solderdude
Grand Contributor
These are separate questions.
You would have to understand how this signal is made.
Basically you have data and clock.
The clock is a continuous changing signal always in the same frequency and it basically tells the receiver at which moment it should look in the data signal to decide the data is 1 and 0.
The data can have consequtive 0's and '1's in a row. For instance 0000110001010011111010.
The only way to tell whether in a row of consequtive 0's or 1's there was a 0 or 1 is by usage of the clock.
That clock signal tells you WHEN to look for the signal level is representing a 0 or a 1.
This is when the signal is stabile so not near the rising or falling edges but in the center of the bit.
So you need a clock and data stream so 2 cables.
The clock is easy to retrieve as the signal changes constantly and in a fixed manner.
The data not so.
Technically it is easier to receive a constantly varying signal than one that flips at different times and lengths (data)
Folks wanted 1 cable. Would have been easier to design a 'stereo' cable with 2 different plugs but it was decided to do this in one cable as that could be a standard 50 or 75 Ohm shielded cable and easy to convert to optics and or modulate a single signal instead of 2 (or more as left and R signals and extra data is also in the data).
So technical solution is to 'combine' the clock and data signals in a nifty way.
In the receiver the signal thus changes constantly but differently for a 0 or a 1.
This way it is easy to retreive the clock, by just synchronising a clock on the receiver end on the rising and/or falling edges of a signal.
clock is thus easily 'recovered' using simple electronics.
Also when a clock is recovered and due to the way the 0's and 1's are embedded (encoded) in the combined signal it is easy to retrieve what was a 0 and 1.
Data thus also recovered.
There are different types of en/decoding these types (there are severall ways to do this) of signals SPDIF is one of them.
so no guessing, no algorithms, just simple electronics that can unravel a re-conditioned datastream.
In reality there are quite a few technically different ways to recover and synchronise the clock each with stronger and less strong benefits.
Clocks on the receiving end differ with those on the transmitting end and here too many different technical solutions exist. One better than the other in some areas.
Has been answered above. A decision was made and everyone now has to comply or come up with other formats (HDMI, USB, I2S etc)
This is so all digital processing, DAC's, ADC's all sample at exactly the same time and all data words (so 16 or 24 bits) are generated/read at the exact same time.
When this was not done the drift in each clock of every 'device' would make them all different in frequency (very slightly) and timing.
So the clock here is only needed for this task.
I still don't get this reclocking thing. How does the dac recover the clock, it receives the 0 and 1 and it guess the timing, just by some algoritme?
You would have to understand how this signal is made.
Basically you have data and clock.
The clock is a continuous changing signal always in the same frequency and it basically tells the receiver at which moment it should look in the data signal to decide the data is 1 and 0.
The data can have consequtive 0's and '1's in a row. For instance 0000110001010011111010.
The only way to tell whether in a row of consequtive 0's or 1's there was a 0 or 1 is by usage of the clock.
That clock signal tells you WHEN to look for the signal level is representing a 0 or a 1.
This is when the signal is stabile so not near the rising or falling edges but in the center of the bit.
So you need a clock and data stream so 2 cables.
The clock is easy to retrieve as the signal changes constantly and in a fixed manner.
The data not so.
Technically it is easier to receive a constantly varying signal than one that flips at different times and lengths (data)
Folks wanted 1 cable. Would have been easier to design a 'stereo' cable with 2 different plugs but it was decided to do this in one cable as that could be a standard 50 or 75 Ohm shielded cable and easy to convert to optics and or modulate a single signal instead of 2 (or more as left and R signals and extra data is also in the data).
So technical solution is to 'combine' the clock and data signals in a nifty way.
In the receiver the signal thus changes constantly but differently for a 0 or a 1.
This way it is easy to retreive the clock, by just synchronising a clock on the receiver end on the rising and/or falling edges of a signal.
clock is thus easily 'recovered' using simple electronics.
Also when a clock is recovered and due to the way the 0's and 1's are embedded (encoded) in the combined signal it is easy to retrieve what was a 0 and 1.
Data thus also recovered.
There are different types of en/decoding these types (there are severall ways to do this) of signals SPDIF is one of them.
so no guessing, no algorithms, just simple electronics that can unravel a re-conditioned datastream.
In reality there are quite a few technically different ways to recover and synchronise the clock each with stronger and less strong benefits.
Clocks on the receiving end differ with those on the transmitting end and here too many different technical solutions exist. One better than the other in some areas.
The old Sony dac or the Linn Karik/Numerik used a separate cable for the clock, why? Because the reclocking was not invented at the time? Sony was making some cost no object at the time, they could have found a better solution..
Has been answered above. A decision was made and everyone now has to comply or come up with other formats (HDMI, USB, I2S etc)
Why in the studio they still use a word clock?
This is so all digital processing, DAC's, ADC's all sample at exactly the same time and all data words (so 16 or 24 bits) are generated/read at the exact same time.
When this was not done the drift in each clock of every 'device' would make them all different in frequency (very slightly) and timing.
So the clock here is only needed for this task.
Last edited: