No reclocking DAC. Are they fine?

[beppe61]

Dear Sirs,


I read somewhere that some external dacs don't reclock the incoming digital signal ( they are slave to the transport).
Someone say that this could give a more natural sounding dac.
I would like to hear some experts' opinion.

Thank you very much indeed.

Kind regards,

beppe61

[beppe61]

Dear Sirs,


I would like to hear some comments about the reclocking process.
What is your experience?
I have an external dac (Behringer SRC 2496) and prefer the sound without reclocking (i.e. using the clock embedded in the incoming SPDIF digital signal).
It seems indeed more natural and less digital to me.

Kind regards,

beppe61
ITALY

[danb1974]

If the source has a stable clock and low jitter on output, then it's no problem sync'ing on it.

If not, better get a dac with buffer and internal clock.

The idea is to get the bits to the dac at the exact time they are supposed to, no matter which method you use to accomplish this.

[Guido Tent]

Originally posted by danb1974
If the source has a stable clock and low jitter on output, then it's no problem sync'ing on it.

If not, better get a dac with buffer and internal clock.

The idea is to get the bits to the dac at the exact time they are supposed to, no matter which method you use to accomplish this.

It IS a problem because the way SPDIF is coded, much jitter is induced in the interface, so a wide bandwidth PLL at the DAC is a disaster, no matter how clean the drive is.

The conversion process is of utmost importance as jitter induces distortion in the DAC, hence the DAC clock should be of lowest jitter possible. In addition all signals entering the DAC chip should be reclocked as to reduce on-chip crosstalk.

best

[stolbovoy]

Originally posted by beppe61
I read somewhere that some external dacs don't reclock the incoming digital signal ( they are slave to the transport).

Actually, most of them

Originally posted by beppe61
I would like to hear some comments about the reclocking process.
What is your experience?

What I got implementing synchronous reclocking:
1) Resolution enhancement.
2) My computer 10m away and CD transport sound identically.
3) Digital cables sound the same

[EC8010]

Originally posted by stolbovoy
3) Digital cables sound the same

I thought that, then had to lengthen a digital (S/PDIF) cable, and it didn't sound as good. As an engineer, I thought, "It's only 1s and 0s, so why the difference?" So I looked at the characteristic impedance of the cable and its termination. Whoops. The connection was generating jitter because of the Manchester coding violation alluded to by Guido plus mismatched impedances. What you need is a proper cable with guaranteed matched source and desitination impedances; then they'll all sound the same.

[stolbovoy]

The connection was generating jitter because of the Manchester coding violation alluded to by Guido plus mismatched impedances.

Typically jitter does nothing with Manchester encoding violation. One will receive the same data as sent with almost any junk cable/termination at short distances. Problem is in a clock recovery.

[Jocko Homo]

Try using a longer cable, with proper terminations. Then tell me that you don't hear a difference.

Jocko

[beppe61]

Dear Sirs,


I want to thank you all for your extremely kind and very valuable reply.
For the first time I get very valuable responses to my doubts.
Sadly the task to connect adequately a transport and a dac seems very difficult to make.
Actually I hoped it was simpler.
Anyway you have been very helpful for me understanding something more of the phenomena.

Thank you very much sincerely.

Kind regards,

beppe61

[bocka]

"It's only 1s and 0s, so why the difference?"

Every bandlimited system with non static input data signals induces jitter, because the non harmonic part of the differential equation is not zero or better said not equal for every time interval. Of course there are other differences like signal dispersion, unstable reference points by noisy power supplies, ground bouncing and probably some more.

Timing also in the digital domain is an analog phenomenon as there are no electrical '0's and '1's but only (analog) signal levels.