Is I2S Connection Better Than S/PDIF - Why ?

Hi,

I am trying to find information explaining benefits of the I2S connection over the S/PDIF. I heard that the I2S is better only in a noisy envirionment.

I only found this one.

from http://www.northstar-audio.co.uk/why_i2s.html

What about the native embedded-clock S/PDIF transmission (no I2S-to-SPDIF conversion) ?

Thanks.

 

Hi bordin,

The main advantage of I2S is the separation of the word clock from the rest of the digital signal. Theoretically, this will allow for lower transmission jitter, same as the separate word clock connection on some pro gear offers.

Of course, this facility must exist at both ends of the connection.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Hi Simon,

That is what theory suggests. I have not made a direct comparison so I don't know. None of my gear uses I2S.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

I2S skips the S/PDIF transmitter and S/PDIF receiver. I2S is actually a format to use internally within digital audio equipment. When equipment is interconnected via I2S, usually line buffers and receivers are specified so that it can drive the cable capacitance and reject noise. However, the drivers and receivers are not perfect, so there is still a small amount of interface jitter, but it's much less than with typical S/PDIF.

 

I use I2S gear in both my main system (transport & DAC) and headphone rig (jitter reduction unit to DAC). The lower jitter of the I2S connection as compared to both AES/EBU and 75 ohm BNC connected S/PDIF has a smoother, glare free sound; I much prefer my gear hooked up with I2S. Kal Rubinson wrote an article years ago in the Audio Amateur on how to add an I2S bus to a CD player. I hope to someday convert a few of my other components. Fortunately, my transport in my big rig was already designed around the I2S connection.

 

Thanks Barry.

Could you explain a bit why it can bring jitter lower.

 

Hi bordin,

It isn't so much that "it can bring lower jitter" as it avoids higher jitter.

When the clock is embedded in the digital audio signal, the likelihood of jitter increases. This can be avoided when the clock signal is carried separately from the audio, as it is with I2S or in some pro gear with separate word clock connections,

For a decent explanation, see:
http://www.anthemav.com/OldSitev1/pdf/i2Srev1.pdf

Hope this helps.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

In properly designed system, I2S is preferable. Dac unit can slave cd transport so dac clock controls transport. This way clock is separate instead of mixed in with audio data and clock and dac and clock chips can be placed very close together. Much less chance for interface jitter.

 

This actually eliminates the effect of interface jitter, since the DAC's clock does not go through the interface. This is another advantage of using separate clock and data lines (as I2S does); it can be implemented so the clock goes one way (DAC to transport) and the data flows the other (transport to DAC). This could be done, regardless of the data format, so long as the DAC has a clock output and transport has a compatible external clock input.

 

Hi Guys,

I have another question. Some people have reported using a high-precision word clock fed into an interface card such as the RME Fireface 400/800 can improve sonic quality of the generated SPDIF output.

I wonder this is valid or can be valid in which conditions.

See >> PC-Source System Sounds As Good As the Esoteric P-05 Disc Transport.

• WinVista PC with RME HDSP 9632 sound card and word clock module. The RME sound card is slaved to the G-03x master clock generator
• foobar 0.9.5.4, output using WASAPI. Active DSPs includes the Secret Rabbit Code Resampler set to 176.4 kHz while playing regular 16/44 lossless files.
• The D-05 is fed by the RME sound card with one AES/EBU cable and slaved to the G-03x clock. The 176.4 feed can be upsampled to DSD.
• This PC-Source system is set against the P-05 disc transport, which is also slaved to the G-03x clock. The P-05 is connected to the D-05 with an Esoteric firewire cable. It can also output dual-AES or a single AES/EBU to the D-05. I've found that the dual-AES output is similar in quality to the firewire link.

It is explained that the external clock is better than the internal crystal of the audio interface.

However, someone have explained that the SPDIF output is totally *independent* from the digital source, i.e. the Firewire words, that are pieced in nature after being received.

This implies that a external word clock does *not* improve quality of sound playback !

Thanks.

 

Hi bordin,

Having the converters inside the computer environment, in my view, makes considerations of where the clock comes from moot. The converters must be removed from the computer first.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Totally agreed. The basic setup would be as follows:

PC's Lossless Song ----- Firewire ----> Fireface 400 -- SPDIF --> DAC
44.1 - 192 kHz Precision Clock -------> Fireface

This looks similar to the separate clock line of the I2S link.

 

To me, it looks as if the jitter introducted by the S/PDIF step would negate any gains made by having a precision clock. For lowest jitter, you want the clock as close to the DAC chip as possible. Suitable buffering (data buffering, not electrical level shifting) and/or ASRC right before the DAC can reduce the effects of interface jitter to inaudible levels.

 

Hi bordin,

I don't believe this is similar. As tps pointed out, you've got an SPDIF link in the chain. The point of a separate word clock is defeated by having integrated clock in the SPDIF link.

In my view, too many links in this chain. Firewire can be a great means of transporting digital audio. Why pollute it with SPDIF?
Again, I agree with tps: the clock should be near the DAC, ideally integrated as it is in the best designs I know.

Personally, I opt for a Firewire interface containing a good DAC and integrated clocking.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

I think this site used to even have a warning next the product saying it wouldn't improve all cards and to ask them if it was even a good idea. Maybe they had too many in-depth questions and had to change the warning.

http://www.blacklionaudio.com/microclock.html

 

Many DACs still take only SPDIF.

This is still my curiosity. A precision clock makes "better" sounds....

As I know, the internal clock operates at MHz while an external wordclock runs at kHz. Making a precise MHz clock may be harder (and cost more) than a low speed one ?

The Black Lion Audio Micro Clock mk2 is among the "players".

 

The DAC chip requires a MHz clock, there's no way around it. If your "reference" is word clock, then the support circuitry for the DAC chip has to use a PLL to generate the MHz "master clock" signal locked to the KHz word clock signal. So there is nothing to be gained by using a KHz reference clock, it just adds more sources of error.

 

Hi bordin,

I mentioned Firewire only because you had it listed in your tentative setup. If you can use Firewire, it does digital audio better than many other means of transporting the bits.

To my ears, better clocking = better sound.
I have heard the affects of an improved clock in a number of systems but in all cases, the changed clock was part of the device.

Best regards,
Barry
www.soundkeeperrecordings.com
www.barrydiamentaudio.com

 

Thanks tps & Barry !

Improving the internal clock would be a better choice.