I'm not talking about latency. I'm talking about clock skew between endpoint. IEEE1588v2 is the closest you will get to having a single master clock for all channels, only clocking multiple end points in separate locations. But of course this is only 1 of the several advantages.
You could get in touch with Hypex to discuss this.
I am not into DSP active loudspeakers, but my gut feeling is that
- building and tuning an active loudspeaker is not per se "easier" than classic passive designs;
- the difference between the different DSP platforms may become less important when the active design is not optimal to begin with.
These are mere tools; when not used properly they will not change game in speakers IMHO.
I am not into DSP active loudspeakers, but my gut feeling is that
- building and tuning an active loudspeaker is not per se "easier" than classic passive designs;
- the difference between the different DSP platforms may become less important when the active design is not optimal to begin with.
These are mere tools; when not used properly they will not change game in speakers IMHO.
And these tools make it possible to perfectly dial in the crossovers using a GUI, rather than switching out passive components. It allows for perfect phase response, time alignment, room correction, and the multitude of other advantages of taking the passive components out of the signal path. But of course, you need to have the ability to use tools properly. Just like you do with anything.
Another option (although not feasible for these plates) to take advantage of IEEE1588v2 is to use the latest generation of SHARC 589 series DSP chips. They have the protocol built right in. So with a simple link cable between units, you can take advantage of the nanosecond clock skew sync, without paying any licencing fees to 3rd party AES67 compatible OEM board suppliers.
ADSP-SC589 Datasheet and Product Info | Analog Devices
Only problem is you need to have the knowledge to program it.
ADSP-SC589 Datasheet and Product Info | Analog Devices
Only problem is you need to have the knowledge to program it.
You'll also need a special chair that clamps your head to appreciate your perfectly aligned system to it's full potential
Of course. You can give a man a Lamborghini, but this doesn't mean the car is poor if he doesn't know how to drive it. Same with a laser for doing brain surgery.
Last edited:
You'll also need a special chair that clamps your head to appreciate your perfectly aligned system to it's full potential
Sounds like old fashioned passive based systems. With advanced DSP you can tune the system for perfect sound from any listening position
You'll also need a special chair that clamps your head to appreciate your perfectly aligned system to it's full potential
Well, yes, considering one nanosecond corresponds to less than a thousandth of a millimeter.
Sounds like old fashioned passive based systems. With advanced DSP you can tune the system for perfect sound from any listening position
https://www.youtube.com/watch?v=Oz8RjPAD2Jk
Based on my limited knowledge, Bruno P. and affiliates (Kii, Grimm, etc.) could make great drive even from barn door
Did you read whitepapers on Grimm's site? Or eventually did you see DLCP datasheet? They all syncing their high end targeted DSP's to different-than-digital audio frequencies and that Ethernet interconnection, I suppose, is synchronising local oscillators in the PLL. Is there a place for errors? I think no, they probably know what they are doing. Unfortunately to DIY'ers, I can't see this Ethernet sync in these new Fusion amps so that is why it cannot be named as a "game changer" to me. Just ordinary DSP chip and syncing to external clock only, with a glimpse of new NCore amps which, indeed, are astonishig, I know that.
Last edited:
Based on my limited knowledge, Bruno P. and affiliates (Kii, Grimm, etc.) could make great drive even from barn door
Did you read whitepapers on Grimm's site? Or eventually did you see DLCP datasheet? They all syncing their high end targeted DSP's to different-than-digital audio frequencies and that Ethernet interconnection, I suppose, is synchronising local oscillators in the PLL. Is there a place for errors? I think no, they probably know what they are doing. Unfortunately to DIY'ers, I can't see this Ethernet sync in these new Fusion amps so that is why it cannot be named as a "game changer" to me. Just ordinary DSP chip and syncing to external clock only, with a glimpse of new NCore amps which, indeed, are astonishig, I know that.
I have no doubt that they utilized the technology they are using to it's full capabilities. But I also have confidence the same great minds would be able to use superior technology to it's best ability as well. The clock sync advantage is only 1 of the several advantages to making these plates AES67 compatible.
I like the idea of simply plugging each speaker into an Ethernet cable and having it pop up as a sound device on my computer. You can run Ethernet cables through the walls, and install one of those wall plates with both AC and Ethernet jacks on them right behind each speaker. Simply plug in 1 Ethernet cable, and the AC cable to the wall behind the speakers and you're in business. 0.25ms latency from the VSC on the computer server, until the I2S outputs on the Dante Broadway chip. No Netflix sync errors to deal with or anything.
And there's no denying that having CMOS I2S being fed direct from the Dante chip to the inputs of the DSP mere millimeters away wouldn't be superior to using a server-USB cable-USB to AES/EBU bridge-PLL generated clock-XLR cable-PLL to I2S decoder chip- SRC chip-DSP. Then you can also use 1 master to clock the Dante chip, DSP and DAC chip in each plate.
Last edited:
It sounds good but unfortunately it rarely will meet consumer audio level, especially on audiophile market which still is dwelving into cable / capacitor dilemmas, rather than looking ahead technology progression so fast, that even big gamers like Hypex can't keep up to this. In combination with Fusions targeted to minor OEM's (yes, domestic and still 'audiophile') I can understand Hypex that there is no point in developing AES67 into that plates, unfortunately to us. They probably come up with something spectacular with their mysterious 2by4 adressed, I think, to big names in pro industry.
They probably come up with something spectacular with their mysterious 2by4 adressed, I think, to big names in pro industry.
It sounds good but unfortunately it rarely will meet consumer audio level, especially on audiophile market which still is dwelving into cable / capacitor dilemmas, rather than looking ahead technology progression so fast, that even big gamers like Hypex can't keep up to this. In combination with Fusions targeted to minor OEM's (yes, domestic and still 'audiophile') I can understand Hypex that there is no point in developing AES67 into that plates, unfortunately to us.
Well AES67 is here to stay. It's a new standard adopted by the AES, to replace all of the old outdated standards. It's not going anywhere. I really don't see any advantage to being a late adopter.
Then a dual DLCP, Grimm LS1 or Kii Three would not work either......
I've been studying how the Grimm LS1 does things. Similar setup, but done with very great care. The master clocks have a very low phase noise, and high stability. Then the link between the left and right channels is done with a high quality CAT 5 cable. So because of the great care taken, and high quality components used, the speaker clocking should be fairly even. But at the end of the day I don't see any measures to ensure perfect sync between the 2 clocks, like there is with IEEE 1588v2. It's basically putting 100% faith that the clocks are dead nuts and won't drift apart over time. I think the system could be improved upon using Ravenna. Also greatly simplified.
Attachments
In a stereo pair of speakers, what would be the actual problem/effect of a minor long term drift in clock?
Within each speaker everything operates off the same stable clock.
I can't see why the equivalent of a very slow movement of one speaker should matter
When there's 2 master clocks timing the system, you don't have the natural flow the same as if you're using just 1 master clock. Each speaker is beating with a different heartbeat. When you take great care and use incredible clocks like in the LS1 it's not so bad. But if you're using cheap clocks it is. Even with clocks like the best Crystek's phase noise can vary up to 10db between them. Unless individually measured and perfectly matched you're playing Russian roulette. Clocks like what Grimm uses is a must. But if you combine clocks of that caliber, with the assistance of IEEE 1588v2, this is next best thing to having a single master clock possible circa 2017.
The holy grail is a single master clock, but there's a problem with having a single master clock in a active system like this. It's impossible to clock the DSP, and DAC chip's with only millimetres of PCB traces right before the amplifiers in each speaker. If you want to do things the very best way, you don't want to use PLL's. PLL's add major jitter, and you end up wasting a good clock. With IEEE1588v2, the PLL is only used to sync the clocks, and keep the hearts beating within a few nanoseconds of each other. It doesn't add any jitter as each DSP/DAC is still clocked CMOS direct off its own master.
This is my understanding anyways. But if someone much smarter than me wants to correct me, that's fine. There's a lot of folks on this forum who know far more about clocking than me.
Last edited: