I hope it will get you some idea how to go about it.
To prove it can work with measured impulses I've used the theoretical sub and an actual measurement of the line array in an Overlay view of their filtered IR, where I've adjusted the timing of the array to fit the subwoofer:
See how it's still easy to align the shapes (identify them), even with an actual measured result at the listening spot? The slight wavy shape in front of the array measurement actually is pre-ringing of my phase correction, still the wave shape comes through undisturbed and is easily recognised and used for alignment without getting the timing wrong.
I know of no other way that could give you this kind of visual confirmation. It basically takes the guess work out of the equation.
To prove it can work with measured impulses I've used the theoretical sub and an actual measurement of the line array in an Overlay view of their filtered IR, where I've adjusted the timing of the array to fit the subwoofer:
See how it's still easy to align the shapes (identify them), even with an actual measured result at the listening spot? The slight wavy shape in front of the array measurement actually is pre-ringing of my phase correction, still the wave shape comes through undisturbed and is easily recognised and used for alignment without getting the timing wrong.
I know of no other way that could give you this kind of visual confirmation. It basically takes the guess work out of the equation.
Attachments
Last edited:
I have not tried the DACT, but several other high quality stepped volumen controls and even though they are sounding good they never filled the bill for me , as the discrete steps were too coarse. Brunos preamp and the lightspeed attenuator are sounding a bit better, but no by a wide margin.
Blatant knowledge theft
To time-align sub with main (main fixed, sub delay changeable)
Align as A or B - never got that really... on the peak or on the start?
Super question imo
Get's at the root of understanding phase and timing....a question Wesayso helped me solve/understand several years back.
The simple answer.
Correct time alignment is when starts are aligned.
Not peaks.
The well noted problem is that our measurement programs key off impulse peaks.
And peaks get less and less defined, as high frequency response is taken away. Which makes defining impulse starts even more difficult than tagging peaks.
Ime, impulse responses are the wrong tool to use for fine time alignment.
They are useful as a quick start to get into the correct time alignment ballpark, by helping to remove excess delay.
But once that excess delay is removed, phase traces are an easier/better way to achieve impulse start alignments, than trying to continue to analyze impulse (or step, or group delay, etc).
When phase traces lay on top of each other as closely as possible through the xover region, impulse start alignment has been accomplished as well as can.
Aligning phase traces is a task that dual channel FFT's handle much better than sine sweeps, again ime...
They make better phase traces down low, since dual FFT allows continuous real time transfer functions. Temporal averaging of the continuous measurement can be used for very good stability and repeat-ability.
Even better, real time phase traces allow you align phase by having one section's (say sub) captured trace up on screen, while measuring the other's (main) trace in real time, and simply dialing processor delay on the leading section until the phase traces overlay.
After using the match phase trace method, I've used REW to confirm and peer into bandwidth filtered impulses like in Wesayo's good explanation.
All jives together fine. Both methods work.
But wow, is dual FFT easier. just another fwiw
Interesting and affordable suggestion... here I was only looking at volume controls like this:
Simple, basic but expensive...
Wow ! and yikes ! I couldn't imagine doing something like that.
I might have been misleading when I was talking about gain staging and the need for analog attenuation after the digital stage (to match amp sensitivity.)
When I look at block diagrams of DSP processors, there is usually a DAC input block, then a DSP processing ouput block (EQ's xovers, dynamics, etc), then a final DAC output block for digital attenuation that essentially becomes analog attenuation.
Instructions are invariably to keep the DAC input block just below 0dBFS, and then the same for adjusting output of the DSP processing block.
The last block becomes the the "volume control", and is still digital.
Whether there are actual separate blocks I dunno...I leave that to folks like Fluid, Kolby, BYRTT, et al...
I do know my processor allows meters to be placed at those points (real or virtual) for optimizing gain staging.
Really good tutorial Ronald. You may recall explaining it to me a few years ago and I couldn't get it to work for me. In my defense then it turns out I had a ceiling reflection coming in at the LP and causing phase difficulties at XO. But this explanation is perfectly clear and even I can follow it. But I ended up aligning on impulse start. It would be interesting to compare results with the two methods.
Changing the subject - thanks all for interesting discussion on volume control
Audio Bitdepth - JRiverWiki
You can't run out of digital headroom but if you use any of it for volume control, that will bring the DAC's output signal closer to the DAC's noise floor and thus degrade the SNR. Does it matter enough to merit design and construction of a multi-channel preamp? Better to design a DAC board with volume control integrated into the DAC's current to voltage conversion and buffering stages.
Changing the subject - thanks all for interesting discussion on volume control
Audio Bitdepth - JRiverWiki
You can't run out of digital headroom but if you use any of it for volume control, that will bring the DAC's output signal closer to the DAC's noise floor and thus degrade the SNR. Does it matter enough to merit design and construction of a multi-channel preamp? Better to design a DAC board with volume control integrated into the DAC's current to voltage conversion and buffering stages.
Should have known/remembered this stuff is up your alley!
So it looks like there is an actual distinct volume control block post all DSP processing.
You know, here's an off the beaten track recommendation for a home audio multi-channel DAC, I/O box, and bit of a DSP processor.
The Behringer XR-18 stagebox mixer.
18x18 USB. 18x8 balanced I/O.
Decent EQs and Xovers.
Full fledged mixer/routing capability.
Very good ethernet control via PC or Mac, Behringer apps..
Excellent wireless tablet control with built in wifi, and $5 Mixing Station Pro app.
All for under $600..
Good sound card for REW too
I don't think my plan here with a volume control is that wild. Do remember that up to 15-17 dB of my digital headroom is locked up in DSP (used to be about 25 dB even) with a 20 bit available in the DAC.
Get another DAC would be a simple and truly reasonable suggestion. But... not all amps sound alike and certainly not all DAC's sound alike. I do know what I've got here (and value it).
So the analog attenuation has become something on my mind due to the most recent experiences. Once you use up too many bits inside that digital room you don't get the output you should have. No way around that with the current means. I don't have a system here that needs only a little wiggle room, I use up a lot of internal space. And I like it that way, but I can win quality at lower levels by taking the analog attenuation outside of the digital realm.
Get another DAC would be a simple and truly reasonable suggestion. But... not all amps sound alike and certainly not all DAC's sound alike. I do know what I've got here (and value it).
So the analog attenuation has become something on my mind due to the most recent experiences. Once you use up too many bits inside that digital room you don't get the output you should have. No way around that with the current means. I don't have a system here that needs only a little wiggle room, I use up a lot of internal space. And I like it that way, but I can win quality at lower levels by taking the analog attenuation outside of the digital realm.
Let's be honest here, we all do what we do, and even though we might have similar systems, each of us has made different choices for a similar problem.
None of us use the same equipment. Why is that?
My way forward is and will continue to be to identify a weak spot and fix it. One thing at a time. Introducing an all new DAC is changing way more than one variable. In other words: I cannot control that. Changing one variable at a time and confirm what it does (roll back if needed) has been my method of choice.
Changing more than one variable at a time means we do not learn what was fixed. I like to learn.
None of us use the same equipment. Why is that?
My way forward is and will continue to be to identify a weak spot and fix it. One thing at a time. Introducing an all new DAC is changing way more than one variable. In other words: I cannot control that. Changing one variable at a time and confirm what it does (roll back if needed) has been my method of choice.
Changing more than one variable at a time means we do not learn what was fixed. I like to learn
.
Last edited:
Super question imo
Get's at the root of understanding phase and timing....a question Wesayso helped me solve/understand several years back.
The simple answer.
Correct time alignment is when starts are aligned.
Not peaks...
Minimum phase slopes there
for linear phase slopes its peaks not starts, right
Throw away enough bits for processing and see what happens with your signal while trying to use it as volume control as well: not good.
Up and down 5 dB won't be a problem. But attenuate it another 15 to 20 dB and degradation starts.
It all depends on what you do with what's available. And your own personal reasons why, I guess
. Fixing most of the digital level SPL and doing the big jums analog is the plan here.A little up and down volume work can still happen digitally, with my phone as the remote
. The dial presented here wouldn't even be in the same room . And it would get the Universal Buffers behind it to power the amps...
Last edited:
Which is why I said that this method displayed here works for both, without worrying about it
Yes superior nice guide there wesayso thanks.
Throw away enough bits for processing and see what happens with your signal while trying to use it as volume control as well: not good.
Up and down 5 dB won't be a problem. But attenuate it another 15 to 20 dB and degradation starts.
It all depends on what you do with what's available. And your own personal reasons why, I guess
A little up and down volume work can still happen digitally, with my phone as the remote
All depends agree, but as youself also told before there is still possibility gain some digital bits back for Jriver volume if one make a simple opamp filter that do the rough character low end boost shelve for arrays
Yes, but I'm not an electronics wizard, you know .
We each use what we've got and what we can get or build...
There will always be limits....
What I suggest here, if the DACT makes true on it claims, keeps every thing in check and might provide me better enjoyment at low listening levels.
Something I've never used before and did not care for even. But if i can make it happen while not upsetting what I've got....
Remember, with unlimited space, money and means I'd be sporting a very efficient horn setup with arrayed bass. Way too big to be practical in my real world.
So I use what can work for me, and work hard to get the best out of it. I don't think it is a large compromise, I do enjoy the level I've got at immensely.
And always working on gaining more out of it. By studying it and thus: changing one thing at a time. Controlled progress, but progress non the less.
.We each use what we've got and what we can get or build...
There will always be limits....
What I suggest here, if the DACT makes true on it claims, keeps every thing in check and might provide me better enjoyment at low listening levels.
Something I've never used before and did not care for even. But if i can make it happen while not upsetting what I've got....
Remember, with unlimited space, money and means I'd be sporting a very efficient horn setup with arrayed bass. Way too big to be practical in my real world.
So I use what can work for me, and work hard to get the best out of it. I don't think it is a large compromise, I do enjoy the level I've got at immensely.
And always working on gaining more out of it. By studying it and thus: changing one thing at a time. Controlled progress, but progress non the less.
Last edited:
Thumb up had always myself drooled looking those nice hardware DACT units, but implementation is not in sync saing i'm not an electronics wizard, but guess you are humble there and really think you a master with your hands and able to familiarize yourself with the complicated technical solutions of the world
Minimum phase slopes there
Yes, good catch.
(and another reason to simply stick with easy to read phase traces imo...
I think very nature of the question, impulse start vs peak, pretty much implies minimum phase though.
And it's kinda seldom anyone tries to employ linear phase for subs to mains...
Hey, you guys ever use ARTA or some other dual channel ?
Anybody else find dual channel and phase traces so much easier, or is it just me?
Maybe we should all just fully go lin phase so everything gets plain easy
Throw away enough bits for processing and see what happens with your signal while trying to use it as volume control as well: not good.
Up and down 5 dB won't be a problem. But attenuate it another 15 to 20 dB and degradation starts.
It all depends on what you do with what's available. And your own personal reasons why, I guess
A little up and down volume work can still happen digitally, with my phone as the remote
Aren't many digital processors using 40bit or better signal processing? For instance, the little Behringer mixer I mentioned does.
I don't think being down 15 to 20 dB is a problem then, even after allowing 17dB headroom. I'm down that much all the time given the high efficiency drivers being used. (with amps at +26dB).
I guess i was going yikes about the multi-pot, because I've had more trouble with pots going crusty over time, than any other problem with high-end home gear. Conrad-Johnson, even Manley. (I don't live in a corrosive environment)
But sorry if i yiked out of place
It really is all about what we want and enjoy, huh?
Internally JRiver is doing everything in 64 bit space. No problems there.
It is the wide stretch that is used that has to go from digital to analog that is eating away from the available digital space that counts. And that space happens to be 20 bits available for me with my DAC/sound card.
So internally? No problem at all.
But listening to the Goldmund amp, before I had the Atom pré amp in place and listening at lower levels than used for critical listening I was amazed at what I heard, compared to what the Pioneer did at such levels.
As said, if I can keep the volume setting at that level all the time, playing it loud or soft (with analog attenuation for big stels, digital for finer steps) will always give me that level of detail. Not throwing away bits that are responsible for carrying that signal helps.
If it were just like most setups, just a driver with reasonably flat output handing of to another driver with reasonably flat output you'd never run into that problem, period. That's just not what I have.
It is the wide stretch that is used that has to go from digital to analog that is eating away from the available digital space that counts. And that space happens to be 20 bits available for me with my DAC/sound card.
So internally? No problem at all.
But listening to the Goldmund amp, before I had the Atom pré amp in place and listening at lower levels than used for critical listening I was amazed at what I heard, compared to what the Pioneer did at such levels.
As said, if I can keep the volume setting at that level all the time, playing it loud or soft (with analog attenuation for big stels, digital for finer steps) will always give me that level of detail. Not throwing away bits that are responsible for carrying that signal helps.
If it were just like most setups, just a driver with reasonably flat output handing of to another driver with reasonably flat output you'd never run into that problem, period. That's just not what I have.