ASR Acourate users

[dped90]

Nah, it's pretty straightforward. I wrote a free Acourate guide and there's also @mitchco's book. My guide is more a collection of Acourate recipes with some explanation, and Mitch's book is a more thorough exploration of DSP, acoustics, and so on ... with a very easy to follow step by step Acourate guide. I am reasonably proficient with Acourate, and I can create a simple set of filters in about 30 minutes. But it would take me much longer than that to create a "refined" filter because of all the measurements involved.

Thanks for this.
https://drive.google.com/file/d/1ml4657gYFLbV3bOVHQN1yJRec3ENUdJw/view , though perhaps you should rescan in OCR format for enabling the use of the standard "Ctrl+F" (or Find) function to quickly locate information within the document. But what's described in your beginners version is the closest thing to what I'd first want to do. This is the first guide I've found that lets me dare to get my feet wet with active filters for passively crossed speakers-and also "bass management", which I have no hands on experience with yet either. So, if I understand it all after I finish reading it, it would be wonderful if I can set up the system on my own, if perhaps with some coaching, get the recommended microphone and proceed as directed.

Hopefully I will later understand enough of theory behind DSP based crossovers, and why and how using them is better than passive ones.

But later on when I have to decide on a multichannel DAC for my mains and subs, do you see any reason why Acourate software couldn't be used with the Danville DSPNexus, even though Danville and/or Hollis Labs specific its use with Audio Weaver?

Audio Weaver Sessions #1: Designing a Simple DSP Crossover - Danville Signal

Danville Signal Processing is a technology partner for high performance audio solutions. We provide SHARC-based DSP modules for a variety of OEMs, ODMs and oth

danvillesignal.com

Outboard DSP: Danville Signal dspNexus

The first in a series of reviews of outboard DSP processors available in the market and of interest to loudspeaker engineers, this article explores an extremely versatile and powerful DSP audio processor and DAC, crossover, and control center for any audio system. The dspNexus comes with...

audioxpress.com

Very grateful your encouragement and from the ground up help.

 

[Keith_W]

Thanks for this.
https://drive.google.com/file/d/1ml4657gYFLbV3bOVHQN1yJRec3ENUdJw/view , though perhaps you should rescan in OCR format for enabling the use of the standard "Ctrl+F" (or Find) function to quickly locate information within the document.

Ctrl-F for find works depending on what PDF reader you are using. May I suggest you try a different PDF leader?

But what's described in your beginners version is the closest thing to what I'd first want to do. This is the first guide I've found that lets me dare to get my feet wet with active filters for passively crossed speakers-and also "bass management", which I have no hands on experience with yet either. So, if I understand it all after I finish reading it, it would be wonderful if I can set up the system on my own, if perhaps with some coaching, get the recommended microphone and proceed as directed.

That Acourate guide only contains recipes. It does not attempt to teach you what is right and what is wrong. For the sake of brevity, the guide only tells you to correct the on-axis measurement at the listening position. But in reality you can take any measurement you want and use that as the basis of correction. The ins and outs of taking correct measurements is waaaaaaay beyond the scope of a simple free PDF, you need a book for that.

Hopefully I will later understand enough of theory behind DSP based crossovers, and why and how using them is better than passive ones.

You are looking at the wrong document for DSP based XO's, you want the other one (also free) for active speakers.

But later on when I have to decide on a multichannel DAC for my mains and subs, do you see any reason why Acourate software couldn't be used with the Danville DSPNexus, even though Danville and/or Hollis Labs specific its use with Audio Weaver?

Acourate can not be used with the Danville DSP Nexus because Acourate is a specialised FIR filter DSP tool and has very limited biquad export functionality. Biquads are not portable, they need to be generated for each specific DSP hardware. This is why:

... REW's EQ section lists all those different manufacturers. The reason is because each manufacturer has a slightly different interpretation of how the coefficients should be implemented. So if you take coefficients generated for MiniDSP and try to use it with Xilica, you will get different results. What this means is that biquads have to be specifically created for different hardware, and this requires cooperation between the software publisher and those manufacturers. Read more about it here. If you want to make filters for DSPNexus, use Audio Weaver, or you could use Eclipse Audio's FIR Designer and export to Audio Weaver.

Bear in mind that the DSP is fundamentally different. If you need a quick primer on the difference between FIR and IIR, read this thread. Acourate is FIR with limited biquad exports. REW has powerful IIR biquad functionality, but less suitable for making linear phase FIR. It is possible, but it is extremely manual and painful and I don't think that some of the things Acourate does is possible in REW. As in - I tried to use REW to make linphase filters, and I got stuck because there is simply no equivalent tool in REW. REW was designed for a different purpose, and it is great at what it does. I mean, you could theoretically chop down a tree with a hammer and chisel, but you are better off with an axe. But you wouldn't use an axe for delicate wood carving. That is what a lot of people don't understand.

 

[3ll3d00d]

You may well be able to use acourate to design an fir filter that fits in the available taps of that device, if you were just doing it for mid to high frequency and used iir for the bottom end then it should work ok. Assuming you can export to a useable format that is, I guess audio weaver might just take a raw double format.

I would think it only makes sense to do that if low latency is critical though

Having said that, audio weaver does have a whole load of functionality I haven't seen elsewhere including fir filter decimation for multirate processing so that would be a fun exercise (if you're into dsp) to get working

 

[dped90]

Having said that, audio weaver does have a whole load of functionality I haven't seen elsewhere including fir filter decimation for multirate processing so that would be a fun exercise (if you're into dsp) to get working

Yeah, the trouble is that I'm a total DSP newbie-and possible dummy, depending on how well I can understand Emilson's explanations. https://www.linkedin.com/in/emilson-enrique-6334a69 If not, dspNexus designer Al Clark said that I would need to hire Rich Hollis and work with him remotely to set up the whole system. I don't mind paying him but everyone says how powerful but complex Audio Weaver is. It really stinks that Danville had to mate what's likely very good sounding hardware like dspNexus with software with what may be an impossibly steep learning curve for some of us, rather than with something more like DIRAC Live.

 

[Keith_W]

Yeah, the trouble is that I'm a total DSP newbie-and possible dummy, depending on how well I can understand Emilson's explanations. https://www.linkedin.com/in/emilson-enrique-6334a69 If not, dspNexus designer Al Clark said that I would need to hire Rich Hollis and work with him remotely to set up the whole system. I don't mind paying him but everyone says how powerful but complex Audio Weaver is. It really stinks that Danville had to mate what's likely very good sounding hardware like dspNexus with software with what may be an impossibly steep learning curve for some of us, rather than with something more like DIRAC Live.

If you want Dirac Live, get a MiniDSP. I suspect the whole point of DSPNexus is to offer a product that DSP nerds will appreciate.

 

[dped90]

If you want Dirac Live, get a MiniDSP. I suspect the whole point of DSPNexus is to offer a product that DSP nerds will appreciate.

Not entirely, as it was obviously also Al Clark's intention to use premium AKM DAC chips which not only measure well but evidently sound better than the often used ESS chips. The processor was also designed with Class A biased output stage and low noise grounding and clocking schemes. I discussed this at some length with an expert on DAC design. https://www.diyaudio.com/community/members/markw4.373860/

I am very grateful for giving the math challenged like me the best chance for learning enough about DSP to use it for multi subwoofer management and to evaluate rooms and enhance their performance beyond acoustic room treatment. https://www.audiosciencereview.com/...erstanding-the-state-of-the-dsp-market.62323/

And perhaps, if I become suitably proficient, I may be able to improve my speakers' passive crossover performance, or even build software based active crossovers to biamp my system. I just hope that the quality of many of favorite recordings will still be enjoyable, as active crossovers would go further than beryllium drivers in revealing their flaws.

 

[Keith_W]

Not entirely, as it was obviously also Al Clark's intention to use premium AKM DAC chips which not only measure well but evidently sound better than the often used ESS chips. The processor was also designed with Class A biased output stage and low noise grounding and clocking schemes. I discussed this at some length with an expert on DAC design. https://www.diyaudio.com/community/members/markw4.373860/

Wellllllllll ... you could read this 542 page debate on ASR which is regrettably closed:

'How can DAC's have a SOUND SIGNATURE if they measure as transparent?'

Oct 4, 2019

I know there are people out there that think cables affect sound, which is much worse, but there really is no response to something like that, but just to smile and nod. But what about people who talk about DACS as if they were headphone drivers or speakers, and talk about the SOUNDSTAGE, IMAGING, and MIDRANGE of a DAC? I actually don't know what to say to people to not be rude. If you try explaining that a DAC isn't something that actually changes the sound, they accuse you of having "a hard-on for measurements", as if it were the measurements themselves that tell you that DACs don't...

• Willhelm_Scream
• Replies: 10,921
• Forum: DACs, Streamers, Servers, Players, Audio Interface

• 

Even if there was a tiny audible difference between AKM and ESS, the difference would be swamped by DSP. And I mean, totally and utterly swamped, like throwing salt into the ocean to make it saltier. When you start making filters, you will see what I mean. I can make filters with a measurable difference, but not be able to hear a difference. The measurement just looks nicer. You really can correct to below the limits of audibility.

I would advise you to stop worrying about DAC's, just get the DAC that you want which has the features that you need. And if you decide to ignore this advice ... with software based DSP it is possible for you to buy four DCS or MSB DAC's and gang them together to make an 8 channel DAC. But IMO it's a very expensive way to soothe an irrational audiophile anxiety. You are better off focusing your energy on learning about room acoustics and what you can and can't do with DSP.

 

[dped90]

Your positioned is certainly well reasoned. However, while audibly perceived differences between DAC chips compared using the same main board might be slight-compared to audible differences between speaker driver designs-the only way to affirm this seems to be via the following:

At least from my original question to you about material, just swapping for material in the same size isn't what would satisfy. I think you'd have to make the speakers so that they MEASURE exactly the same - say Beryllium or Flax or Titanium speakers - in every measurement that we know in today's time. Then see if people observe differences in a blind study. If a significant amount of people [not just one pair of ears like Amir's] observe differences, yet they measure the same in all known methods, then I would say you have a case for this "texture." Otherwise, if they measure the same in every way, then they are the same to our ears.

https://www.audiosciencereview.com/...s/understanding-subwoofers.63662/post-2334275

Unfortunately, such standalone (stereo or multichannel) DAC model measurement/listening reviews are not conducted using multiple listener panels here at this or probably any other forum, nor of course, for any reviews published by Stereophile, Enjoy the Music, 6 moons TAS, et al. And obviously, such DAC chip shootouts witnessed by a multiple listener panel would be 10x less likely, so no opportunities for such comparisons.

And for the record, while my overall budget is likely a fare bit larger than many DIYers, >>/=$8K DCS or MSB stereo DACs are hardly my thing.

Turning to a possibly far more important concern for systems using multichannel DACs for DSP crossed mains and/or managed subs, how serious can per channel processor output voltage be? For example, the DSPNexus balanced outputs are only ~ 4V, while those of the Hapi are at least 9V. Can use of Acourate or DIRAC Live software-even prior to acoustic correction-cause enough gain loss to require use of external preamps (UGH!!)-even with >95db SPL/w/m main speakers?

Meanwhile, my immediate concern is selecting a main horn and drivers for two or three way system above my midwoofers. https://www.diyaudio.com/community/...erred-directivity-pattern.429264/post-8075180

 

[gfinlays]

Your positioned is certainly well reasoned. However, while audibly perceived differences between DAC chips compared using the same main board might be slight-compared to audible differences between speaker driver designs-the only way to affirm this seems to be via the following:

At least from my original question to you about material, just swapping for material in the same size isn't what would satisfy. I think you'd have to make the speakers so that they MEASURE exactly the same - say Beryllium or Flax or Titanium speakers - in every measurement that we know in today's time. Then see if people observe differences in a blind study. If a significant amount of people [not just one pair of ears like Amir's] observe differences, yet they measure the same in all known methods, then I would say you have a case for this "texture." Otherwise, if they measure the same in every way, then they are the same to our ears.

https://www.audiosciencereview.com/...s/understanding-subwoofers.63662/post-2334275

Unfortunately, such standalone (stereo or multichannel) DAC model measurement/listening reviews are not conducted using multiple listener panels here at this or probably any other forum, nor of course, for any reviews published by Stereophile, Enjoy the Music, 6 moons TAS, et al. And obviously, such DAC chip shootouts witnessed by a multiple listener panel would be 10x less likely, so no opportunities for such comparisons.

And for the record, while my overall budget is likely a fare bit larger than many DIYers, >>/=$8K DCS or MSB stereo DACs are hardly my thing.

Turning to a possibly much more imporant concern for multichannel DACs used for DSP crossed mains and/or managed subs, how serious can per channel processor output voltage be in DSPed systems? For example, the DSPNexus balanced outputs are only ~ 3.5V, while those of the Hapi are at least 9V. Can use of Acourate or DIRAC Live software-even prior to acoustic correction-cause enough gain loss to require a use of external preamps (UGH!!)?

Meanwhile, my immediate concern is selecting a main horn and drivers for two or three way system above my midwoofers. https://www.diyaudio.com/community/...erred-directivity-pattern.429264/post-8075180

You're overthinking this. The difference in sound between competent DACs is way beyond human hearing perception when properly level-matched.

DAC output voltage levels need to be appropriate to the input sensitivity of the power amplifier. Most modern power amps have selectable input sensitivity, so pick what works best.

DSP has a massively greater influence on sound than the "texture" of driver membrane materials.

There are several routes for a multi-channel DAC. RME or Motu interfaces for example, or the Octo Research DAC8, which can be ordered with a customised output voltage. Noise and distortion levels for all of them are way below human hearing perception.

Build some steep slope crossovers in Acourate, linearise your drivers, time align, room correct, finalise your linear-phase FIR filters and let your jaw hit the floor with what you hear. DAC sound signature? Milligram of sea salt added to an ocean......

 

[Keith_W]

Turning to a possibly far more important concern for systems using multichannel DACs for DSP crossed mains and/or managed subs, how serious can per channel processor output voltage be? For example, the DSPNexus balanced outputs are only ~ 4V, while those of the Hapi are at least 9V. Can use of Acourate or DIRAC Live software-even prior to acoustic correction-cause enough gain loss to require use of external preamps (UGH!!)-even with >95db SPL/w/m main speakers?

That is a fair question. All DSP sacrifices SPL for linearity. So the more nonlinearity you have to correct, the more SPL you have to sacrifice. I think I said somewhere in my guide that you need to make a real effort to get your system as linear as possible before you start. This means:

- adjusting the volume of each driver so that the summation of drivers is as flat as possible
- choosing XO orders and Q's so that they sum as flat as possible (if using minphase)
- repositioning subwoofers, speakers, and listening position so that bass variation is as small as possible
- think carefully about what you are correcting. It is tempting to linearise as much of what you see as possible, but the more you linearise the more SPL you lose and the greater chance of introducing side effects. These usually show up in the time domain as ringing. Do not correct an inaudible problem by introducing an audible problem to make the measurement look better.

If your main speakers are >95dB efficient it is very unlikely that you will suffer so much gain loss that you will require external preamps. But there are a lot of variables which prevents me from answering your question for certain. These include:

- choice of music. Thanks to the loudness wars, some recordings are mastered much softer than other recordings. For e.g. classical recordings are typically >10dB quieter than other types of music.
- size of the room / listening distance. The bigger your room, the louder your speakers need to go. The further you listen, the more SPL you need - thanks to the inverse square law.
- problematic room. If you have multiple room dimensions which are similar to each other and an unfortunate choice of listening position, you will get giant room modes which require a huge volume cut.
- amplifier output power.
- power handling of your speaker. Even if you have high efficiency speakers, if you make poor choices with the bandpass you will get a lot of distortion. In the low end, feeding too much power into the speaker's Fs will make it exceed the Xmax. At the top end, don't exceed the speaker's cone breakup frequency.
- preference for playback volume

With my system, I have an overall volume loss of -4.5dB after the DSP process. My early efforts had a much greater volume loss. These days I am more aware of what I am doing and every decision I make along the way is aimed at preventing excessive volume loss.

None of what I said is Acourate specific. It applies to ALL DSP products, because the fundamental principles are the same.

 

[dped90]

or the Octo Research DAC8, which can be ordered with a customised output voltage. Noise and distortion levels for all of them are way below human hearing perception.

This I had asked Pavel Krasensky about and the numbers do show a substantial difference in noise numbers moving up to the 12v range option, which for horn speakers like mine could be quite audible. USB rather than a CAT cable would be more convenient but this looks like the safest bet.

Merging Technologies | Horus & Hapi Mic-Pre & AD/DA Converter Specifications

Horus & Hapi mic preamp and ADDA converter has some of the best specifications of any I/O on the market today. Have a look at the numbers here.

www.merging.com

Also, unlike the DSPNexus's remote, I would need to use my wireless Logitech keyboard for master volume control/mute, though I think JRiver claims it's compatible with some wireless remotes.

 

[3ll3d00d]

the mentioned diya user also seems to believe in audible differences between speaker cables, it seems fair to say a discussion with such a person will go quite differently to a discussion with most posters here

there's a comment in your linked thread that seems on the nose to me

I stopped reading your original post half way through, but it seems like you're very far into the weeds on all these details based on what other people are saying but without ever trying any of this stuff to see what you actually like?

you will get so much more out of just building *anything* and trying it in your room then running a series of threads on different forum trying to perfect what to build, seriously just build anything and try it out

re noise, it's a non issue if you manage gain appropriately and use remotely modern kit. I run my 113dB sensitive CDs straight from a UCD-180 (aka not an especially new amp) out of a antelope audio orion studio synergy core, added some inline attenuators so most of the difference is managed in analogue domain rather than DSP, totally silent with your ear pressed to it (same was true using a motu 1248 or a rme fireface 800)

 

[OCA]

there is simply no equivalent tool in REW.

All filters REW generates are min phase. But you can multiply them with their "Inv phase A" and convert them to linear phase if you need to. Question is why would you ever need to?

 

[Keith_W]

All filters REW generates are min phase. But you can multiply them with their "Inv phase A" and convert them to linear phase if you need to. Question is why would you ever need to?

It's actually "Invert Phase A" followed by "A*B" to convert it to linear phase. I did not say that REW is lacking in all tools, it lacks some other tools you will find in Acourate. Like pre-ringing compensation or excess phase compensation. When I showed you the step response and asked you if you know how to do that, your response was that it's unnatural and going to sound awful. So I gave up on asking you and tried to figure out how to do it myself in REW and rePhase. FYI, it's an intense multi-hour job if you want to do it for 8 filters, a step that takes a few seconds in Acourate.

 

[OCA]

It's actually "Invert Phase A" followed by "A*B" to convert it to linear phase. I did not say that REW is lacking in all tools, it lacks some other tools you will find in Acourate. Like pre-ringing compensation or excess phase compensation. When I showed you the step response and asked you if you know how to do that, your response was that it's unnatural and going to sound awful. So I gave up on asking you and tried to figure out how to do it myself in REW and rePhase. FYI, it's an intense multi-hour job if you want to do it for 8 filters, a step that takes a few seconds in Acourate.

Pre-ringing compensation sounds like a function forced by popular user demand. In my most recent view, you should only ever correct for excess phase of the speaker (and the minimum phase of the room). Even for a very large speaker that's never longer than the first few ms which will not produce pre-echo. I can do both in REW although admittedly EP inversion is easier in rephase and am getting my best ever results lately.

 

[Keith_W]

Pre-ringing compensation sounds like a function forced by popular user demand. In my most recent view, you should only ever correct for excess phase of the speaker (and the minimum phase of the room). Even for a very large speaker that's never longer than the first few ms which will not produce pre-echo. I can do both in REW although admittedly EP inversion is easier in rephase and am getting my best ever results lately.

If I wasn't clear, it was the EP of the speaker that I was talking about. It's a multi-hour process if you want to correct 8 channels.

And FWIW it is quite possible to get more than 20ms of pre-ringing if one is clumsy with correction, particularly bass correction. Also in my view, it is better to avoid getting that pre-ringing by being more careful with filter design in the bass region than to be a cowboy and use pre-ringing compensation (PRC) to compensate. So although PRC is a nice tool to have, I try to avoid using it as much as possible.

 

[CK.]

Hello @Keith_W and everyone,

I have used Acourate quite a few times with Roon, but yesterday I went to edit the target curve in an existing project, run again macros 0-5 and the resulting convolution filter when imported to Roon produced a highly distorted loud sound that was clipping with a -10dB headroom. I have attached the relevant files.

I have imported the wav file to APO and seems fine, also my previous Roon filter is working fine.

Cannot understand what is wrong, very strange, would appreciate all the help I can get.

Thank you!

Constantinos

PS Playback chain is Wiim Pro (Tos link @ 24-bit 192Khz) ---> miniDSP Flex --> power amp

 

[OCA]

Hello @Keith_W and everyone,

I have used Acourate quite a few times with Roon, but yesterday I went to edit the target curve in an existing project, run again macros 0-5 and the resulting convolution filter when imported to Roon produced a highly distorted loud sound that was clipping with a -10dB headroom. I have attached the relevant files.

I have imported the wav file to APO and seems fine, also my previous Roon filter is working fine.

Cannot understand what is wrong, very strange, would appreciate all the help I can get.

Thank you!

Constantinos

PS Playback chain is Wiim Pro (Tos link @ 24-bit 192Khz) ---> miniDSP Flex --> power amp

Roon DSP will require about 6dB headroom just for importing impulses "normalized", if you are upsampling, most conversions will also require at least another 3dB headroom. And the remaining 1dB is probably your filter boost.

 

[CK.]

Roon DSP will require about 6dB headroom just for importing impulses "normalized", if you are upsampling, most conversions will also require at least another 3dB headroom. And the remaining 1dB is probably your filter boost.

Thanks for the reply but my filter has -6dB gain and my roon headroom also set to -10, I have used many acourate filters before with 0 issues but this time the resulting sound is like 100dB and distorted, something really wrong is going on. the sampling rate for the filter is 48Khz

 

[OCA]

I have no experience with Acourate filter exports but exporting with IR windows applied and normalized or not from REW effects required Roon headroom as I stated above. But if the sound is very loud and distorted even after -10dB headroom, this points to something else gone wrong with the filter. Your filters are not in .wav format so I couldn't open them (neither can Roon btw).