OK, I've seen a few posts on audioasylum about this thing, but have been unsuccessful in finding information on this forum. Anybody ever heard of it, or tried it? The reviews I've read are pretty convincing... but not enough to spend $500 on it.
I'm wondering if the amazing minds over here could possibly reverse engineer the concept... The concept makes sense to me, but I just can't figure out how passive device can first scramble the jitter, and then proceed to modulate it. How does one 'isolate' jitter? It's not really "part" of the signal, but rather, it "IS" the signal. I'm confused...
Perhaps this is the answer to the jitter problems of SPDIF ???
www.jitter.de/english/jisco.html
Please advise
I'm wondering if the amazing minds over here could possibly reverse engineer the concept... The concept makes sense to me, but I just can't figure out how passive device can first scramble the jitter, and then proceed to modulate it. How does one 'isolate' jitter? It's not really "part" of the signal, but rather, it "IS" the signal. I'm confused...
Perhaps this is the answer to the jitter problems of SPDIF ???
www.jitter.de/english/jisco.html
Please advise
I would say one of these things would be fairly easy to build yourself, at a very small fraction of 500$.
My guess is you need a 40 MHz clock generator (derived from the 25nS delay time in the datasheet), a 74HC74 flip flop, and a 74HC04 for input / output signal conditioning.
Total cost of parts: 3-4$ plus the nice box and BNC plugs.
My guess is you need a 40 MHz clock generator (derived from the 25nS delay time in the datasheet), a 74HC74 flip flop, and a 74HC04 for input / output signal conditioning.
Total cost of parts: 3-4$ plus the nice box and BNC plugs.
I figured it would be easy to do...
Until now, I thought that it was a passive device. But, now I see that there is an "external power supply" in the spec sheet discription. This makes me feel much better.
Lars, could you elaborate a little more on the application of the flip-flop? I'm assuming he's just doing amplitude modulation with a summing circuit...
I still don't understand how the jitter is "isolated" to be modulated. The whole signal would be modulated from what I can see, and while the PLL will have the high frequency noise reduction to cancel out the 40MHz, the (primarily large) 10kHz jitter would still pass through. What am I missing...???
Thanks,
Until now, I thought that it was a passive device. But, now I see that there is an "external power supply" in the spec sheet discription. This makes me feel much better.
Lars, could you elaborate a little more on the application of the flip-flop? I'm assuming he's just doing amplitude modulation with a summing circuit...
I still don't understand how the jitter is "isolated" to be modulated. The whole signal would be modulated from what I can see, and while the PLL will have the high frequency noise reduction to cancel out the 40MHz, the (primarily large) 10kHz jitter would still pass through. What am I missing...???
Thanks,
From the description it seems clear, that the principle of this device is:
'If you can't remove the jitter, add some instead.'
And it might have an effect, because the jitter is not modulated with music signal, but overloaded with 'pink time noise', resulting from the modulation between the 40 MHz and the asyncronous SP/DIF signal.
I don't know for sure if it works this way, but it appears so, if you read between the lines of the data sheet.
It's a little bit like a Triode amplifier, the harmful (high order) distortions are masked by 10 times louder nice sounding (low order) distortions.
You might get out of the 'phonebooth effect' with this unit, however i don't think it is the solution for a wide and precise soundscape. For this you need to remove the jitter all together (which is very possible).
'If you can't remove the jitter, add some instead.'
And it might have an effect, because the jitter is not modulated with music signal, but overloaded with 'pink time noise', resulting from the modulation between the 40 MHz and the asyncronous SP/DIF signal.
I don't know for sure if it works this way, but it appears so, if you read between the lines of the data sheet.
It's a little bit like a Triode amplifier, the harmful (high order) distortions are masked by 10 times louder nice sounding (low order) distortions.
You might get out of the 'phonebooth effect' with this unit, however i don't think it is the solution for a wide and precise soundscape. For this you need to remove the jitter all together (which is very possible).
I agree with part of what he says......
The part about you have to have low jitter at the DAC.
Even if his "decorrelates" the incoming signal, the PLL in the RX will generate its own anyway.
The only real way to do this is with a secondary PLL with a very stable VCXO as its reference.
Takes a few flip-flops, a good op-amp, clean VCXO, knowledge of PLLs..............
Jocko
The part about you have to have low jitter at the DAC.
Even if his "decorrelates" the incoming signal, the PLL in the RX will generate its own anyway.
The only real way to do this is with a secondary PLL with a very stable VCXO as its reference.
Takes a few flip-flops, a good op-amp, clean VCXO, knowledge of PLLs..............
Jocko
For those who can make something of it: on the Jisco website (www.jitter.de) you can find a picture of how the device looks from the inside. It's in the pdf-file 'jitter artikel II' marked as 'new', top-right on the main page. (The text in that document is a nice explanation of the jitter phenomenon, but only in German.)
Very interesting ...
I found the Quartx oscillator on Google, it appears to be a standard 100 ppm TTL oscillator. Costing €1.45 = $1.75.
http://www.hofstaedtler.at/sonder/oszillatoren.htm
Two chips for SP/DIF signal recovery, presumably 74HCU04 or 74HC04, and the GAL 16V8 (www.latticesemi.com) for flip-flopping. (These programmable devices are often used only with the intent to improve the mystification factor, hence stopping diy'ers from copying the circuit. It is a programmable Logic Array, but the fact is, the device is so simple, that the program can pretty much only be used to direct input signals to the 8 internal logic gates / flip flop's - no big deal).
Total cost of active parts ... around 7-8 US$
Plus the nice box still ..)
I found the Quartx oscillator on Google, it appears to be a standard 100 ppm TTL oscillator. Costing €1.45 = $1.75.
http://www.hofstaedtler.at/sonder/oszillatoren.htm
Two chips for SP/DIF signal recovery, presumably 74HCU04 or 74HC04, and the GAL 16V8 (www.latticesemi.com) for flip-flopping. (These programmable devices are often used only with the intent to improve the mystification factor, hence stopping diy'ers from copying the circuit. It is a programmable Logic Array, but the fact is, the device is so simple, that the program can pretty much only be used to direct input signals to the 8 internal logic gates / flip flop's - no big deal).
Total cost of active parts ... around 7-8 US$
Plus the nice box still ..
)
Jisco Jitter decorrelator
Hi Wim,
for more information see:
http://l2.espacenet.com/espacenet/viewer?PN=DE19954248&CY=gb&LG=en&DB=EPD
and
http://l2.espacenet.com/espacenet/bnsviewer?CY=gb&LG=en&DB=EPD&PN=DE10006880&ID=DE++10006880A1+I+
Persononally I believe any jitter reduction device should be placed after the digital inputreceiver and also after the digital filter, if used, not before.
Unfortunately I did not find how exactly the Jisco works but I presume just as my Asynchronous Reclocker with a 40 MHz oscillator and a few flip-flops.
http://home.student.utwente.nl/f.s.bouwman/audio/images/nonoz-schematic.gif
and
http://www.geocities.com/react_71/audio/
(schematic will be soon on Pedja's site)
The decorrelation is a nice theory why the asynchronous reclocking improves the sound.
Still better is, I believe, to avoid the SPDIF interface all together using I2S direct or a one-box player.
http://www.diyaudio.com/forums/showthread.php?postid=249220#post249220
Hi Wim,
for more information see:
http://l2.espacenet.com/espacenet/viewer?PN=DE19954248&CY=gb&LG=en&DB=EPD
and
http://l2.espacenet.com/espacenet/bnsviewer?CY=gb&LG=en&DB=EPD&PN=DE10006880&ID=DE++10006880A1+I+
Persononally I believe any jitter reduction device should be placed after the digital inputreceiver and also after the digital filter, if used, not before.
Unfortunately I did not find how exactly the Jisco works but I presume just as my Asynchronous Reclocker with a 40 MHz oscillator and a few flip-flops.
http://home.student.utwente.nl/f.s.bouwman/audio/images/nonoz-schematic.gif
and
http://www.geocities.com/react_71/audio/
(schematic will be soon on Pedja's site)
The decorrelation is a nice theory why the asynchronous reclocking improves the sound.
Still better is, I believe, to avoid the SPDIF interface all together using I2S direct or a one-box player.
http://www.diyaudio.com/forums/showthread.php?postid=249220#post249220
You can also remove jitter from the SP/DIF line in another way:
1..Mount the ref clock in the DAC (rather than the transport).
Now the DAC has low jitter clean clocking.
2..Sync the transport with a coax running from the DAC to the
transport. Add 75 Ohms resistors on in-a nd out put of cable to avoid standing waves.
It works great IF your transport and DAC run on the same frequencies, like 11.2896 MHz ( the most common for DAC's) or 16.9344 MHz if your DAC uses the YM3623 receiver chip. (The receiver chip decides the operating clock frequency of the DAC) All Crystal (Now Cirrus Logic) and Philips receivers run 11.2896 MHz.
In some cases it works also when the transport and DAC run on different clock frequencies, but that is the twilight zone. Be careful, this should only be attempted by experienced diy'ers.
1..Mount the ref clock in the DAC (rather than the transport).
Now the DAC has low jitter clean clocking.
2..Sync the transport with a coax running from the DAC to the
transport. Add 75 Ohms resistors on in-a nd out put of cable to avoid standing waves.
It works great IF your transport and DAC run on the same frequencies, like 11.2896 MHz ( the most common for DAC's) or 16.9344 MHz if your DAC uses the YM3623 receiver chip. (The receiver chip decides the operating clock frequency of the DAC) All Crystal (Now Cirrus Logic) and Philips receivers run 11.2896 MHz.
In some cases it works also when the transport and DAC run on different clock frequencies, but that is the twilight zone. Be careful, this should only be attempted by experienced diy'ers.
Over my head...
That patent was fun to see, I'll see if I can try to understand anything of it... but I'm afraid most of the digital design stuff here is over my head.
I've actually emailed with the Jisco maker a few times, he seemed knowledgeable and confident in his product. And even though it's expensive (which has held me back so far), the few responses that can be found on the web have all been positive. Might still give it a try.
Currently I've got a digital chain of Marantz CD50 -> Monarchy DIP -> Behringer Ultracurve -> Behringer Ultramatch -> ART DI/O dac.
The Ultramatch is going out as soon as I've ordered the proper coax to make an aes/ebu -> s/pdif adapter cable, it's an ASRC that locks jitter (I've confirmed that by putting some of those gadgets in a different order). I'm gradually modding the di/o, now halfway. The cd50 will be tackled soon, with the usual transport mods to start off with, just ordered the service manual.
I want to keep the DIP in for now, and the Ultracurve equalizer is badly needed with my terrible room acoustics. With such a long chain, I guess that rules out any option of I2S-like solutions, right?
That patent was fun to see, I'll see if I can try to understand anything of it... but I'm afraid most of the digital design stuff here is over my head.
I've actually emailed with the Jisco maker a few times, he seemed knowledgeable and confident in his product. And even though it's expensive (which has held me back so far), the few responses that can be found on the web have all been positive. Might still give it a try.
Currently I've got a digital chain of Marantz CD50 -> Monarchy DIP -> Behringer Ultracurve -> Behringer Ultramatch -> ART DI/O dac.
The Ultramatch is going out as soon as I've ordered the proper coax to make an aes/ebu -> s/pdif adapter cable, it's an ASRC that locks jitter (I've confirmed that by putting some of those gadgets in a different order). I'm gradually modding the di/o, now halfway. The cd50 will be tackled soon, with the usual transport mods to start off with, just ordered the service manual.
I want to keep the DIP in for now, and the Ultracurve equalizer is badly needed with my terrible room acoustics. With such a long chain, I guess that rules out any option of I2S-like solutions, right?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.