Agreed, If any of the bark's reproduced frequency components are out of a match in time with the natural bark, the bark will not sound like the same bark anymore.
So how do we keep that from happening?
Whether there is much phase shift depends on a system's number of ways, or rather number of xovers, and their types and orders.
But the bark itself will behave similar to a band limited IIR system. There's no one there to adjust it's phase behavior 😀. In other words: natural sounds act according to the IIR rules. They don't stick to a specific bandwidth, say 20 Hz to 20 KHz, they are what they are, in all their IIR glory.
Member
Joined 2009
Paid Member
I was referring to Mark's comment on how the much phase shift it would take for the bark to not sound like a bark anymore. It would take a lot, and normal crossovers simply don't have that much phase shift to warp the sound.
That is not to say that crossovers don't have phase shift or that improper phase overlay between drivers through the crossover region isn't audible--it very much is audible. But it makes itself audible through the ripples in the amplitude response (on- and off-axis), and not through the phase turning through the crossover.
This leads to the question of can you hear FIR correction of crossover phase change. My experience is no, but why not fix it when you can using FIR filters?
Wesayso may have thoughts on this... he's done a ton of experimentation on this aspect.
I just don't think you can design a crossover in a way to make the bark not sound like a bark. Drivers and crossovers don't behave that way.
BTW, how did we get here?
There is nothing like being able to change any crossover parameter wirelessly from the listening seat and hear exactly what is happening in real time without ever moving your head. As far as I’m concerned, there is no comparison. (Note the monitor)
Attachments
Last edited:
ra7 said:
Not going to go there, but I'll admit: yes I do use FIR filters to manipulate my results.
There are many things that would go higher up the list of things to worry about. But as you said: if you can, why not? 🙂
It (a system with linear phase crossovers) would be able to reproduce a more 'right' version of that 'bark', technically. Whether you're able to hear it or not. But even that depends on so many more variables...
Nope, a dog will bark whenever it feels like it. If one records that sound, you'll be able to see its frequency response. That will show as a band limited sound, complying to IIR rules.
One could argue if you play that sound back on a system with one (or more) crossover(s) or somehow with a too limited bandwidth to completely cover the spectrum of the original, then it isn't a good representative of the original sound.
But to capture a dog's bark, without capturing it's natural surroundings would be a daunting task all by itself. Playing it back in a different environment... there we go...
Last edited:
Hi Wesayso,
This is what i don't think is true....that the bark itself will behave similar to a band limited IIR system.
I think the bark doesn't have a behavior that can be described by mag and phase response, as the bark is not a response at all...to any kind of stimulus signal.
The measurement of the bark however, will make the bark look like it came from a band limited IIR system. It's the only way the FT-IFT measurement system can make sense of a bandlimited response.
The measurement automatically assume the bark is a response, when it really isn't.
The bark is its own unique stimulus, that can't really be compared to anything else i think.
It's just a bark....maybe up a tree like me, hahaha
Ps if we do think a bark responds like a band limited speaker, shouldn't we be saying the bark behaves like minimum-phase, rather than IIR ?
Please never mind this...most all IIR are min phase....wasn't worth mentioning.....
Last edited:
What is making the bark noise, think about that. Wouldn't that comply to all the rules we encounter in reproduction of sound? Think about that swing again, that AllenB mentioned.
Yes, minimum phase would be a more proper term to use.
Yes, minimum phase would be a more proper term to use.
The bark sounds like it does due to the harmonics it produces. What you can tell as "real" vs "recorded" is due to the harmonic structure of the original sound (bark, piano, whatever). Adding in all other manner of content further works against the realistic reproduction.
That's exactly what i've spent ages pondering....what is making the bark?
What makes any sound that occurs in nature, and does that differ from reproduction?
Have also spent considerable time studying our measurement programs, and the assumptions they rest on. The programs all work (afaict) to find minimum phase via the Hilbert Transform.
The programs are all meant to measure system response, the reproduction response of how a system responds to a stimulus.
I believe it is a miss-application trying to make natural sounds fit the Hilbert-Transform, fit minimum phase. I believe natural sounds are what they are, with no need to fit them into response math.
What makes any sound that occurs in nature, and does that differ from reproduction?
Have also spent considerable time studying our measurement programs, and the assumptions they rest on. The programs all work (afaict) to find minimum phase via the Hilbert Transform.
The programs are all meant to measure system response, the reproduction response of how a system responds to a stimulus.
I believe it is a miss-application trying to make natural sounds fit the Hilbert-Transform, fit minimum phase. I believe natural sounds are what they are, with no need to fit them into response math.
Not really what I mean in this regard. The bark itself will contain certain harmonics that make it up. But it will contain a group of frequencies, it won't ring forever up to the thousands of KHz. Yet, it's origin of sound was a vibration of the dog's vocal cords, shaped by it's mouth movement. Nobody limited it, but it's limited in frequency content.
The mouth of the dog could be seen as a filter, if you think about it. creating a band pass...
Neh.... no I'm taking it way too far!
I keep thinking about Roger Waters, strangely enough... 😀
The mouth of the dog could be seen as a filter, if you think about it. creating a band pass...
Neh.... no I'm taking it way too far!
I keep thinking about Roger Waters, strangely enough... 😀
The Achilles heal of DSP will always be time domain error, not just jitter itself in both conversion steps, but errors introduced by the DSP in the software side. Then theres the input lowpass filtering process, which may cause relative phase issues in the top end.
You guys who only play back digitial source have the advantage of already being done with the first conversion step. The studio that mastered the recording will likely have access to the best gear available to minimize this issue, but those errors (hopefully not to many) are basically already there and you're stuck with them. You only have to deal with the last conversion step back to analog, which isn't as hard to get right as A/D conversion. That makes DSP so practical since you're already dealing with digital source, but there will however always be time domain errors (jitter) built into it along with processing artifacts.
Us analog source guys (specifically vinyl) don't want to risk screwing up our carefully derived audio with two digital conversion processes, plus extra DSP. Its sort of like buying a remastered version of your favorite analog recording, then finding out it was digitally mixed/remastered, which negates the whole purpose of a fully analog signal chain. In addition to that, it would have likely been compressed and EQed to death as well, so the audio has already been mutilated, which you're permanently stuck with. The analog stages in the DSP may have a low noise floor, but the artifacts and errors are the worst part, not so much the hiss and noise. Using a higher sample rate won't really matter either if the top end of the audio has originally been sampled at regular 44.1k (or 48k) after being run through the unavoidable brick wall lowpass filter.
Analog crossovers may be noisier and limited in processing power, but I'll take those things any day (within reason) over the digital (non musical) junk the floats on top of the original signal. Don't get me wrong. There are many decent sounding digital recordings out there, but these are far fewer in comparison to all the great analog recordings. Most high end audio lovers of 60s, 70s, 80s, music will tell you that the original analog recordings are typically far better sounding than their digital re-mastered versions. There are of course some exceptions, but again, very few in comparison. If you lack good vinyl playback capability, you're definitely missing out on some amazing stuff. No it's not as convient as streaming digital source, but its definitely a more musical listening experience when dealing with originally analog recorded music.
You guys who only play back digitial source have the advantage of already being done with the first conversion step. The studio that mastered the recording will likely have access to the best gear available to minimize this issue, but those errors (hopefully not to many) are basically already there and you're stuck with them. You only have to deal with the last conversion step back to analog, which isn't as hard to get right as A/D conversion. That makes DSP so practical since you're already dealing with digital source, but there will however always be time domain errors (jitter) built into it along with processing artifacts.
Us analog source guys (specifically vinyl) don't want to risk screwing up our carefully derived audio with two digital conversion processes, plus extra DSP. Its sort of like buying a remastered version of your favorite analog recording, then finding out it was digitally mixed/remastered, which negates the whole purpose of a fully analog signal chain. In addition to that, it would have likely been compressed and EQed to death as well, so the audio has already been mutilated, which you're permanently stuck with. The analog stages in the DSP may have a low noise floor, but the artifacts and errors are the worst part, not so much the hiss and noise. Using a higher sample rate won't really matter either if the top end of the audio has originally been sampled at regular 44.1k (or 48k) after being run through the unavoidable brick wall lowpass filter.
Analog crossovers may be noisier and limited in processing power, but I'll take those things any day (within reason) over the digital (non musical) junk the floats on top of the original signal. Don't get me wrong. There are many decent sounding digital recordings out there, but these are far fewer in comparison to all the great analog recordings. Most high end audio lovers of 60s, 70s, 80s, music will tell you that the original analog recordings are typically far better sounding than their digital re-mastered versions. There are of course some exceptions, but again, very few in comparison. If you lack good vinyl playback capability, you're definitely missing out on some amazing stuff. No it's not as convient as streaming digital source, but its definitely a more musical listening experience when dealing with originally analog recorded music.
Maybe I'm just behind the times and I need to leave everything that has worked for thus far behind and embrace DSP. And maybe if I have couple grand disposable income laying around I will go right out and get me one of those high end units you mention, even though mini DSP can be got for under a grand, and maybe it will be the start of a beautiful relationship but then again maybe I don't like the way it sounds and I sell it on e-bay. That's a lot of maybes. 🙄😀
I dunno....profiguy said:
I can agree some of the magic from vinyl and really good moving coil cartridges... and a premium analog signal chain...is missing from my best digital source / dsp systems.
But those analog systems also lack some of the the magic from my best digital source and dsp systems.
I keep both types of systems running, to keep my ears open.
Honestly, for me it's almost album by album, or CD by CD, track by track, which type system sounds best.
Most of the time, i think it comes down to the system that matches a track's inherent SPL and dynamics needs for convincing reproduction.
9 times out of 10 lately, i prefer the clean robust power of my dsp systems.
Maybe has more to do with my music type preference than anything.
Like said...easy to say dunno..
Member
Joined 2009
Paid Member
I do think its a preference thing for most of us. There are tradeoffs with both and from a convenience aspect, digital will always be superior.
It gets kind of frustrating when you have multiple copies / pressings of various albums and trying to decide which gives the best overall delivery of the musical experience. More often than not the recording just simply sucks and there's not much you can do. I'm not the kind of high end music guy that listens to crappy music (or even test signals LOL) just because its recorded well or it shows off extreme dynamics. I'd rather listen to a so-so recording of really good music (within reason).
I've given up on trying to have one system do both to my satisfaction. Having separate high end systems is cost prohibitive. I can see why people just settle for digital since its much more convenient and user friendly. I just have too much tied up into my vinyl collection to ignore and decent analog playback gear is very expensive, plus the selection of really good mid priced cartridges is disappearing... I refuse to pay 5 figure sums for a cartridge. My most expensive MC is $2000 and thats on an equally priced turntable minus the cost of the arm. Then there's record cleaning and maintenance... not a cheap hobby.
It gets kind of frustrating when you have multiple copies / pressings of various albums and trying to decide which gives the best overall delivery of the musical experience. More often than not the recording just simply sucks and there's not much you can do. I'm not the kind of high end music guy that listens to crappy music (or even test signals LOL) just because its recorded well or it shows off extreme dynamics. I'd rather listen to a so-so recording of really good music (within reason).
I've given up on trying to have one system do both to my satisfaction. Having separate high end systems is cost prohibitive. I can see why people just settle for digital since its much more convenient and user friendly. I just have too much tied up into my vinyl collection to ignore and decent analog playback gear is very expensive, plus the selection of really good mid priced cartridges is disappearing... I refuse to pay 5 figure sums for a cartridge. My most expensive MC is $2000 and thats on an equally priced turntable minus the cost of the arm. Then there's record cleaning and maintenance... not a cheap hobby.