Baldr
Sponsor: Schiit Audio
- Joined
- May 14, 2011
- Posts
- 943
- Likes
- 7,513
As I sit here listening to the very energetic 13th Floor Elevators I am inspired to complete further thoughts on digital audio connectivity applied to we who build Schiit. (It would probably be more fun to comment on the genus of Psychedelic Rock being Texan, as opposed to San Franciscan, but I will save that argument for later if anyone is interested.)
I saved the least talked about for first. Direct connection – (What is that??) This is the direct connection of digital clocks and data, which has been around since even before S/PDIF, but more recently come to be known as I2S, as if that is the only direct transfer protocol. I2S complicates the transfer protocol with a one bit delay of the word clock; I cannot help but look askance at this delay when considering ways to optimally remove jitter from this most critical of clocks. These have neither electrostatic nor electromagnetic isolation. I suppose this could be the ideal way to do things in an alternative universe where someone bothered to standardize and isolate (with only small distances tolerated between source and DAC), but in the current physical universe, not do-able.
Then we have S/PDIF; it has been around forever. It is an audio interface, designed only for that application. For that reason, it is reliable. Indeed, the most reliable. TOSlink versions degrade the jitter, but add isolation, albeit over short distances between source and DAC. In addition, TOS is only specced up to 96K sample rates – there are, however a few combinations of links/hardware combos which MAY work at 192KHz, however non reproducibly. I believe, however, there is no official support for this 10lbs of data in a 5lb bag. Far better to use coax.
Below we enter into computer based interfaces which were NOT designed for audio. These would be USB and AOIP. In order to use them, we have to deal with ugly layers between the source and the destination, for example ASIO. It is why many makers of software decoders players have an exclusive mode which bypasses most of the very thick rubber in the OS layer. The problem is the reliability with exclusive modes, which is very much a YMMV area. AOIP requires networks, routers (whether built-in or not). What they have in common is that is forces audio makers such as ourselves to support Windoze, Apple, and Android devices which we had no hand in making. What the user needs to remember is that this raises the price of audio devices made not just by us, but for everyone else we compete with.
Now, I hate being in the computer customer support biz, but it is what we must do as long as folks want to keep using interfaces specifically not designed for audio. (Kinda like doing the unspeakable to sheep.) But wait! Surely there must be some silver lining in using these misapplied interfaces. AHA! AOIP gives you two advantages! Isolation and distance. Isolation, true, honest to God isolation, both electrostatic and electromagnetic. There is a thread on this forum and some users who really believe in a certain hardware brand as sounding better than any other interface. Now I believe where there is smoke, there is probably fire.
Distance! This means you can setup your source machine in your garage, your destination machines in your bedroom, living room, and run between them all to stay in shape! Well, not. Pro uses, yes. Home automation installs in 36,000 sq. ft. homes, yes. That's just not our market. We build for the ROW (rest of the word). This not to say I am an AES67 denier. I know there applications for audio over internet (home automation) and some of our clients are willing to pay for it. It is just I am not willing to make (and worse yet) support it.
Now I know that the clients who believe in the exalted sound of the better AOIP systems believe there is something truly special, indeed magical about the sound. Give the reliance on the layers in the OSes between the file and the network output, I wondered.
So my goal was to build a universal product of isolation for a non distributed audio system that is truly as isolated as AOIP. If all those who have heard them say they are sonically competitive in the context of all of the above, look out.
I saved the least talked about for first. Direct connection – (What is that??) This is the direct connection of digital clocks and data, which has been around since even before S/PDIF, but more recently come to be known as I2S, as if that is the only direct transfer protocol. I2S complicates the transfer protocol with a one bit delay of the word clock; I cannot help but look askance at this delay when considering ways to optimally remove jitter from this most critical of clocks. These have neither electrostatic nor electromagnetic isolation. I suppose this could be the ideal way to do things in an alternative universe where someone bothered to standardize and isolate (with only small distances tolerated between source and DAC), but in the current physical universe, not do-able.
Then we have S/PDIF; it has been around forever. It is an audio interface, designed only for that application. For that reason, it is reliable. Indeed, the most reliable. TOSlink versions degrade the jitter, but add isolation, albeit over short distances between source and DAC. In addition, TOS is only specced up to 96K sample rates – there are, however a few combinations of links/hardware combos which MAY work at 192KHz, however non reproducibly. I believe, however, there is no official support for this 10lbs of data in a 5lb bag. Far better to use coax.
Below we enter into computer based interfaces which were NOT designed for audio. These would be USB and AOIP. In order to use them, we have to deal with ugly layers between the source and the destination, for example ASIO. It is why many makers of software decoders players have an exclusive mode which bypasses most of the very thick rubber in the OS layer. The problem is the reliability with exclusive modes, which is very much a YMMV area. AOIP requires networks, routers (whether built-in or not). What they have in common is that is forces audio makers such as ourselves to support Windoze, Apple, and Android devices which we had no hand in making. What the user needs to remember is that this raises the price of audio devices made not just by us, but for everyone else we compete with.
Now, I hate being in the computer customer support biz, but it is what we must do as long as folks want to keep using interfaces specifically not designed for audio. (Kinda like doing the unspeakable to sheep.) But wait! Surely there must be some silver lining in using these misapplied interfaces. AHA! AOIP gives you two advantages! Isolation and distance. Isolation, true, honest to God isolation, both electrostatic and electromagnetic. There is a thread on this forum and some users who really believe in a certain hardware brand as sounding better than any other interface. Now I believe where there is smoke, there is probably fire.
Distance! This means you can setup your source machine in your garage, your destination machines in your bedroom, living room, and run between them all to stay in shape! Well, not. Pro uses, yes. Home automation installs in 36,000 sq. ft. homes, yes. That's just not our market. We build for the ROW (rest of the word). This not to say I am an AES67 denier. I know there applications for audio over internet (home automation) and some of our clients are willing to pay for it. It is just I am not willing to make (and worse yet) support it.
Now I know that the clients who believe in the exalted sound of the better AOIP systems believe there is something truly special, indeed magical about the sound. Give the reliance on the layers in the OSes between the file and the network output, I wondered.
So my goal was to build a universal product of isolation for a non distributed audio system that is truly as isolated as AOIP. If all those who have heard them say they are sonically competitive in the context of all of the above, look out.
Stay updated on Schiit Audio at their sponsor profile on Head-Fi.
|