Tweaker's Rash

[sima66]

Reading this thread is similar to reading a Google translation!

Last time I was this confused is when I read a Google translation from a German audio forum.
It says that "using toilet is a much better idea" and "make sure that you use 75-ohm BNC cable for the toilet"!
After I realized that they used the short for World Clock, which is WC, and of course, WC = toilet!

As far as I can see you guys are not gonna use any toilets in the future! ;-)

[darkpink]

It sucks being late to the party, just invested in a new pricier linear powered audiopc. And then your read this thread! Back to square one -ahh the things I do for best possible audio :)

[jkeny]

Reading this thread is similar to reading a Google translation!

Last time I was this confused is when I read a Google translation from a German audio forum.
It says that "using toilet is a much better idea" and "make sure that you use 75-ohm BNC cable for the toilet"!
After I realized that they used the short for World Clock, which is WC, and of course, WC = toilet!

As far as I can see you guys are not gonna use any toilets in the future! ;-)

Sorry, sima, yes I'll be all bunged up well into the future :)

I2S is an audio communication standard that is usually used between digital chips - it literally stands for IIS or Integrated Interchip Sound
Much like SPDIF is also another standard for audio communication between devices but over longer distances

Every communication pathway has to have an agreed protocol in order to effectively communicate.
In the official protocol for I2S there are 3 signal lines involved
Bitclock = is the timing for each bit in the signal
Word clock = an indicator if this is the left channel bits or right channel bits (remember the signal is stereo)
Data = the actual data bits

There is usually a master clock (MCLK) from which these signals generally are referenced to or synchronised to.
It's required to be provided by lots of DAC chips & therefore has to be dealt with for these chips.

Maybe this helps, maybe not - a bit of wine has lubricated this post - maybe not a good idea to try to post technical explanations?

[mjock3]

Thanks John for clarifying these terms. It sure is nice to read posts like this so that I have some understanding.

I have been interested in checking out Linux for a year or so. So I am pretty rev'ed to try this out. Of course following the footsteps of those who know what they are doing. :-)

[sima66]

Thanks John, that was very nice put and explained, even for the guys like me with no background in digital audio.
It definitely helps.

[jkeny]

Glad it made sense
So what we are trying to handle is the I2S signal coming from these device RPi or BBB.
The RPi doesn't output it's internal MCLK timing signal (even though there must be one used internally) & the BBB does output MCLK

This means that the RPi is limited to connecting only with DAC chips that don't need a MCLK signal input (some DAC chips don't - ESS ones & PCM5102 are two of the ones being mentioned here).
Whereas the BBB seems to have an MCLK signal exposed & it can be used by more DAC chips - the ones that need MCLK input.

However, exact timing in digital audio is a critical & any inexactness in this timing is called jitter & lead to less than optimal sound. This is a hotly debated topic on forums - just how much timing inexactness results in audible degradation of the sound.

My take on it is that timing is important but only when all the other factors that degrade the sound are dealt with. It may be that PC/PS/USB introduce other factors which degrade the sound & this is what we may find out from these tests/experiments. Mark is already hearing a far better sound by taking I2S directly to his Buffalo DAC (ESS DAC) rather than through USB to his WaveIO (USB to I2S converter) & we are heading down that road ourselves.

Interesting times ahead, I reckon!

[nige2000]

if we wanted or felt it was necessary to add a separate master clock is it necessary/how do we sync the i2s to it

[sima66]

Glad it made sense
So what we are trying to handle is the I2S signal coming from these device RPi or BBB.
The RPi doesn't output it's internal MCLK timing signal (even though there must be one used internally) & the BBB does output MCLK

This means that the RPi is limited to connecting only with DAC chips that don't need a MCLK signal input (some DAC chips don't - ESS ones & PCM5102 are two of the ones being mentioned here).
Whereas the BBB seems to have an MCLK signal exposed & it can be used by more DAC chips - the ones that need MCLK input.

However, exact timing in digital audio is a critical & any inexactness in this timing is called jitter & lead to less than optimal sound. This is a hotly debated topic on forums - just how much timing inexactness results in audible degradation of the sound.

My take on it is that timing is important but only when all the other factors that degrade the sound are dealt with. It may be that PC/PS/USB introduce other factors which degrade the sound & this is what we may find out from these tests/experiments. Mark is already hearing a far better sound by taking I2S directly to his Buffalo DAC (ESS DAC) rather than through USB to his WaveIO (USB to I2S converter) & we are heading down that road ourselves.

Interesting times ahead, I reckon!

Looks like we have a New Era ahead of as. The idea is looking very promising.

BTW, John I like when you "lubricate" your thoughts with wine! Most people talk more gibberish after wine, but in your case it's the opposite......you've never been clearer! ;-)

[jkeny]

if we wanted or felt it was necessary to add a separate master clock is it necessary/how do we sync the i2s to it

This is the problem, Nige - it is necessary to synch the master clock with the rest of the I2S signal timings. The ESS DACs can handle an MCLK that is not synchronised with the rest of the I2S signals because it has an inbuilt ASRC (asynchronous sample rate converter). There are ASRC chips that can be used in front of DAC chips that don't have them in-built. In general, ASRCs tend to be somewhat detrimental to the sound, for various reasons. The ESS DAC's ASRC is one of the best ones but even so ESS DACs sound slightly better when not using it's internal ASRC.

So an optimal solution is something to consider down the road & not easy to explain/achieve - let's see how this solution sounds first!

[jkeny]

Looks like we have a New Era ahead of as. The idea is looking very promising.

BTW, John I like when you "lubricate" your thoughts with wine! Most people talk more gibberish after wine, but in your case it's the opposite......you've never been clearer! ;-)

LOL!