Best Digital Interface
Whether a USB or traditional digital output sounds better are completely dependent on the Streamer / DAC components you have. It’s basically a question of which component is better (simplified answer). I have used both. On my headphone system I have a USB only DAC, so I use that. It is a very good DAC. On my main system I use traditional digital... it sounds ever so slightly better... I mean so little it is silly to talk about. |
The answer is very simple. You have to listen to your system, using various cable connections and decide what sounds best to you. There is no correct answer for everyone. I decided on a music streamer with BOTH USB and AES/EBU digital outputs to my DAC. I listened AND listened to both USB and AES/EBU connections for several weeks. After weeks of switching back and forth, I prefer the AEB/EBU connection from my Aurender music server to my DAC. To my ears, the music sounds slightly better using the AES/EBU Connection. |
Without USB isn't the clock in the Node 2 running the show? Yes, but: 1 - You can't put USB in the middle and have it magically control the output of the coaxial cable. The coaxial signal (or optical) is one way only. 2 - The only reason this was ever needed in the first place was to improve upon mediocre clocks in the transports by replacing them with the superior clocks coming out in the DACs. So, if the transports had perfect clocks, asynchronous USB would never have been needed. These clocks are now better than they used to be. 3 - There are clock regenerators, like Wyred4Sound's Remedy, and it does work and it is effective but only up to a point. It works a lot better with old DACs than new DACs which have better clock locking circuits. |
If your transport is providing S/PDIF, it's never a good idea to convert this to USB. The target DAC will have to re-convert back to S/PDIF and I2S anyways, but now you have timing problems since USB is really a data packet interface and not an audio/timing interface. The RME --might-- be re-clocking the S/PDIF data, but the S/PDIF data from the Node 2 is already clocked using the sample rate of the original music data. |
30.3 Digital Clocks: Internal, SPDIF In Jitter suppression of external clocks: > 50 dB (2.4 kHz) Effective clock jitter influence on DA conversion: near zero PLL ensures zero dropout, even at more than 100 ns jitter Additional Digital Bitclock PLL for trouble-free varispeed ADAT operation Supported sample rates for external clocks: 44 kHz up to 200 kHz Internally supported sample rates: 44.1 kHz up to 768 kHz Maybe I'm a bit dim, but this section of the manual states that this DAC controls jitter of all formats. |
| Post removed |
Is there a Coax to USB adapter available? Does this make sense?The beauty of async USB interface is that by desynchronizing clocks it allows DAC to operate at its own independent stable clock. It does it by signaling back to computer to increase or decrease (upon buffer under or overflow) size of next frame. With adapter there is nothing to signal back to. Computer cannot provide more or less data since coax from computer to adapter is one way only. There is no way to tell computer to stream data faster or slower. Perhaps adapter can have large memory to store extra samples, but it would require big delays. If you start with coax I would stay with it. |
I believe what your original inside outside statement means is that the RME will accept an external clock signal over the SPDIF interface (usually a BNC connector but does not have to be). If using the USB connection, then the RME defaults to its internal clock, which it does with any of its inputs "unless" there is a clock signal present at an SPDIF interface, which is probably set by a menu choice or internal DIP switch? |
I believe what your original inside outside statement means is that the RME will accept an external clock signal over the SPDIF interface The S/PDIF interface requires the data to be clocked from the source. This is the only way it will work and it is part of the interface specification (the digital pulses representing data will need to be transmitted using the timing of the original music data, such as 44.1Khz or 96Khz, etc.) However, there are some DACs that will re-clock those digital pulses in effort to reduce jitter and increase accuracy. Sometimes this provides an improvement, but not always. You may be getting confused with DAC/transport systems that use a "master clock" for S/PDIF. In this scenario, there is a separate device that generated exactly timed "clock pulses" to BOTH the transport and dac. The transport and DAC do their audio encoding/decoding based on the timing of the pulses from the master clock device. This is usually a closed proprietary system which requires the transport, dac and master-clock devices to all work together. This is an option usually on $$$$ level equipment. |
@auxinput The S/PDIF interface requires the data to be clocked from the source. This is the only way it will work and it is part of the interface specificationWhy would a DAC in this day and age rely on a source's data stream which could be flawed rather than use an internal clock? It makes sense to not include an internal clock if a master clock is being used, as in studio setups. These specs indicate that jitter is being suppressed internally, so there is some device reducing any jitter. Clocks: Internal, SPDIF In Jitter suppression of external clocks: > 50 dB (2.4 kHz) Effective clock jitter influence on DA conversion: near zero |
I2S even though my PS Audio DirectStream Sr. (Sunlight) is input Agnostic (according to Ted Smith & my ears). He is correct but my ears find the USB a tad thin (It's Placebo effect IMHO) even he does all that funky USB cleanup inside the DAC. That's all I got and I've been doin' this Digital thang since my first Phillips CD150 with the 14bit/176.4kHz DAC chip. |
Rsjaurr, Yes, I know. That's the point I've been trying to make. The OP posted from the manual... With USB the internal clock is used, with SPDIF the external one".This statement indicates that only USB is being reclocked. But, according to the spec sheet I posted, all signals are being clocked inside the DAC. My earlier post was to say it doesn't make sense to rely on the stream from the source without correcting it. I believe this Dac is designed to reclock all digital streams. |
@mahler123. “I thought the issue was that usb delivers info in packets, and that t.....” . I think you are thinking of ethernet. Network connections are packetized. On the other side of the music server comes either USB or traditional digital streams (SP/DIF, AES...). But yes on USB being a multi-use protocol. |
Both ethernet and USB are data-packet based interfaces. However, the actual communication protocols are different between the two. " USB data is sent in packets Least Significant Bit (LSB) first. There are 4 main USB packet types :Token, Data, Handshake and Start of Frame. Each packet is constructed from different field types, namely SYNC, PID, Address, Data, Endpoint, CRC and EOP. The packets are then bundled into frames to create a USB message." |
Hi mofojo, 1. These two statements in the manual are consistent: With USB the internal clock is used, with SPDIF the external oneand Jitter suppression of external clocks: > 50 dB (2.4 kHz)Jitter suppression of external clocks does not mean the DAC is reclocking the SPDIF signal, just that it’s suppressing some of the jitter. But like your manual and you have said, the Node 2 is running the timing show in this case. I think you’re right to think about reclocking, but I think you might be better served by running a SPDIF reclocker rather than USB reclocker. 2. I will disagree with you about Asynchronous USB being the "best interface for digital transfer to a USB DAC." Sure, Async USB is better than SPDIF with no reclocking. But, for now, I fit into the camp that says USB suffers from more jitter than SPDIF coax. My understanding is that Async USB was designed for the "computer audio" era, when people would run their home computers directly into DAC’s. I believe that the main reason some people continue to use USB at the high end is because of USB’s ability to play large file types like 32bit PCM and DSD. I would suggest that we have now moved from the computer audio era to the streaming era, where most people are running 24/192 and lower only. Although SPDIF coax introduces less jitter than USB, many SPDIF streamers need help. I think many members on this forum think too highly of their Bluesound Nodes and think too highly of their DAC’s capabilities. Why would you want a reclocker for every input in your DAC? That would greatly inflate the cost of the DAC, right? It seems to me like a better practice would be to build the reclocker into the streamer, or to place a separate reclocker between the streamer and the DAC. On his YouTube channel, John Darko states that the Raspberry Pi with SPDIF DigiOne Signature reclocking hat outperforms the Node (~$500) and Volumio Primo (~$900) streamers. To demystify the Primo, it’s just an Asus Tinker Board inside. If you run any off-the-shelf board by itself (Pi or Tinker Board) directly into a DAC, it can only sound so good. These streaming boards are noisy little computers, and they need help. Not all streaming boards are equally noisy, not all master clocks are of equal quality, and not all DACs are of equal quality. More manufacturers are putting out master clocks now. Look into some. Just like you're bypassing the DAC in the Node with an external DAC, think about bypassing the reclocker in your DAC, which is limited to the USB input. It's probably not the best quality anyway. But it's something for people who use that DAC with a computer. I think you’re right, that if you ran a USB from a computer into your Async DAC, that may actually sound better than running SPDIF out from your Node with no reclocking. I would suggest that you try adding a reclocker into your current chain. You may be pleasantly surprised at the improvement. iFi and Wyred4Sound make the most affordable reclockers that I know of. Maybe start there. I use an iFi. On the What’s Best Forum, some members there report benefits running two reclockers in their systems. I haven’t tried that, but it’s an interesting idea. |
@classdstreamer, Good comments, so the DAC is not actually reclocking SPDIF, only suppressing jitter. That’s a pretty vague statement. Like you, I’m using an iFi to reclock the stream from my Node2i and the improvement is audible. I recommend adding a LPS. Even with advancements in jitter reduction, I want the master clock at the destination and not at the source. Thanks for the comprehensive explanation. |
Thank you auxinput for some reality. For audio, USB connections are only required if you are using a computer as a music source. In order to use USB data for music it has to be converted into a data stream and clocked. A DAC with a USB input will do this for you. If you do not have a DAC with a USB input then you will require a USB to SDIF converter. I use the Berkley Alpha USB for this. There are various ways to transfer SPDIF streams, they are RCA by 75 Ohm Coax, AES/EBU by balanced 110 ohm cable, BNC by 75 Ohm Coax and finally optical by fiberoptic cable. There is no sonic difference between these formats. It is not music being transferred, it is clocked data, numbers. You do not have music until after the DAC. As for USB being "flawed" ? That is a ridiculous notion. USB does what it is supposed to do wonderfully well. It can transfer data packets much faster then one requires for music data. It is so fast a computer can transfer an entire 10 minute 24/192 song in about 10 seconds. A 10 minute 16/44.1 song would take less than a second. Digital audio devices do not need to use a USB connection because they are transferring the data already clocked in real time like playing a record. They can music data between them in the digital audio formats mentioned above plus HDMI. There is no Coax to USB adapter because they carry entirely different signals. There are many sub $1000 streamers that have USB connections. They are called Laptop computers. |
@kennyc, I think you're right to separate out "Best Digital Interface" and "Best Digital Performance." Best digital interface is only one consideration, while best digital performance will take into account any strengths in our particular streamer outputs and DAC inputs. This summer, I was planning to built a Pi and run USB to see if I heard a difference with my Async DAC. From here, I would like to include a reclocking hat and run SPDIF to again see if I heard any difference. |
Apples to oranges apt description. Many if not most dac manufacturers optimize usb input, starting to see I2S implemented more often. I2S theoretically better, implementation is the key. I've found maximizing streaming experience is extremely complex. My streaming equipment chain is almost certainly one of one in the entire world. Perhaps the future may bring more certainty to best interface. |
During the 90ties and 00decade, i enjoyed a wadia x32 by spdif with a micromega Duo CD 3.1. Since 2013 my preferred dac is a Mytek DSD 192, using the FW-Output/Oyaide-cable! First, i used a windows-PC/Windows 7 Pro with a FW-PCI-Card. Since 2020, i've got a streamer by apolitan.de and a PCIe-Card/Chipset by NEC (TI no longer avaiable), on going using the mytek. Comparing to a PS Audio DSD, i'm very satisfied by the mytek/no upsampling. Frontend by ATMASPHERE and high efficiency speakers (4way-full horn), you can hear every breathing! Are there any other members with similar experiences? |
@hgeifman -- The answer is very simple. You have to listen to your system, using various cable connections and decide what sounds best to you. There is no correct answer for everyone. +1 I decided on a music streamer with BOTH USB and AES/EBU digital outputs to my DAC. I listened AND listened to both USB and AES/EBU connections for several weeks. For over 13 years my source has been PC-based/HDD; I've used a DIY audio server with AES/EBU output via an RME soundcard, then a USB to S/PDIF converter, then a streamer solution by SOtm, then re-visiting the USB to SPDIF converter (Audiophilleo + PurePower), and now back again essentially to where I started - an AES/EBU output, this time via a Marian Seraph D4 pro soundcard connected to the XLR input of my DAC/preamp with a braided Mundorf Silver/gold digital cable. Even with the DAC clock-slaved by the Marian card it's my preferable sonic scenario. With this AES/EBU connection the music is presented less congealed, more effortless/unforced, and just slightly fuller and better balanced throughout. |
As USB and Ethernet are packet based, you need to clock the timing of packets as well as the digital to analogue conversion. Reclockers only address the USB/Ethernet packet part of clocking accuracy. The actual digital to analogue conversion clocking has major impact on spacial transparency and attack/reverb of the resulting analogue signal. While this has significantly improved in the last 5-10 years, the remaining room for improvement, particularly on Upsampling Delta-Sigma converters cannot be underestimated. This is why USB reclockers will always be only a partial solution. For this part, low phase noise, high accuracy clocks, whether atomic or OCXO have audible benefits. Since most Dacs don‘t have BNC Clock inputs, this may require both additional word- and masterclocks. In an ideal world, having the master clock at the DAC, i.e. just before the conversion to analogue would be best, since buffering, error correction and noise suppression could all happen there. Unfortunately USB and Ethernet are both somewhat compromised because they both transport power and data alongside each other with resulting need for galvanic isolation and noise suppression. In sum, there are good and bad implementations and compatibilities of interfaces and clocking. It follows that there is currently no ‘best’ way of connecting dacs and servers. This is also why the likes of Lumin and Naim are offering integrated DAC/streamer/Server solutions. |
