TOSLink was only designed for 96k and that’s all the most common TOSLink transmitters and receivers are speced for. On the other hand when you exceed the speced bandwidth you may just get a sloppier signal: for example, one that’s got more jitter and/or less amplitude. With a TOSLink cable that doesn’t compound these issues too much it may still work just fine. I’ve seen some that work at 176.4k but not 192k. Also I’ve seen some sort of work on faster signals but introduce enough errors to sound “fuzzy”. If there’s any suspicion of errors you might run the bit perfect test for a while and see if there are any problems, but staying in DoP mode almost surely means there are few (if any) errors at 176.4k.
It is this simple and don’t let audio makers scare you with false marketing information just because they want you to believe in what they can do best (But I can understand for them: this is a small and crowded market with more makers and less consumers so a simple thing like a femto grade clock of 20-30$ has to be put in marketing information like a rocket science to make people believe). I am not against anybody; i buy stuff from them, but I buy from those who offer more truths on the products.
Unless you don’t understand Ethernet then you can conclude Ethernet is no good for music. In NAS case, Ethernet transfers your data from NAS to music server, just DATA, no timing information. At music server, it is the part (i.e USB board/chip set) which clocks out your data to DAC starts to add timing problems to the music. It is all about the design of this music server is to blame (Some music servers use normal PC board with Intel CPU and using the computer clock to generate clock for USB via software and software is written badly, not good board design to eliminate cross-talk, no clean power supply, no separate power supply for that part, software on music server is not well written to add delay, they cannot maintain good team of software and hardware testers because unlike Apple/Microsoft/Texas Instruments, etc. selling products to million people, not enough prototypes before mass production due to cost cutting, you name other challenges). So instead of blaming NAS, do home work to choose the right source (music server/ streamer). Most makers shift the problems to other gears which they don’t make. For instance, I bought a well-known DAC with information in a company’s website that it rejects jitters very well etc. and good reviews on 6moons (6moons is the marketing department of makers anyway). But in reality, it cannot reject anything, it even cannot work with quite a decent music streamer. I called the key designer and co-owner of that company and he said that I need a good source with little jitters to work with that DAC. If that’s the case, don’t massage the information on the jitter part of the product in their website in the first place. I consider that as providing consumers with false information. Of course I had my money back because they were totally wrong.
If Ethernet signal on cable is noisy a bit, isolate it from power supply part for the USB board on streamer/music server.
If NAS creates noise to power line of your house, everything else also creates noise like aircon, etc. PS Audio claims P10 is good, use P10. I am using a P10.
So the fact to blame NAS is just wrong. If NAS is bad because of the spinning devices (HDD), go with a server of all SSD instead of HDD and still feed your streamer via Ethernet. Your music is just using very small bandwidth compared to high bandwidth of Ethernet .My point in the critical path is in designing this USB or whatever interface which sends out your data to DAC.
Many exaggerate things to sell and many listen to them blindly without critical thinking from engineering points. Hearing/listening to music to select products is good but hearing/listening alone to judge without analyzing things would cost lots money in trials and errors.
@wglenn you wrote “I believe that we, as the collective computer audio community, will be the ones to figure out how to get digital equipment to do what we want, not the hardware producers themselves.” —> true. But if you can look into how their products are designed and how they solve the challenging matters then you add one more step in decision making than just listening alone. We learn and go but end of the day, we educate ourselves; truthful makers can help us in the learning curve; most cannot educate us anything accept for giving half-truth information or false information. Most audiophiles boil everything down to “it sounds good” and equip themselves with all wrong knowledge fed by marketing information–> not enough to be audiophiles.
Ted Smith said TOSLink was only designed for 96k and that's all the most common TOSLink transmitters and receivers are speced for. On the other hand when you exceed the speced bandwidth you may just get a sloppier signal: for example, one that's got more jitter and/or less amplitude. With a TOSLink cable that doesn't compound these issues too much it may still work just fine. I've seen some that work at 176.4k but not 192k. Also I've seen some sort of work on faster signals but introduce enough errors to sound "fuzzy". If there's any suspicion of errors you might run the bit perfect test for a while and see if there are any problems, but staying in DoP mode almost surely means there are few (if any) errors at 176.4k.Yes I have experienced that...at 192 it actually stops playin for a split second..I thought it was my Internet connection even though it's hard wired..I thought maybe it was a switching issue with my router. ..I'm going to try the coax and see if it's better...is coax better than toslink?
Coax will work at 192k, with most systems is has less jitter than TOSLink which can make it sound better, on the other hand it can create a ground loop which is worse. All in all coax will be more reliable and probably be better sounding than TOSLink, but not for everybody.
Coax works fine for me, and it’s the only connection I use on my DS.
I’m currently using PC over usb and notice the difference between Qobuz streaming (warmer sound) and cd/Foobar (more balanced). Both should be bit perfect – so the differences lie in the noise carried in the system.
Perfect signal conditioning should make all bit perfect sources sound the same. It would be good to condition the bit perfect feed to the Directstream so as to eliminate colouration of sources. Views?
The PC is anachronistic, generating bit perfect audio, but smothering it in software and hardware generated noise. Cleaning up the PC, while helpful, is ultimately doomed – there are simply too many noise sources. A better strategy is to leave the PC noise behind and clean up the feed to the Directsream.
I got surprisingly good results cleaning up usb by interposing a usb hub (thanks @Ted) with its own 5V supply and modifying cheap cables by cutting the usb 5V from the PC (you do need to pull up the usb cable 5V to a different 5V supply – Yale Final needs this).
The next step, not so cheap, is to galvanically isolate the PC usb, re-clock the signal, and convert usb to I2S, eliminating a couple of stages of usb processing and going straight to the FPGA. The only device I know that does this is the Audiobyte Hydra Z, for about $850.
So, questions. Given the approach above, will switching to the Hydra Z make an appreciable difference? On the one hand it should – isolate galvanically from the PC noise, re-clock to eliminate usb timing differences, I2S feed to the Directstream. On the other hand, if the Directstream’s internal buffering helps isolate the effects of noise on inputs, any effect may be marginal.
Comments and experiences welcome.
Lonan
(Note - I edited this to clarify the core points )
lonan said I’m currently using a PC source over usb and notice the difference between Qobuz streaming (warmer sound) and cd/Foobar (more balanced). Both should be bit perfect –They should be, but have you run the bit perfect test to make certain this is true?
Ta. When I run the bitperfect test on Foobar it’s fine. How to test bit perfect on Qobuz is a problem? I think PS Audio would have to log their Bitperfect Test track with Qobuz to allow it to be downloaded and played through in the normal way. Don’t know another wayd.
Lonan
Ted Smith said Most NASs are build to be fast, e.g. they use multiple fast drives. Fast drives make audible noise and to go fast they need big bursts of current. Those bursts of current are noise that the NAS puts on the power lines and noise that shows up on the output wires. Also most (but not all) NASs are optimized to be reliable and fast not to be electrically quiet, i.e. they are often built for business or industrial environments. Anyway that noise does cause jitter (which will get greatly attenuated both in the bridge and then again in the DS.) But it also can be radiated by the cables, etc. Different systems will react differently to these sources of noise. My system (which is fairly immune to many sources of noise) gets a hard edge to the sound when connected to the NAS by a network cable. I'm sure that if I used a longer cable, had more switches between the NAS and my system and/or had the NAS more remotely located the noise would lessen. But since I had other reasons to put the NAS on wireless I chose that as a way of isolating it. Don't get me wrong, wireless is often a network reliability problem and obviously it's an RF source, but 1 gig wireless doesn't seem to be a reliability problem and like everything else it's a compromise, but a compromise that works for me at this point in time.I’m not trying to scare anyone away from NASs, it’s just that virtually anything has some downsides and being nice to a music system was the last thing on the minds of the designers for most NASs. (Tho that’s changing.) For many systems a NAS is can be better than a bunch of USB drives or a bunch of drives in the music computer proper.
Ted, is this risk of noice by a NAS over Network cable to the bridge also valid in case the NAS uses a different AC network than the DAC?
I just try to sum up my insight from Ted’s useful info:
Toslink can do from 96-176k, has higher niose and lower ground loop risk
Coax can do up to 192k or 1xDSD, has lower noise and higher groud loop risk
USB can do any samplerate, has higher noise and higher groud loop risk
I2S can do up to 352k or 2xDSD, has lower noise and lower ground loop risk, but designed for short lenghts
I also got, that in general USB and Toslink are considered as the worst options (except for USB’s ability to do all sampling rates) while Coax and I2S can be the best solution for limited sampling rates depending on the rest of the equipment.
And I got, that going by LAN or WLAN into the Bridge has less risks of noise or ground loop than using any direct input into the DAC by any other option (USB. i2s, Tos, Coax) and is probably the best sounding solution, independent of possible noise from a NAS, too.
Correct?
lonan said How to test bit perfect on Qobuz is a problem?Yes, I forgot it is a streaming service (I do not use streaming of any kind).
jazznut said I just try to sum up my insight from Ted's useful info:Just because I'm a pedantic kind of guy and abhor oversimplifications:Toslink can do from 96-176k, has higher niose and lower ground loop risk
Coax can do up to 192k or 1xDSD, has lower noise and higher groud loop risk
USB can do any samplerate, has higher noise and higher groud loop risk
I2S can do up to 352k or 2xDSD, has lower noise and lower ground loop risk, but designed for short lenghtsI also got, that in general USB and Toslink are considered as the worst options (except for USB’s ability to do all sampling rates) while Coax and I2S can be the best solution for limited sampling rates depending on the rest of the equipment.
And I got, that going by LAN or WLAN into the Bridge has less risks of noise or ground loop than using any direct input into the DAC by any other option (USB. i2s, Tos, Coax) and is probably the best sounding solution, independent of possible noise from a NAS, too.
Correct?
The term “Noise” isn’t being used consistently in this synopsis - for example TOSLink typically has the lowest electrical noise but somewhat higher jitter than AES/EBU, S/PDIF and I2S.
Because jitter is less of a problem for the DS the common knowledge about these formats isn’t entirely true for the DS, in particular I use TOSLink instead of S/PDIF or AES/EBU when 24/192 TOSLink is reliable in my system. (In fact I did all of my demos of my prototype hardware with a cobbled PC → USB → S/PDIF → TOSLink → DAC signal chain.)
Also with the DS if all other things are equal (e.g. from the same source/hardware to the DS) I2S, AES/EBU and S/PDIF have very similar performance - If a reliable TOSLink connection isn’t available and all other things are equal I would use I2S for short connections because it typically has robust shielding and good grounding, for longer connections I’d use AES/EBU for the better common mode rejection and wider compatibility - but any of them can work well.
The Bridge isn’t inherently a lower noise solution than others (and certainly not lower noise than TOSLink). Like all of this the Bridge’s benefits are quite system dependent but for many people the bridge provides excellent sound quality. With a NAS in the same room as the rest of my system the Bridge was the noisiest (in more than one way) and worst sounding connection - but with the NAS in another room and plugged into a remote outlet the Bridge sounded much better.
So, just to beat a dead horse, all of these things are system specific and you can get great sound quality with any of them - personally I choose the one that best matches my lifestyle / system layout / user interface / available sources concerns and then tweak it to work as best I can up to my frustration and budget thresholds.
Ted Smith said
personally I choose the one that best matches my lifestyle / system layout / user interface / available sources concerns and then tweak it to work as best I can up to my frustration and budget thresholds.Exactly, that makes sense! In my case I currently don't want to have a PC or Mac aside music listening but will use the bridge and I try to have optimized sound with a separate power line for the stereo equipment than for the rest of the house-electricity incl. NAS, which is located in the basement aside the router.
Like all of this the Bridge's benefits are quite system dependent but for many people the bridge provides excellent sound quality. With a NAS in the same room as the rest of my system the Bridge was the noisiest (in more than one way) and worst sounding connection - but with the NAS in another room and plugged into a remote outlet the Bridge sounded much better.Assuming equal system setup with isolated NAS or non isolated NAS and just related to electrical not acoustical noise:
What is the theory why a NAS should cause more noise and produce worse sound when connected to the bridge than when connected to a PC which is connected to the DAC by cable or to the bridge over network?
Allthought we will certainly all make our own experiences, I wonder that there are no general rules or transferable experiences for best sound quality related to the few implementation options between sound files and DAC. All this seems to be much more a puzzle than CD or vinyl playback ever was
But the sound of the DS is worth the effort for sure!
jazznut said What is the theory why a NAS should cause more noise and produce worse sound when connected to the bridge than when connected to a PC which is connected to the DAC by cable or to the bridge over network?Allthought we will certainly all make our own experiences, I wonder that there are no general rules or transferable experiences for best sound quality related to the few implementation options between sound files and DAC. All this seems to be much more a puzzle than CD or vinyl playback ever was
But the sound of the DS is worth the effort for sure!
I think getting the best sound out of a system is always “hard”. I don’t think there are general rules for picking cables (of any kind), for choosing speakers, etc.
But yep, having a computer in one’s audio system magnifies such problems more than most think.
I didn’t mean to imply that a bridge/network is inherently worse than other methods, I just don’t agree with some that it has any strong inherent advantage. I do think it works better out of the chute for many people as far as sound quality, but there can be a host of issues that are harder to get under control when things aren’t working/sounding great.
In my case using USB from my PC to the DAC and a wireless network from my PC to the NAS in another room sounds better than having the NAS in the room with the audio system no matter how it’s connected to the DAC, bridge or not. Further my network system is too broken/weird for the Bridge to work reliably in other rooms, but it suits my purposes in all other ways so I just snake an Ethernet cable down the hall when I have to debug something related to the Bridge.
Ted … Are there any SQ reasons you use wireless instead of Ethernet from your NAS?
No, in general wireless is more hostile to an audio system than a wired connection. But the convenience of wireless connections to the router and the NAS for all of the laptops in the house wins. I can’t believe that a wireless connection can sustain more than a Gigabit/second from the NAS, but it does.
I use mostly USB to the DAC. I’ve never liked the UPnP player experience so I’ve never pushed thru cleaning up my network enough to reliably support the Bridge.