be darned if I could figure out how to tell JRiver that the low pass path in the .cfg file should be redirected to outputs 7&8 of the card. 

Is it possible that no one using JRiver so far has tried to do this with true stereo subwoofers?

6) If I have to reroute in JRiver, please tell me how to change the numbers in the existing .cfg file so I don't have to take measurements all over again. Maybe I have to reset JRiver to some kind of 8 channel mode to begin with, but it multiplies the number of things I have to set up and get ready just to debug the system initially. I was really hoping to evaluate a simple stereo convolution and XO setup to begin with. 

7) Note that I'm beginning to realize that even when we master all of the above, switching back and forth between playing files from JRiver or playing external sources (such as a CD player with digital output, or the Logitech Transporter with Digital output, or my Sequoia DAW...) is going to require my switching the Lynx card from internal to external sync as well. If the card is left on external sync, JRiver tries to play files at different sample rates but it can't and the pitch comes out all wrong!

Love it...   Don't you love being a geek  :-).

BK

[Bernt Ronningsbakk]

I tested the automatic filter sample rate switch in JRiver in parallel with issuing 4.6. Eventually I learned that this switch didn’t work as I expected.

 

You need to pick a config file from the CORRECTION FILES folder instead of the files in the CORRECTION are filterlist files.

 

1 &2: If you use Sourceforge Convolver in an environment that can handle automatic format and sample rate shift, SfCon will pick the config file that fits with the input format. But in JRiver things works a bit different. If you enable the sample rate switching filter support in JRiver, the program will look for another config file with the same name, except that 441 is replaced with 96, when you switch from a 44.1kHz recording to a 96kHz recording. This wasn’t quite as I expected when I did the final preparation on 4.6 and I am sure that this question will come up a few times in the near future.

 

4: A rough guess was that you only had straight pass-through without correction because the JRiver convolver couldn’t find the filters.

 

3: I’d like to hear how that is done too.

 

5: I’m not sure how JRiver does remixing. If you set up JRiver for enough channels to cover any input format and the number of output channels used, and also specify that JRiver should play silence on unused channels, you can downmixing and bass management in Audiolense.

 

 

 

Kind regards,

 

Bernt

--
Audiolense User Forum.
http://groups.google.com/group/audiolense?hl=en?hl=en
To post to this group, send email to audio...@googlegroups.com
To unsubscribe, send email to audiolense+...@googlegroups.com

[mojave]

1.  You use the filters in the CorrectionFiles folder. If you keep the default ending of the file name such as "xxxx2.0_441 or xxxx5.1_48" then JRiver will automatically switch files to match channels and sample rate if you check the "Automatically switch filter . . ." box in JRiver's convolution engine. Audiolense and JRiver use the WAV file.

5. JRiver first requires you to set the number of channels in the Output Format DSP. This opens the number of channels that can be used for playback. I always leave mine set to 7.1. You can playback 2.0, 2.1, 5.1 or 7.1 through this. You don't need to reassign speakers to channels in JRiver. I let Audiolense do this for me. You can do it in JRiver by using the PEQ and Mix Channels. However, you can't rename channels. JRiver allows you to do active crossovers, PEQ, bass management, mix channels, etc. Some users use JRiver for everything you want to do with your setup and use manual PEQ filters instead of convolution.

When I setup Audiolense, it defaults to having stereo subwoofers on the last two channels. However, when you go to make a measurement you can change the default channel order. Once the measurement screen is up go to Advanced Settings and check Enable Output Settings Override. Now in the Output Channel section by each speaker/driver/sub you can change the channels. I use a Steinberg UR824 for output and actually start with channel 10 since the first 9 go through the built in mixer. This way I bypass the mixer. I then use 10 and 11 for mains, 12 and 13 for subs, 14 and 15 for rears, and 16 and 17 for sides. I am using a phantom center so the center output is actually used for my left subwoofer. 

In JRiver's Options > Audio > Audio Output > Output Mode Settings I set the channel offset to 10 so that JRiver starts playback with channel 10 on my DAC.

If you don't want to redo your cfg files, you can use the Parametric Equalizer in JRiver and Mix Channels to move info from one channel to another. For example, you can have your output set to 7.1 and use Mix Channels to move channel 3 to 7 and channel 4 to 8. Remember, though, you have to use the original names rather than channel numbers. 

3.  In File > Open Live you can choose to use WASAPI loopback or ASIO Line-in. Perhaps these are what you are looking for.

[Bob Katz]

This is a great page with replies by Bernt and Mojave leading the way. Maybe I'll get correct playback by the end of tonight!

We have to start adding selections like these to an FAQ. Please visit the primitive WIKI page that I established and see if it's got the structure to allow us to do this and expand it step by step. If you want to participate in the wiki project, please send me a message at bobkatz24bit[atsign]digido.com and I'll set you up as an administrator for the wiki page. 

BK

[mojave]

I've got a meeting for the rest of the evening so I won't be of much help. I'll check the wiki out later.

[Brad]

Hi Bob,

I tried that email address for helping with the wiki but my message bounced "no such user".

Thanks for the posts to make us think of a few new ideas.

Brad

[Bob Katz]

Oops, I made a mistake, I meant to give my gmail address, which is

bobkatz24bit[replace left bracket through the right bracket with an @ sign]gmail.com

Sorry about that!!!! I'll put a note about it on the home page of the wiki. 

Bob

[Bob Katz]

I'm making progress, big progress but there are still a few obstacles. I can play files and audio CDs and DVDs now in JRiver through the Convolution engine. The sound is remarkably good, very transparent, with good depth and purity of tone, and I haven't even finished the final tweak of the HF target at 20 kHz! One remarkable thing is how much wider the sweet spot is now in stereo, and I have no idea how that is happening. What I mean is that the image is not pulling as rapidly to the left speaker or the right speaker if someone sits left of center or right of center. It's some fringe benefit of the TTD correction? Perhaps because it's cancelling or reducing a few reflections that were causing some slightly ambiguous imaging. That's pretty hard to accomplish in this room which is so well treated that the log ETC BEFORE correction is better than 15 dB down all the way past 30-40 ms. It's really 20 dB down with two tiny spikes to -15 dB. And it's clean to -20-25 dB in the first 10-15 ms. Again, before correction. 

I'll have something to say later about some measured anomalies in the log ETC due to the TTD calculation that may or may not be audible defects. Objectively speaking, the post correction simulated ETC is slightly worse, but it's still excellent in terms of the ratio between the peak of the impulse and the highest spike in the first 20 ms. 

But first the weak links:

1) I can't get the automatic config file sample rate business to switch. The status window in the convolver is always showing whatever config file I chose first. And the audible difference when using the wrong config file is quite obvious. So in the meantime I configured JRiver to always upsample or downsample everything to 96 kHz and I'm using a 96 kHz config file. I don't think that JRiver's sample rate conversion sounds quite as good as the best that I've heard (and I know what the best sounds like!) so I'd like to bypass that in the near future. But when playing 96 to 96 without realtime conversion it sounds very nice. 

2) I can't get live playback to work. I get a failure message from JRiver when I try. It says that you should use a different interface for the input (what a drag) and I tried that as well, but it just is not cooperating. I should ask on the JRiver forum, but if anyone here has a hint I'd appreciate it. Thanks, Mojave, for the pointer to the file menu where that option is implemented. Unfortunately, you have to tell JRiver what sample rate is coming in and that is a big drag in a production studio where I need to patch many different sources into the player which are all at different sample rates. J River should be able to detect the incoming S.R. and switch its parameters without a menu. Oh well, more work for Matt  :-(. 

3) Latency: What happens if I choose a shorter filter length? I don't see any places where the filter length choice interacts with any of the TTD parameters in the setup, so it must affect the quality of the TTD correction after a certain number of milliseconds. Would be good to get the latency down. Filter length affects latency, right?

That's it for now. I am excited about where we stand at this moment. 

[walter.fo...@gmail.com]

Dear Bob,

I already "said welcome" in another thread so I will just cut to the chase.

1) have a look at the Convolution thread I (TheLion) started over at JRivers. Lot's of information. Automatic filter switching is a rather new feature and I don't use it (because I have multiple different filter per sampling rate). I would recommend using different Zones for different content and ZoneSwitching. Therefor you can have a Stereo Zone for 44.1khz content, another for 88.2khz, another for 48khz 5.1,...

2) the live playback feature of JRiver needs alot of additional work to make it useful. This will take some time. I use Acourate Convolver for this task which is perfect for this task and an ASIO interface (but this only works with Acourate filters)

3) JRiver automatically adjusts for audio delay (buffer and filter) when you are doing video playback. If you need low Latency (eg. for live playback) use Freq. only correction instead of TTDC. The filter delay depends on "backed in channel delays" (gets larger when the necessary inter channel delays are large), the filter length (eg. try 32k instead of 64k taps) and the FFT partition of the convolution engine (which cannot be selected in JRiver).

Let us know how it works out!

Best regards

Walter

[Bob Katz]

On Saturday, January 19, 2013 9:40:08 AM UTC-5, Walter wrote:

Dear Bob,

I already "said welcome" in another thread so I will just cut to the chase.

1) have a look at the Convolution thread I (TheLion) started over at JRivers. Lot's of information.

Thanks, Walter. I read that thread partially once and then I read it again at your urging today. Much very useful information in there. 

 

Automatic filter switching is a rather new feature and I don't use it (because I have multiple different filter per sampling rate). I would recommend using different Zones for different content and ZoneSwitching. Therefor you can have a Stereo Zone for 44.1khz content, another for 88.2khz, another for 48khz 5.1,...

I'm a big fan of ergonomics and when I master audio during the day I'm constantly switching between 44.1k and 96k for example from two different digital playback sources. Navigating to another application and switching switches in my monitor chain to accomodate that while I'm also trying to be a music producer and make comparisons and listen critically---- would give me a big headache! So eventually we have to conquer the integrity of live input and automatic filter switching. I'll try to keep working with Matt and JRiver and all the excellent participants in that forum to see if we can get that going. 

 

2) the live playback feature of JRiver needs alot of additional work to make it useful. This will take some time. I use Acourate Convolver for this task which is perfect for this task and an ASIO interface (but this only works with Acourate filters)

I really don't want to spend the money on another convolver and measurement tool, especially since Audiolense appears to integrate DRC, sophisticated XO and bass management, and channel mapping/matrixing all in one tool, at 64-bit precision. It's pretty hard to beat. And Bernt is such a nice guy! While I'm waiting to overcome JRiver's weaknesses I need an interim tool for convolution and live playback. 

As a workaround I might settle on a second instance of my Sequoia workstation loaded on the other ASIO card in my computer, but what a waste of CPU resources. And use a VST plugin like Voxengo Pristine space. Or try purchasing and installing Console, http://www.console.jp/en/about/about_4.html, to run the VST, which would minimize CPU load for the time being. But I doubt the Voxengo is 64-bit and I still would have to implement dithering to 24 bit in another plugin. And it doesn't select filters for different sample rates automatically either. But it would conquer the current JRiver weakness. And switching from live input to internal playback means quitting and launching JRiver...  so I would stick with live input and switch sources into it. It's not a perfect world. But it's an adventure!  Do you know of any other convolution tools that would provide live playback and be a bit more flexible? 

I have to finish mastering a prog. rock album today and I'll get back to trying to kick the tires of Audiolense. I'm already hooked on the improved sound quality. It's a matter of sonic coherency. My current system sounds great, it's an A grade with its analog filters doing analog-level correction and crossover. But I can already tell that Audiolense will bring it to an A+ level and also introduce a new era of control and consistency. Those damn analog parametrics are a drag to adjust---what took me a day to get close to right now can only take an hour or two and get me much much closer. Audiolense is compensating for a subwoofer time lag that was always bothering me a bit and integrating the system very well. Still have to decide and listen critically to either frequency-domain correction or true TTD. We shall see!

Best wishes and thanks,

Bob

[Brad]

Just use Convolver VST. It works perfectly with Audiolense config files and filters.

Convolver VST

Do you know of any other convolution tools that would provide live playback and be a bit more flexible? 

Bob

[Bob Katz]

Fabulous. I may try it today! Off hand, Brad, do you know if it takes the initial config format that's in the first folder Audiolense saves to, or the second config format that's in the second folder. As you can see, I don't have the names of the terms on the tip of my tongue (yet).

Can you see why the mute and dim buttons in Avocet are so useful  :-).

Thanks,

Bob

[Walter]

Bob,

the Convolver VST by itself doesn't provide live playback - but you can integrate it into your DAW setup via VST. So live playback may be in reach with it. It also allows setting the FFT partition i was talking about earlier in order to lower latency.

It actually handles the "Audiolense config files" perfectly as Audiolense uses the Convolver syntax ;-) Matt simply used the very same syntax for his convolution engine for simplicity reasons.

With the ZoneSwitch feature in JRiver you can have automatic filter switching - I didn't make this point clear enough I guess. I can define rules which Zone is automatically switched to/used. So you can build Zones with different filters and set rules like "for 44.1khz content switch to Zone XY". Once defined the switching is completely automatic.

[Bob Katz]

Dear Walter:

Thanks. Oy--- I didn't see the forest for the trees. I didn't realize that the convolver included in JRiver is a standard VST plugin which I could use in another host to do live input until JRiver gets that straightened out! Silly me.

Your trick for switching filters automatically sounds cool. Of course I know nothing about JRiver's zone switching and I'll have to step into that learning curve, maybe next weekend, or maybe the week after  :-). Mary and I are going out to the movies tonight.

Take care and thanks again,


Bob

[Brad]

The only thing I've ever used is the .cfg file in a folder called correctionfiles. Works fine with convolverVST. Where's the other one?

Brad

[Bob Katz]

Walter, sorry. I did a search for the convolver plugin and all I see are some abbreviated .dlls in the JRiver plugins folder. Do you know which one is the convolver? I want to try it in my Sequoia DAW to get live input going.

Thanks,


Bob

[Brad]

I'm not sure that it is. I think you need ConvolverVST. I'm don't know if the JRiver convolver can be used as a VST plugin.

Brad

I didn't realize that the convolver included in JRiver is a standard VST plugin............

[Bob Katz]

Maybe it's a direct x plugin. I'll have to play with the plugins in the folder and see if I can psyche which one is the jriver convolver

[Mitch Global]

Hi Bob,

 

I have used http://convolver.sourceforge.net/vst.html before as a standalone app using http://www.hermannseib.com/english/savihost.htm

 

With Savihost, you can run convolverVST as an.exe and patch into your monitoring chain.

 

While convolverVST has the capability to auto sample rate switch: http://convolver.sourceforge.net/config.html - scroll to bottom section "Filter List", I could not get it to work.  In exchanging a few emails with the software's author, he could not get it to work either.  Last release was 2006.  Maybe he would like to open source it?

 

Regards,

 

Mitch

On Sat, Jan 19, 2013 at 8:04 PM, Bob Katz <bobkat...@gmail.com> wrote:

Maybe it's a direct x plugin. I'll have to play with the plugins in the folder and see if I can psyche which one is the jriver convolver

[Bob Katz]

There's a folder called Correction and another called CorrectionFiles. I get the impression that they are in slightly different config formats for different types of convolvers but I could be speaking out of turn. Others here are more experts on this one! I do know that when I attempted to use the Correction folder in JRiver it didn't work right, and someone kindly guided me to the CorrectionFiles folder. I'd have to study the config format more carefully to figure it out. Bernt?

BK

[Walter]

Hi Bob,

sorry, I guess I caused more confusion than actually being of help for you.

I didn't mean to suggest the JRiver Convolver is an VST plugin. If we talk about "ConvolverVST" we mean  http://convolver.sourceforge.net, which is availbale as VST, DS and DMO plugin. I guess your DAW setup already provides live feedback cabability AND support for VST plugins. Therefor it is a way to get convolution into your present DAW setup.

The only "issue" with the ConvolverVST is that it is limited to 32bit fp processing. Therefor I made a feature request to Matt from JRiver to make his own convolution engine tightly integrated into the JRiver 64bit fp audio engine. That's how the JRiver "Convolver" (sadly using the same name) came to be. The configuration file syntax introduced by ConvolverVST is used for Ausiolense and JRiver (for simplicity reasons Matt and Bernt didn't invent the wheel twice ;-)

This configuration files (basically doing the channel routing) are available in the "Correction File" folder. All configuration files for a particular filter (with different sampling rates, channel routing) are "linked" in the filter file you find in the "Correction" folder. This "filter file" should enable auto filter switching. But I am not sure how Matt implemented it and if it works.

Best

Walter   

[Bernt Ronningsbakk]

You can use the files from the correction folder when you use Sourceforge Convolver in an environment that supports automatic sample rate shift. With JRiver you need to use config files from the correction files folder.

 

Kind regards,

 

Bernt

 

From: audio...@googlegroups.com [mailto:audio...@googlegroups.com] On Behalf Of Bob Katz

Sent: Sunday, January 20, 2013 6:11 PM
To: audio...@googlegroups.com

--

[Brad]

Hi Bernt,

Is the "correction folder" new to 4.6? Will it show up in the same place as the "correction files folder"?

Sorry, I've not run a new solution for awhile.

[Walter]

Brad,

these two different folders exist as long as I can remember using Audiolense.