| Originally Posted by Francis_Vaughan AES/EBU is nearly(!) identical in format to S/PDIF, with a different (and curiously inferior) electrical interface. So any chip that delivers S/PDIF could be used. PCM2706 etc. You need an external pulse transformer, or high frequency balanced line driver to convert the single ended S/PDIF to the balanced AES/EBU. Not everything is happy with S/PDIF if it expects AES/EBU. AES/EBU is limited to 20 bits sample depth, and has other format issues that can cause problems. If you do need absolutely perfect AES/EBU format - you will need to use one of the format conversion chips. Crystal do some - try the CS8416. (If you need sample rate conversion CS8420.) They will suck I2S and spit AES/EBU - although you still need the balanced output driver. So a PCM2706 and CS8416 will get you bit perfect AES/EBU. |
![DaKi][er](https://cdn.head-fi.org/avatars/m/6/6700.jpg?1273083357){kind=avatar}
| Originally Posted by Francis_Vaughan The problems with S/PDIF are intrinsic to the data format - in that it carries the clock with the data, and it is impossible to recover the clock without some data correlated jitter artefacts - which yields jitter related distortion components of a peculiarly horrible nature. Jocko Homo's experience seems to suggest that the difficulty in reducing these artefacts is greatly under-appreciated and that jitter in the clock recovery is influenced by an astonishing range of issues - many related to the RF performance of the link. So much so that even very small reflection artefacts due to impedance glitches - even as simple as a slight crushing of the coax cable - is detectable. Now an XLR connector was never designed with the transmission of RF in mind - it has no guaranteed characteristic impedance - and the same is true for the balanced cable. AES/EBU was really designed with the idea that studios could use existing patch cables. The RF performance of this is at best undefined, and almost universally rotten. So why does music recorded under such circumstances survive? Mostly because in a professional setting the clock is not recovered from the AES/EBU signal - it is only used as a digital interconnect - and the audio stays in digital form - with the ADC clock signal derived from a master clock that runs the entire studio. Distribution of that clock is critical - but if it is done with a signal that has no other data on it, one can avoid the worst of the data related jitter issues and recover a pretty clean clock. In a domestic setting, or indeed anywhere where you need to recover the clock, AES/EBU is demonstrably inferior to S/PDIF, and that is starting at a pretty low base. |
| Is tackling the 75ohm problem an all-or-nothing deal? Can I shop and pay for a good cable and forget about what it hooks into? (I am not a Do-it-Myselfer) - as if you couldn't tell. |
