A comparison of convolution engines

[gnarly]

Thanks guys. Like i said at start, i know my question veers off topic....I'll let it lie after this, ....unless others continue it, where a reply would be appropriate.

I've found I can keep track of more complexity with separate instances, than I can navigating through a file structure.
It's a convenience/habit feature for me, not a necessary one.

The $286 cost of Acourate isn't an issue. Functionality and ease of use are the priorities.
It's less expensive than the FirDesigner I'm using;


I was ok with the one time perpetual...until I've been finding out it has a life span of only about 1.5 to 2 years before you need a new version or subscription to maintain full functionality with other software it is designed to work with.
This is the sole reason I'm asking about filter design software. I do not have $300 a year to put into this. (and wouldn't even if $ were no issue)

Keith, like i said earlier..I've searched a lot...most of what i find is about room correction...with not so much emphasis on FIR files at the speaker builder level.

 

[3ll3d00d]

Probably because room is the fuzzier part as well as being more commonly needed. In terms of software, I suppose it depends what you want to correct and how you want to do it. RePhase might be another option for example.

 

[Keith_W]

Thanks guys. Like i said at start, i know my question veers off topic....I'll let it lie after this, ....unless others continue it, where a reply would be appropriate.

I've found I can keep track of more complexity with separate instances, than I can navigating through a file structure.
It's a convenience/habit feature for me, not a necessary one.

The $286 cost of Acourate isn't an issue. Functionality and ease of use are the priorities.
It's less expensive than the FirDesigner I'm using;
View attachment 286415

I was ok with the one time perpetual...until I've been finding out it has a life span of only about 1.5 to 2 years before you need a new version or subscription to maintain full functionality with other software it is designed to work with.
This is the sole reason I'm asking about filter design software. I do not have $300 a year to put into this. (and wouldn't even if $ were no issue)

Keith, like i said earlier..I've searched a lot...most of what i find is about room correction...with not so much emphasis on FIR files at the speaker builder level.

Fortunately both Acourate and Acourate Convolver have lifetime licenses. Acourate does not stop working, you can keep using it for all eternity if you want to. But if Uli upgrades Acourate to a new major version, then you have to pay an upgrade cost if you want it. I think Acourate came out in 2005, and Uli created v2.0 in 2019. So major upgrades are few and far between, and in the meantime you get minor improvements in useability. And the cost of the upgrade was €135 which was very reasonable considering how long i've been using it (since 2016). I could have decided not to upgrade if I didn't want to, and Acourate would have kept on working.

I hate subscription licenses, fortunately none of the Convolvers listed here have that model, apart from Roon (and I don't think anybody buys Roon purely as a convolver). And I was not aware that any filter designer had that model either, until you pointed that out about FirDesigner.

 

[HerbertWest]

Fyi, HLConvolver comes with HLConvolverHost that can also host any/multiple VST3 plugins. You can wire up the plugins pretty much anyway you want...

Wrt VB-Virtual Audio Cable issues, you may want to evaluate this alternative: https://vac.muzychenko.net/en/

Confirmed I use penteo VST upstream of the convolver VST module within the host.

 

[etc6849]

Wow, there are a ton of options now! This thread is great. Back when I tri-amped my system and added multiple subs there wasn't a lot of discussion on this topic.

What I'd like to explore doing is to move my 5.1 system (5 actively tri-amped speakers and 5 subs) to a Windows 10 software based solution that uses normal DACs attached to an RME HDSPe AES sound card (16 channels out). I would need to add a second HDSPe AES card, but RME's software should allow this without issue. This would give me 32 digital channels in and out. One of the RME sound cards would act as a master clock, ensuring that all AES outputs going to the individual DACs would always stay insync.

The convolver software would need to allow for 20 channels of DSP for my 5.1 setup, but I really I'd like a solution that will allow for more speakers later for Atmos decoding. For now, I'd like to eventually replace my three speaker processors (each has 4 inputs x 8 outputs), and just use DACs. I'd want full DSP capability such as: FIR filters, time delay and level. Plus a virtual soundcard driver that is a must.

Which one of these is the most stable that will definitely never ever allow unfiltered audio to reach my tweeters? Are any of these stable enough to where they show up as a virtual sound card that works well with all Windows programs (Netflix and Youtube in browser, JRiver, etc..)?

It would seem if my motherboard's BIOS makes noise during boot up, this could blow my tweeters? What about as Windows 10 loads -is there ever a risk of outputting unfiltered sound to my drivers then?

 

[Keith_W]

What I'd like to explore doing is to move my 5.1 system (5 actively tri-amped speakers and 5 subs) to a Windows 10 software based solution that uses normal DACs attached to an RME HDSPe AES sound card (16 channels out). I would need to add a second HDSPe AES card, but RME's software should allow this without issue. This would give me 32 digital channels in and out. One of the RME sound cards would act as a master clock, ensuring that all AES outputs going to the individual DACs would always stay insync.

If you need a second RME HDSPe AES card, I happen to have one lying around.

The convolver software would need to allow for 20 channels of DSP for my 5.1 setup, but I really I'd like a solution that will allow for more speakers later for Atmos decoding. For now, I'd like to eventually replace my three speaker processors (each has 4 inputs x 8 outputs), and just use DACs. I'd want full DSP capability such as: FIR filters, time delay and level. Plus a virtual soundcard driver that is a must.

Nearly all the convolvers mentioned should be able to output 20 channels, except Roon. But if you use Roon, you can set it to output to one of the other convolvers anyway.

For a virtual soundcard driver, look at the row titled "WDM to ASIO converter". I think it is a serious negative for convolvers to lack this function because I have found that VB Cable, although it is free, is extremely unreliable. The problem I had when I tried it is latency - it drops packets, and then eventually it drops so many packets that sound starts to stutter and eventually the music stops. All this happens within 4-5 minutes, and then a reboot is required. Other people have had no problems with VB-Cable, but it does not work for me. There are alternatives to VB-Cable, I believe some of them are mentioned in this thread (I did not go through the whole thing to check).

Which one of these is the most stable that will definitely never ever allow unfiltered audio to reach my tweeters? Are any of these stable enough to where they show up as a virtual sound card that works well with all Windows programs (Netflix and Youtube in browser, JRiver, etc..)?

All the convolvers I have tested are stable, and provided you set them up correctly, they all work as advertised. The convolvers I have personally tested are: JRiver, Acourate Convolver, Hang Loose Convolver, CamillaDSP, Roon, and HQPlayer. I own licenses for JRiver, Acourate Convolver, and HQPlayer.

It would seem if my motherboard's BIOS makes noise during boot up, this could blow my tweeters? What about as Windows 10 loads -is there ever a risk of outputting unfiltered sound to my drivers then?

Although theoretically there should be no sound reaching your speakers until Windows finishes boot and loads the convolver, I would not take the risk. I only turn on my amps when the PC has finished booting up and JRiver is loaded. When turning off the system, I reverse the order - amps off first, then DAC, then PC.

 

[etc6849]

Hi Keith, thanks for the reply.

JRiver has a virtual sound card included, but I'm not sure how stable it is and haven't used it. One problem with JRiver I have is it hangs up sometimes when I try to close it. My theater has an older version of the software though, but I can't imagine relying on it not to ruin my speakers' drivers.

From what you're saying, it seems there is no all in one reliable solution that includes filter design software, convolver and virtual sound card driver. I'm sure eventually we'll see a solution. I have to think most AV Processors are running linux internally. Perhaps in time, we see a custom linux version dedicated to audio that just works and offers FIR filtering and even Atmos decoding.

The Xilica XD4080's I use sound great, but have limited number of FIR taps. Sound quality wise, I am pretty happy with them though:

https://cn.xilica.com/wp-content/uploads/2022/01/210630054009_XIL003_Specifications_XD.pdf

If I want to try different DACs, I could always use the digital outputs from the XD4080.

For a virtual soundcard driver, look at the row titled "WDM to ASIO converter". I think it is a serious negative for convolvers to lack this function because I have found that VB Cable, although it is free, is extremely unreliable. The problem I had when I tried it is latency - it drops packets, and then eventually it drops so many packets that sound starts to stutter and eventually the music stops. All this happens within 4-5 minutes, and then a reboot is required. Other people have had no problems with VB-Cable, but it does not work for me. There are alternatives to VB-Cable, I believe some of them are mentioned in this thread (I did not go through the whole thing to check).

 

[Keith_W]

JRiver has a virtual sound card included, but I'm not sure how stable it is and haven't used it. One problem with JRiver I have is it hangs up sometimes when I try to close it. My theater has an older version of the software though, but I can't imagine relying on it not to ruin my speakers' drivers.

I have found JRiver to be absolutely reliable, and I have run almost every version of JRiver since MC15 (they are currently on MC31). The only downside is that it uses .CFG files. If your filter design software does not output these (my early version of Acourate that I was using at the time didn't have this feature), then you will have to write it yourself. There is a debate earlier in this thread about .CFG files. The biggest danger with ALL convolvers involves setting them up and assigning channels properly. The best way is to disconnect the speaker, play a test tone, and verify that the amp is outputting signal to the correct driver using a voltmeter. Only then should you connect the speaker.

And BTW, if you are worried about destroying your tweeter, you should consider putting a high pass capacitor in series with your tweeter to protect it.

From what you're saying, it seems there is no all in one reliable solution that includes filter design software, convolver and virtual sound card driver. I'm sure eventually we'll see a solution. I have to think most AV Processors are running linux internally. Perhaps in time, we see a custom linux version dedicated to audio that just works and offers FIR filtering and even Atmos decoding.

No, Dirac Live includes everything that you mentioned. The only downsides of Dirac Live is the expense, the lack of more advanced features, and (this is the real deal killer for me), it outputs proprietary files. Acourate and Audiolense make their own convolvers, which you can purchase at extra cost, but both lack virtual sound card drivers.

I disagree with "no reliable solution", I certainly did not say that. I can vouch for the reliability of the products that I have licenses for and the ones I have personally tested. I have been running my system in this configuration for 7 years, and not once have any of the convolvers I have used over the years sent audio to the wrong channel once it has been set up correctly. If it happens, it is not the fault of the convolver or the playback software, it is usually the mixer software which has been inadvertently assigned to route one channel to another.

 

[etc6849]

Thanks again! These are all great points and information to consider. I will have to give some of these software options a try and test them out.

 

[JG2020]

I am using Acourate Convolver (AC) in combination with VB-Audio Matrix
[ https://www.audiosciencereview.com/forum/index.php?threads/vb-matrix.48989/ ]
in a 3-way active configuration. AC 'sees' the Virtual ASIO in VB Matrix as a soundcard and once setup you can then route to/from any other Windows WDM or ASIO devices. It appears that VB Audio Matrix has replaced VB-Cable.

 

[3ll3d00d]

Which one of these is the most stable that will definitely never ever allow unfiltered audio to reach my tweeters?

"Definitely never" is an extremely high bar, taken literally no software can possibly clear such a hurdle.

 

[somebodyelse]

"Definitely never" is an extremely high bar, taken literally no software can possibly clear such a hurdle.

Most hardware can't either.

 

[dualazmak]

Hello @etc6849,

Based on my limited experiences and actual implementations, today I could find and spare my time responding to your similar inquiries here; please visit my post #861 here (top of page 44) on my project thread.