When you remove the upsampler you have two rows, one from the receiver and one to the dac, the pin from the receiver are output signal for data, word clock, bit clock and master clock, the pin to the dac are the respective input signal. The clock signals is generated from a PLL internal to the receiver and are synchronous to the incoming spdif signal.
If you want to send a high quality local clock to the dac you have to generate it via a VCXO and lock it to the clock signal coming from the receiver with a PLL.
Another solution is to use a free running XO to clock the dac, use the receiver in slave mode (after changing configuration bit on the 8416) and send the same clock back to the source.
Anyway these solution are limited to a fixed sample frequency unless you use two clock (plus shift registers to obtain the right bit clock and word clock), one for the 44.1KHz based frequencies and one for the 48KHz ones, and a "mechanism" to switch between these.
An implementation of the second solution (a very rude one by me) is on this thread
http://www.diyaudio.com/forums/digital-line-level/140232-slaving-audio-board-mck-dac-via-spdif.html
in post #47 a bug fix relevant for this dac.
Ciao
Andrea
Hello Andrea
more or less that's what I wanted to do. The problem here is that there is a lack of high quality standards for source-dac communication. When big companies stopped to do research in digital audio, Hifi became a small-whorkshops-affair.... so you put a cheap-chip in a beautiful box, two ECC88, a big trafo, several huge caps... and then sell at high-end prices!
I am not a digital guy but reading and studying the problem my conclusion is that the best way to do play with digital is the following:
1. output everything at a fixed frequency using (if needed) a good resampler like SOX
2. then have control to the the clock between receiver and DAC.
3. avoid built-in filters in the dac... that's why resampling on the PC always sounded better to me.
In principle this is what I would do. But with this DAC something goes wrong when I remove the resampler, and I think this is because this resampler has some effect on the clock presented at the DAC, do you think so?
yes I read your interesting posts mentioned above but I couldn't join the discussion because I don't have the necessary backgorund. You were ritght in principle but imho the solution proposed by UnixMan to fix the spdif frequency at the computer is still the only feasible simple solution. In order to do things properly to me the only feasible alternative is the ALTMANN Attraction DAC even though is too expensive for what is on the board!
Best Wishes
Pietro
more or less that's what I wanted to do. The problem here is that there is a lack of high quality standards for source-dac communication. When big companies stopped to do research in digital audio, Hifi became a small-whorkshops-affair.... so you put a cheap-chip in a beautiful box, two ECC88, a big trafo, several huge caps... and then sell at high-end prices!
I am not a digital guy but reading and studying the problem my conclusion is that the best way to do play with digital is the following:
1. output everything at a fixed frequency using (if needed) a good resampler like SOX
2. then have control to the the clock between receiver and DAC.
3. avoid built-in filters in the dac... that's why resampling on the PC always sounded better to me.
In principle this is what I would do. But with this DAC something goes wrong when I remove the resampler, and I think this is because this resampler has some effect on the clock presented at the DAC, do you think so?
yes I read your interesting posts mentioned above but I couldn't join the discussion because I don't have the necessary backgorund. You were ritght in principle but imho the solution proposed by UnixMan to fix the spdif frequency at the computer is still the only feasible simple solution. In order to do things properly to me the only feasible alternative is the ALTMANN Attraction DAC even though is too expensive for what is on the board!
Best Wishes
Pietro
just as info Snow the designer of the big dac and has alot of other stuff is here ....
http://www.hifidiyshop.com
http://www.hifidiyshop.com
Damned near all of them, it is the native format used internally in 98% of all digital audio equipment. SPDIF was created as an easy interface means and has many inherent problems. To me, it doesn't make sense to convert the native format to another format of questionable precision just to transmit it a meter or so over a cable of questionable precision, then convert it back to the original format using a receiver chip.
I admit that buffering an I2S signal is much more involved than merely connecting two units together, but sending clock signals back and forth from transport to dac just to improve the SPDIF signal would be like polishing a car painted with primer paint. It is much easier in the long run to eliminate problems, not improve them.
I admit that buffering an I2S signal is much more involved than merely connecting two units together, but sending clock signals back and forth from transport to dac just to improve the SPDIF signal would be like polishing a car painted with primer paint. It is much easier in the long run to eliminate problems, not improve them.
The cdm-4 is a mechanism, the I2S originates in the decoder chip used in the player to process the raw data coming from the laser. Separating a mechanism from a player is not a viable option, you need the control circuitry. You can gut the player of all audio circuitry and use what you wish but the player functions must remain intact.
This is a question for Bill fuss, but anybody else who has knowledge of this feel free to chip in.
Bill, I remember reading that you tapped the I2S signal from a transport (Marantz?) to your DAC. Is it as simple as finding the I2S outputs on the transport decoder and sending them via a 6 (?) pin connector to the relevant DAC inputs?
I take it I2S is reasonably jitter free - and what about the clocking and sampling circuits, do they still come into play?
Sorry if you have already posted these details - just point me to the post if so.
Cheers
Bill, I remember reading that you tapped the I2S signal from a transport (Marantz?) to your DAC. Is it as simple as finding the I2S outputs on the transport decoder and sending them via a 6 (?) pin connector to the relevant DAC inputs?
I take it I2S is reasonably jitter free - and what about the clocking and sampling circuits, do they still come into play?
Sorry if you have already posted these details - just point me to the post if so.
Cheers
Hi Rich,
A great question. Yes, it is that simple IF you are using upsampling, because the upsampler is asynchronous and reclocks the signal. If you aren't then you need to send the master clock to the dac chip also. Normally, it would be supplied by the regenerated clock from the receiver chip.
And just 4 wires are necessary with upsampling, data, bit clock, and LRCK or word select, plus ground, if you can keep the wires short, like 2 or 3 inches.
The I2s signal is the purest form you can get AFAIK, pretty much jitter free. I basically built a CDplayer with a Philips based transport using the existing control circuitry and replacing all the audio circuitry. It's all in the original case and it smokes everything else I've ever owned.
A great question. Yes, it is that simple IF you are using upsampling, because the upsampler is asynchronous and reclocks the signal. If you aren't then you need to send the master clock to the dac chip also. Normally, it would be supplied by the regenerated clock from the receiver chip.
And just 4 wires are necessary with upsampling, data, bit clock, and LRCK or word select, plus ground, if you can keep the wires short, like 2 or 3 inches.
The I2s signal is the purest form you can get AFAIK, pretty much jitter free. I basically built a CDplayer with a Philips based transport using the existing control circuitry and replacing all the audio circuitry. It's all in the original case and it smokes everything else I've ever owned.
Excellent Bill thanks.
Yeah, I forgot that you fitted the DAC into the case with the transport... 2 or 3 inches might be a problem. How important do you think the wire length is? Obviously jitter / RF is added with longer wires?
I use a PC as source and initially wanted to try I2S output from a suitable soundcard (ESI Juli@ or RME) so would mean longer interconnects. Surely there is a solution for sending to an external DAC?
Yeah, I forgot that you fitted the DAC into the case with the transport... 2 or 3 inches might be a problem. How important do you think the wire length is? Obviously jitter / RF is added with longer wires?
I use a PC as source and initially wanted to try I2S output from a suitable soundcard (ESI Juli@ or RME) so would mean longer interconnects. Surely there is a solution for sending to an external DAC?
I guess that that's very diffcult to do unless you buffer the I2S... which again it's not that easy to do. A friend of mine built a mini-itx fanless pc into an ATX case, and he used the space to place the DAC board into the case near the juli@ from which he took the I2S.
IMHO in order to easily work with the I2S you need something which cane be placed near the dac unless you have the experience to buffer the I2S.
This USB receiver board could be a solution
USB Audio 2.0 Reference Design | XMOS
and they comment on this here
Youtube xmos interview
What I understood is that this is MAC OS X native but Win drivers are available *upon request*!... Linux of course is not even mentioned!
Best
Pietro
Last edited by a moderator:
You know I'm merely a hacker with this stuff. When I was first kludging it together I was using a 4 wire twisted bundle about 12" long and it worked fine, but everything I read about I2S said keep it as short as possible. My idea was to put everything in one case anyway so naturally I didn't pursue trying anything any longer.
A PS Audio guy posted the buffer circuit that they use for I2S transmission just a couple months ago. They use a couple chips for transmitter and receiver and a regular cat5 or 6 can be used for cabling. It looked relatively painless. You should be able to search for it, try (PS Audio I2S). He was also very gracious in his postings, offering help to anyone who would like to try it. Nice guy.
I also replaced the crystal in the player with a Tent Labs XO module. It seemed to make a real difference but that is a purely subjective observation. It seems that every improvement I made was in the direction of microdynamics, the soundstage kept getting wider, taller and deeper. Closer and closer to holography, pretty cool !
I have a fairly well treated listening room so my results probably cannot be duplicated in a normal living room environment.
I have a fairly well treated listening room so my results probably cannot be duplicated in a normal living room environment.
Last edited: