Viognier, I don't think I am any different to any other audiophile. Everybody, in every showcase system, is trying their best to improve their systems. I suspect that I am the only one on SNA at the moment who is trying to implement a digital active crossover using Acourate/HQPlayer though. Or at least, IF there was someone else ... they would have piped up long ago. 

  Or at least, IF there was someone else ... they would have piped up long ago. 

 

Your unaware of the DarkSNA then K. Very much like the Darknet, they exist, long time members with years under their belt here but with posts counts that are negligible.

I think they operate via PM occasionally and then correspondence is continued off site. True tale  

Yep, Luc is on the money there.  It's very true.

 

 

You are way more patient than me Keith!  I'm sticking with both silver and black spinners for a good while yet rather than true what you are up too!

Acourate is on my horizon just haven't bought one yet.

I suspect that I am the only one on SNA at the moment who is trying to implement a digital active crossover using Acourate/HQPlayer though. Or at least, IF there was someone else ... they would have piped up long ago.

It's a club of (at least) two. At present I'm using Acourate filters for a humble 2-way active desktop system. My filters do the x/o, driver linearisation and room correction - having got around to time alignment. I was using HQ Player with the NAA but at present just using a single box set-up as I kept having issues with the two machines. All that said I wouldn't say that I'm an expert as I've stumbled through the process making every possible mistake.

Edit: My experience is similar in that the bass is a lot more nuanced and overall sound more coherent. That said, I think it's fair to say that the inherent character of the speakers / system such as dynamics are unchanged

 

Edit. Re time alignment. Seem to recall dallasjustice telling me that he could never get HQPlayer to support the time delays built into the filters and thus he stuck with JRIver. (I've not had the time though to do this step.)

Edited August 24, 20169 yr by zydeco

Over the weekend, I decided to make the system 4 way active instead of 3 way. 

 

Then I discovered that the plasma tweeter only accepts high level input. I suppose I always knew this, but was always hoping it would accept a preamp level input. This meant that I had to buy a pair of amps to drive the tweeters. 

I thought about it and remembered my friend Jayden was selling his monoblocks. A quick phone call, and he agreed to let me have them for a great price. 

 

Welcome to the system: 

 

 

... a pair of Valvet A1R monoblocks. 

 

Very interesting design. 25 Watts Single Ended Transistor in pure Class A. I was somewhat skeptical that it made 25W (given how small they are) but boy do they get hot! 

 

Unfortunately, soon after I plugged them in I started blowing fuses in my left tweeter. I hunted for a replacement, and it blew that one, too. So I decided to give up. I'm off to Jaycar today to get more fuses - if it blows one more fuse, the tweeter will have to go to get it checked out. 

 

Dang and I felt bad about ordering a new UMIK-1 last Friday


Sent from my iPhone using Tapatalk

For those who have not seen the inside of a plasma tweeter before, here are a few pics: 

 

 

 

 

There are three chambers. The left chamber contains the amplifier and crossover circuit. The middle chamber has a PL519 valve, that ceramic thing with coils of wire around it is the resonance coil/RF oscillator, and the brass machined chamber is the ignition chamber and horn for the plasma flame. The right chamber contains the power supply. 

 

This thing is incredibly heavy and you need to undo more than 30 Allan screws just to remove the top panel. So it takes a while to get it open. 

 

Despite the incredible build quality of the unit, the valve is held in position with nothing more than a spring. It only takes a little jolt for the valve to unseat itself, and if it does that - no sound. Not that it damages anything in the tweeter, but it is a massive pain in the backside to undo 30 Allan screws just to push the valve in properly. And then having to do them back up again. 

Keith, are you sure the problem is not with the new amp?

 

 

2 hours ago, AudioGeek said:

Keith, are you sure the problem is not with the new amp?

 

 

 

I'm not sure. Thing is, the output of the new amp is much lower than the valve amp I was using. I had to turn the gain up on the tweeters to match the output of the old amp. I have the new fuses, i'll try again tonight. If it blows again, i'll have a chat with my engineer. 

Maybe try the old amps first?

My Corona tweets are much simpler than the Accapella units it seems...

 

They use the same PL519 tube though, and yes, it's held in horizontal just as poorly too!

I found out that the new amplifier oscillates when it is fed a loud test tone. I fed the tweeter a 3000Hz sine wave, forgot to reduce the volume, and heard a high pitched whine from the speaker. I went up to the amp and turned it off. Whine disappeared. Turned it back on, and it did not come back. This is probably what caused the tweeter fuse to blow. 

 

Now I need to decide what to do with it. Give it back to my friend? Buy another amp? Sighs. 

 

Anyway. I continued my Acourate journey. The first thing I did when I was able to was linearize the whole system. Now that I have control over each driver (one DAC channel and one power amp channel per driver), I decided to take nearfield measurements and linearize each driver. 

 

 

This is the result of linearizing the tweeter. It involves measuring the native response of the tweeter, designing a crossover point and slope, selecting what you want to be corrected, and then convolving the native response with the curve that you design. You then take another sweep to check the result. Black = no correction, red = corrected. Not too bad - the response is nearly flat to 10kHz and +/- 2dB from 10kHz - 20kHz. I doubt if it is worth trying to correct those little wiggles at 12kHz and 16kHz, if it is +2dB at that point I doubt if I can hear it. 

 

 

Once every driver was measured and linearized, we have a set of crossover filters. Note that I am using a different amplifier for each channel (each with different input sensitivities and gain), and all the slopes have been normalized so that their peaks are at 0dB. I have one more step to go - load the filters into Acourate, perform another sweep, and correct the volume discrepancies introduced by using different amplifiers. 

 

 

So far, so good. A quick glance at the slopes shows that the midrange and tweeter have been given a substantial boost in volume. This is in line with my expectations. HOWEVER ... there is a rather obvious problem here which I did not notice at first. I loaded these filters into HQPlayer and started listening. 

 

It sounded really weird. The tweeter was clipping, and the frequency response was all over the place! 

 

I went back and had a good look at the filter. Somehow, Acourate has shifted all the filters to the left! The tweeter can't produce frequencies below 1.5kHz - that's why it was clipping. The horn can't make sound below 400Hz, and there is no way that midrange woofer can generate that much volume below 80Hz! 

 

All this took me all evening. I am bloody tired now and i'm going to bed. I'll worry about it tomorrow. 

Edited September 5, 20169 yr by Keith_W

How did you measure the native response of the tweeter Keith? Gated measurement?


Sent from my iPad using Tapatalk

Yes David, it's a gated nearfield response. Acourate calls it "Frequency dependent windowing" - the theory being that the timing of room reflections varies depending on the length of the wavelength. 

 

If I really wanted to (and had hours and hours to spare), I could do what they call a "quasi-anechoic response using beamforming" - the journal article is here. Basically, you take 10 measurement sweeps along a straight line, moving your mic about 50cm each time on a straight line backwards from the speaker. The theory here is that the direct response from the speaker will stay the same, but room reflections will vary. You then use Acourate, align all the measurements, sum them up and hit the magic button that does the calculations - and presto, the room reflections will disappear. 

 

That's what they say, anyway!

 

When I got started on this journey, I thought I would just use Acourate to generate and time align the filters, then put it away and listen to music. But as I delve deeper and deeper into it, I am discovering more and more functions that I don't understand, let alone how to implement. I only recently heard about a double bass array (DBA) from another forum. Turns out that not only does Acourate support creation of a DBA, but it can also do a virtual DBA. Very cool. 

 

This is why I am so enamored with Acourate. With other hardware based solutions - DEQX, MiniDSP, Lyngdorf - you are tied to their hardware. This often means low powered CPU's and so-so DAC's. Upgrades to their system come every few years, and involve purchasing a brand new box and that box is not cheap. DEQX is >$6000 if I recall correctly. 

 

With Acourate, you roll your own PC and you can use an absurdly fast PC if you wanted to (and if you could tolerate the fan noise). How much you spend on DAC's is up to you. And the software is orders of magnitude more powerful and flexible. The author, Uli, is up to date with all the latest research (see beamforming article above - published in 2014, and it was implemented in Acourate in 2015) - and all it takes is a software update and you have new features. 

 

Having climbed the learning curve to use DEQX and now climbing Acourate - I also have to say that THE major downside of Acourate is it's difficulty to learn. DEQX software presents you with a visual representation of your speaker and crossover. With Acourate, you have to go through different menus, which have rather non-obvious names. Also, there is no manual! There is a certain logic to it, but it's clear the interface was designed by an engineer. 

Ah still having to have a computer turned on constantly is a major bummer for me. Otherwise it sounds like something I'd be interested to try. It's great being able to see how you work with it. Very much enjoying your thread.

When my mic arrives I'm thinking of applying Dirac Live just up to midrange and see how it goes. In the meantime I'm trying out various XO settings of miniDSP


Sent from my iPhone using Tapatalk

10 hours ago, Keith_W said:

I found out that the new amplifier oscillates when it is fed a loud test tone. I fed the tweeter a 3000Hz sine wave, forgot to reduce the volume, and heard a high pitched whine from the speaker. I went up to the amp and turned it off. Whine disappeared. Turned it back on, and it did not come back. This is probably what caused the tweeter fuse to blow. 

 

Now I need to decide what to do with it. Give it back to my friend? Buy another amp? Sighs. 

 

Anyway. I continued my Acourate journey. The first thing I did when I was able to was linearize the whole system. Now that I have control over each driver (one DAC channel and one power amp channel per driver), I decided to take nearfield measurements and linearize each driver. 

 

 

This is the result of linearizing the tweeter. It involves measuring the native response of the tweeter, designing a crossover point and slope, selecting what you want to be corrected, and then convolving the native response with the curve that you design. You then take another sweep to check the result. Black = no correction, red = corrected. Not too bad - the response is nearly flat to 10kHz and +/- 2dB from 10kHz - 20kHz. I doubt if it is worth trying to correct those little wiggles at 12kHz and 16kHz, if it is +2dB at that point I doubt if I can hear it. 

 

 

Once every driver was measured and linearized, we have a set of crossover filters. Note that I am using a different amplifier for each channel (each with different input sensitivities and gain), and all the slopes have been normalized so that their peaks are at 0dB. I have one more step to go - load the filters into Acourate, perform another sweep, and correct the volume discrepancies introduced by using different amplifiers. 

 

 

So far, so good. A quick glance at the slopes shows that the midrange and tweeter have been given a substantial boost in volume. This is in line with my expectations. HOWEVER ... there is a rather obvious problem here which I did not notice at first. I loaded these filters into HQPlayer and started listening. 

 

It sounded really weird. The tweeter was clipping, and the frequency response was all over the place! 

 

I went back and had a good look at the filter. Somehow, Acourate has shifted all the filters to the left! The tweeter can't produce frequencies below 1.5kHz - that's why it was clipping. The horn can't make sound below 400Hz, and there is no way that midrange woofer can generate that much volume below 80Hz! 

 

All this took me all evening. I am bloody tired now and i'm going to bed. I'll worry about it tomorrow. 

 

That's strange. Are you sure that you didn't use the cross-over filters for the 3-way set-up in creating the 4-way set-up? 

29 minutes ago, zydeco said:

 

That's strange. Are you sure that you didn't use the cross-over filters for the 3-way set-up in creating the 4-way set-up? 

 

GOOD POINT! Arghhhhhhh, it does look like the crossover filter for the 3 way setup! I must have been using the wrong working directory! I'll find out when I get home. 

11 hours ago, Keith_W said:

 

 

 

This is the result of linearizing the tweeter. It involves measuring the native response of the tweeter, designing a crossover point and slope, selecting what you want to be corrected, and then convolving the native response with the curve that you design. You then take another sweep to check the result. Black = no correction, red = corrected. Not too bad - the response is nearly flat to 10kHz and +/- 2dB from 10kHz - 20kHz. I doubt if it is worth trying to correct those little wiggles at 12kHz and 16kHz, if it is +2dB at that point I doubt if I can hear it. 

 

 

 

 

Been following this with interest Keith as about to setup my own DIY speaker with PC and acourate.

 

Understand your only starting out with measuring and crossover points but with the above response if black is the tweeter's natural roll off and red is corrected is that not forcing (boosting) the tweeter to operate out of its normal range and create other issues like distortion? or is this ok to do with acourate? i would have thought from your graph crossover point should be around 5kH but looks like your boosting the tweeter to target 2kHz.

 

 

 

Edited September 6, 20169 yr by lusk

36 minutes ago, lusk said:

or is this ok to do with acourate?

 

No, accourate isn't 'special' in that regard.     You're right - boosting a horn below the cutoff like that, isn't typically a good idea.

Hmm, you guys are probably right. Acourate is also able to measure distortion - I will go take a look and see if there is much of it. If there is, I will choose a higher crossover point. 

2 hours ago, Keith_W said:

Hmm, you guys are probably right. Acourate is also able to measure distortion - I will go take a look and see if there is much of it. If there is, I will choose a higher crossover point. 

 

It's not only a (non-linear, ie. harmonic/intermod) distortion issue, in fact it might be completely inaudible at home SPLs.

 

The beam width (coverage pattern, aka polar response) will change (widen) as you go lower in frequency....   this means that your crossover will work at one measurement angle .... but from a different angle, there will be a lump.

 

 

@davidro waiting for the computer to boot is no problem. In fact the computer boots faster than it takes the valve power amp to warm up. Or my old preamp to even turn on. 

 

@davewantsmoore that is another good point. Tomorrow I will go and buy an Earthworks M30 microphone and redo all the measurements and probably choose a higher crossover point. I was thinking 4kHz, since that is when the response starts to really drop off. 

 

On another note - I got home early and the first thing I did was to go and check if the crossover filter left-shift was caused by me loading the 3-way crossover by mistake. NOPE. I correctly loaded the 4 way crossovers. I regenerated the filters, paying close attention to what it was using to make the corrected files. No dice - it still shifts it to the left. I have asked a question on the Acourate forum. 

 

On yet another note - I have been told that the NADAC will be in my hands later this week! Darn, I was hoping to have all my filters ready in time for it. Guess my shiny new toy will have to sit on the shelf until I make those filters! 

8 hours ago, davidro said:

Ah still having to have a computer turned on constantly is a major bummer for me. Otherwise it sounds like something I'd be interested to try.

 

My HTPC goes to sleep automatically after 10 minutes of inactivity and with a fast SSD wakes up instantly after pressing a button called "ON" from my IR universal remote control. You can turn a PC into a Hi-Fi appliance, it can be done. 

 

My HTPC goes to sleep automatically after 10 minutes of inactivity and with a fast SSD wakes up instantly after pressing a button called "ON" from my IR universal remote control. You can turn a PC into a Hi-Fi appliance, it can be done. 

Yeah but it takes money, effort, time, knowledge etc. just look at what the op is going through.


Sent from my iPad using Tapatalk