Michael, Thank you so much for posting this. After doing new measurements I see the miniDSP 2x4HD works perfectly on the 50Hz 4th order low pass filter. I went back and made a measurement today and both reproduced my earlier result as well as confirming yours ??? It seems when I measure the impulse response with Arta and then generate the Frequency Response and Distortion analysis plot, I get what I remembered seeing. An initial sharp roll off followed by a shallow roll off. Plot 1 below. I then used a one of the other Analysis options. The default 16k FFT on the Single Gated FR had some issues. When I increased to the 256k FFT and ran the Single Gated FR analysis, that plot matched your result. Plot 2 below. So I was getting fooled by the Frequency Response and Distortion analysis plot. I have yet to figure out why it shows such a different result. I suspect it has a very narrow FFT window or something. I don't see any settings that change the result. So many ways to go wrong. Thanks again. Dana Olson
Plot 2: Arta Gated Frequency Response smoothed, using the same impulse response data as used in Plot 1 above. FFT 256k
Engineers working (ie. as a profession) in audio embraced DSP at the end of the 20th century.
And could accomplish more than ever.
They never looked back.
And the general consumer and the professional marketplace embraced these products with open arms.
Exhibit A:
https://store.google.com/product/nest_audio
Who is the Original Equipment Manufacturer for this touch sensitive, voice controlled, Bluetooth audio receivable, app configurable, DSP tuned, bi-amped, waveguided two-way speaker built in an recycled aluminium enclosure, which retails for under USD$99 per speaker?
Peerless/Tymphany
Measurements:
Measurements reference:
https://www.audiosciencereview.com/...-nest-audio-spinorama-and-measurements.16464/
Who needs DIYers who are still pontificating about old (solved) issues when you can have a productive conversation with clients who know how to advance the state of the art? Tymphany doesn't.
/rant off
And could accomplish more than ever.
They never looked back.
And the general consumer and the professional marketplace embraced these products with open arms.
Exhibit A:
https://store.google.com/product/nest_audio
Who is the Original Equipment Manufacturer for this touch sensitive, voice controlled, Bluetooth audio receivable, app configurable, DSP tuned, bi-amped, waveguided two-way speaker built in an recycled aluminium enclosure, which retails for under USD$99 per speaker?
Peerless/Tymphany
Measurements:
Measurements reference:
https://www.audiosciencereview.com/...-nest-audio-spinorama-and-measurements.16464/
Who needs DIYers who are still pontificating about old (solved) issues when you can have a productive conversation with clients who know how to advance the state of the art? Tymphany doesn't.
/rant off
Last edited:
Answers:
1. frequency response, sound stage, resolution, neutrality, etc etc.
2. frequency response, sound stage, resolution, neutrality, etc etc.
3. Magic; although it sometimes comes close ;-)
Jan
1. frequency response, sound stage, resolution, neutrality, etc etc.
2. frequency response, sound stage, resolution, neutrality, etc etc.
3. Magic; although it sometimes comes close ;-)
Jan
1. Improve in-room bass response. Improve frequency and phase response.
2. Improve frequency and phase response. Allow the use of smaller cabinets yet maintaining the same low end response (the only way to bend Hoffman's iron law). Protect drivers against destroying themselves. Allow the use of multiple smaller (lower cost, lower power consumption) amplifiers.
3. Reduce non-linear distortion, affect directivity...
PS. @AllenB
Evidence shows that you really can't change people's minds on LOTS of things.
The more you present a counter-argument, the more they dig their heels in. From audio to politics to vaccinations...
2. Improve frequency and phase response. Allow the use of smaller cabinets yet maintaining the same low end response (the only way to bend Hoffman's iron law). Protect drivers against destroying themselves. Allow the use of multiple smaller (lower cost, lower power consumption) amplifiers.
3. Reduce non-linear distortion, affect directivity...
PS. @AllenB
Evidence shows that you really can't change people's minds on LOTS of things.
The more you present a counter-argument, the more they dig their heels in. From audio to politics to vaccinations...
Last edited:
That's my experience.
I'm not interested in turning people whatever way; I'm not a political party.
But having info helps to make a decision.
@tktran303 - you can't change opinions by logical arguments if those opinions are not based on logical arguments to begin with ...
Jan
Hi all, Hi Bryguy
I have entered DSP-audio in 2017, when I had the chance/money do build some big loudspeakers. A three-way speaker with Duelund filters.
I had two main objectives back then:
1. I wanted to go all digital with a PC based system, no more cable- and connector-bla bla bla.
2. I did not want to buy endless numbers of components for passive filters and spend time on experimenting in all kind of directions.
So I found some MiniDSP plate-amps that could be connected (daisy-chained) digitally with AES/EBU to a MiniDSP OPEN DRC as "digital preamp" and a MiniDSP U-DIO8 to connect things to my old NUC i3.
It works fine for me, and I can do regular IIR filtering in the plate-amps or I can bypass the OPEN DRC and go four-way FIR through the U-DIO8 and only use the dac and amp in the plate-amps. (Today, much wiser, I had probably gone an other digital route).
I am now in the process of building a pair of really big MEH´s. It´s going to be a three-way MEH with SUB, so four-ways, and I do feel tempted to go FIR in the long run. We will see. I visited Uli Brueggemann who created Acourate this summer, that was nice!
The more that I am back in DIY and reading threads and learning things, I DO NOT understand why people would want to buy a loudspeaker with a fixed frequency-response!? There is no "standard"-room, so all speakers sound different and more or less correct in different rooms! People listen at different SPL´s. Place the speakers here and there. So many things that I think need to be adjustable/flexible. DSP, that I can program/get access to, gives me freedom to adjust to my taste and my room. One can even use the DSP as on the fly Equalizer.
Only downside is, that I do need to learn and know what and what not to do with a DSP, but that is just part of the hobby.
I am not into DIY to save money, but to make a dedicated Steffen-solution, something that I can not buy! Well I could buy a pair of SH96, but they do not fit into a corner, bummer!
That´s just my two cents.
Steffen
I have entered DSP-audio in 2017, when I had the chance/money do build some big loudspeakers. A three-way speaker with Duelund filters.
I had two main objectives back then:
1. I wanted to go all digital with a PC based system, no more cable- and connector-bla bla bla.
2. I did not want to buy endless numbers of components for passive filters and spend time on experimenting in all kind of directions.
So I found some MiniDSP plate-amps that could be connected (daisy-chained) digitally with AES/EBU to a MiniDSP OPEN DRC as "digital preamp" and a MiniDSP U-DIO8 to connect things to my old NUC i3.
It works fine for me, and I can do regular IIR filtering in the plate-amps or I can bypass the OPEN DRC and go four-way FIR through the U-DIO8 and only use the dac and amp in the plate-amps. (Today, much wiser, I had probably gone an other digital route).
I am now in the process of building a pair of really big MEH´s. It´s going to be a three-way MEH with SUB, so four-ways, and I do feel tempted to go FIR in the long run. We will see. I visited Uli Brueggemann who created Acourate this summer, that was nice!
The more that I am back in DIY and reading threads and learning things, I DO NOT understand why people would want to buy a loudspeaker with a fixed frequency-response!? There is no "standard"-room, so all speakers sound different and more or less correct in different rooms! People listen at different SPL´s. Place the speakers here and there. So many things that I think need to be adjustable/flexible. DSP, that I can program/get access to, gives me freedom to adjust to my taste and my room. One can even use the DSP as on the fly Equalizer.
Only downside is, that I do need to learn and know what and what not to do with a DSP, but that is just part of the hobby.
I am not into DIY to save money, but to make a dedicated Steffen-solution, something that I can not buy! Well I could buy a pair of SH96, but they do not fit into a corner, bummer!
That´s just my two cents.
Steffen
Last edited:
Agreed. I think that the only way to correct room response is with DSP. It could be possible with line level analogue, but certainly not practical at speaker level. I also have frequency dependent compression (DEQ) so any notches in the room response I have filled in won’t overdrive my modest speakers
Brian
Brian
Those trying to understand the problem for their own benefit may get confused if it's implied there's a fundamental difference.
You should be getting the same result either way, or there must be something wrong with the way it was carried out.
There is a fundamental difference.
The difference is although theoretically you can get a good response with using passive high-level circuits and analogue line levels circuits....
Practically, you cannot reach the level of fine tuning that can be attained with DSP- one cannot correct linear distortion to the finest level of granularity that DSP can achieve. Good luck if even if you are using dozens of parts in your passive crossover or dozens of IC in your line level active. I can go smoother flatter etc with DSP... with non DSP at some point I just stop and say, well, I think this is already excellent, I don't think it's worth adding all those extra parts.
eg. "Should I notch out my 27KHz tweeter resonance?"
eg. Should I notch out this low Q peak... but it's already at 30dB down and hardly changes the summed response?
eg. Should I do time alignment of drivers? Should I add a ladder delay network?"
The ASP answer is..-
"Maybe, maybe not. It depends on the increased complexity and potential downsides and cost of extra parts. I need to build two crossovers to compare..."
DSP- "why not? Just dial it in. Click it in/out and listen for yourself in real time"
As for phase correction, ie. FIR, you cannot do that @CharlieLaub alluded to.
The difference is although theoretically you can get a good response with using passive high-level circuits and analogue line levels circuits....
Practically, you cannot reach the level of fine tuning that can be attained with DSP- one cannot correct linear distortion to the finest level of granularity that DSP can achieve. Good luck if even if you are using dozens of parts in your passive crossover or dozens of IC in your line level active. I can go smoother flatter etc with DSP... with non DSP at some point I just stop and say, well, I think this is already excellent, I don't think it's worth adding all those extra parts.
eg. "Should I notch out my 27KHz tweeter resonance?"
eg. Should I notch out this low Q peak... but it's already at 30dB down and hardly changes the summed response?
eg. Should I do time alignment of drivers? Should I add a ladder delay network?"
The ASP answer is..-
"Maybe, maybe not. It depends on the increased complexity and potential downsides and cost of extra parts. I need to build two crossovers to compare..."
DSP- "why not? Just dial it in. Click it in/out and listen for yourself in real time"
As for phase correction, ie. FIR, you cannot do that @CharlieLaub alluded to.
BTW just regarding the forum, if you edit your post to page a member it doesn't register. It has to be posted at the original press of 'Post Reply' to work.
DSPs do have some irrefutable advantages like:
A DSP that is powerful enough to reduce its FIR latency to (at least) match the filter delay (which is fixed) looks really useful. For example, a DSP with a computational delay of 10ms for another 10ms of 'linear phase' filter / crossover delay. However, if the filter function is an FIR replication of an IIR filter, the IIR filter maybe used directly to obtain negligible latency and a 10ms overall delay.
- Programmability - Load/save/transfer settings, recalibration etc.
- Repeatability with high channel-to-channel matching.
A DSP that is powerful enough to reduce its FIR latency to (at least) match the filter delay (which is fixed) looks really useful. For example, a DSP with a computational delay of 10ms for another 10ms of 'linear phase' filter / crossover delay. However, if the filter function is an FIR replication of an IIR filter, the IIR filter maybe used directly to obtain negligible latency and a 10ms overall delay.
Are you telling me you're unable to achieve an arbitrary response using conventional circuit methods?
Also for Thanh..
Also for Thanh..
Actually too much of any form of correction towards flat can make a mess of things. As acoustic issues become better managed, the response anomalies change with them. The right way to do this is to fix the acoustic problem.
Sorry Alan, that's a content-free statement. Just a bunch of words but it doesn't say why or how.
What mess, how created?
And yes, I can make an arbitrary response with 'conventional' means. (I assume you mean 'means that were cutting edge last century'). Do I want to invest a few years of my live and half of my retirement savings? No, I'd rather do it on a rainy afternoon with a $200 sound card and free software.
And what you call 'the right way', that's just - again - an unsubstantiated personal opinion.
Jan
I am in complete agreement about fixing very bad acoustic problems.
Allow me to expand-
Suppose your speaker has some major problems.
Exhibit A:
Reference: https://www.erinsaudiocorner.com/loudspeakers/behringer_b212xl/
According to some thinkers out there, because this speaker has a generally smooth and gradually rising directivity index, it can be equalised.
Here’s a suggested DSP based correction:
Reference:
https://www.spinorama.org/speakers/Behringer Eurolive B212XL/ErinsAudioCorner/index_eac-pattern-90x60.html
Well yeah that looks somewhat smoother.
But is it gonna sound any better?
It probably DOES have the “life sucked out of it” - you know all that funkiness going on between 1Khz to 10Khz. This is probably what gives it an err… “special or unique character”, if I’m going to be kind about it… And then give it a subdued midrange and top end.
I’d fix whatever underlying issue is there. That could be a cone and/or cabinet resonance.
Yes fix the underlying acoustic issues. Get a better behaved driver. Invest in a more structural inert cabinet.
Here’s another wonder of a speaker:
Reference:
https://www.audiosciencereview.com/...ds/avantone-pro-mixcube-monitor-review.42817/
For this speaker, Have a guess where
“the life of the party” is.
Holy smokes no baffle step compensation and resonances you can see for miles!
Why even bother fixing that with any EQ, (DSP or not). So Just because you can do MORE with DSP, and it’s easier, doesn’t mean you should.
This is the Horrortone- the saying goes- if you can make your Music sound good on this speaker, it can sound good on any ol speaker..
No, I wouldn’t bother with DSP here.
Same with rooms.
Put some top quality speakers in an Aussie backyard dunny? Fuggedaboutit…
But to be clear, I’m happy to agree to disagree with you on this Allen, if we are doing to go further on this issue…
Last edited:
Even while using DSPs, it is generally not recommended to 'suck the life' out of anything. Instead it is recommended to first solve mechanical & acoustic problems like distances / reflections / resonances to get a smooth (vs. flat) response and then use gentle (low-Q) EQ to obtain flatness.
While it is possible to come back and EQ another layer of kinks in the response (and so on) to iteratively get closer and closer to perfection, any small ones are best left alone.
And, of course, I understand how this could be a thing of personal preference to many people.
While it is possible to come back and EQ another layer of kinks in the response (and so on) to iteratively get closer and closer to perfection, any small ones are best left alone.
And, of course, I understand how this could be a thing of personal preference to many people.