[sabloke]

From what I know, the best thing is to avoid using USB altogether as it was never developed with high definition sound in mind. Best option is using a proper streamer with I2S or dual AES output.

 

[Lolito]

I know my macbook has optical coming out of the headphone jack, but don't know if that is clean enough. I also have Airplay with an airport express that does red book via wifi, and makes it optical toslink, that goes to the dac... sound darker than usb, and with 3 secs of delay...

Anyhow, for my topping L30 DAC (LOL), what would be best option to clean the USB signal? is there any DDC thread on the forum not to waste space here? any decent galvanizator of those or ddc to make it coaxial, if that would make it better, would be nice. I am ready to waste my money.

 

[]eep]

From what I know, the best thing is to avoid using USB altogether as it was never developed with high definition sound in mind. Best option is using a proper streamer with I2S or dual AES output.

yes and no. Universal serial bus means it is a connection type that can do all sorts of things. It isn't a protocol as much as it is a couple of wires that can transport digital data or analog signal. If you use the proper driver/protocol it just delivers the file verbose in it's own pace. There is no streaming with a clockrate, just: 'heres a chunk of data. Checksum ok? No? Repeat. Yes? Next.' So there is no error correction or redundancy needed. No trouble with bad distorted square wave forms because it just repeats. In streaming you can't resend data because its in real time (well, technically if you use a buffer you can but that's not relevant here). So it isn't the same as FI i²s that is a bunch of wires that each have a seperate stream where the data needs to be uninterrupted (but slower). So what usb does is bring the data as close to the dac and clock as possible, error free. I don't think that's a bad idea.

 

[sabloke]

USB seems to be overly difficult to implement properly. Curious to see how well the new Bluesound Node does it. So far, USB output is only a promise that will be made available via a future firmware update.

 

[sajunky]

Anyhow, for my topping L30 DAC (LOL), what would be best option to clean the USB signal? is there any DDC thread on the forum not to waste space here? any decent galvanizator of those or ddc to make it coaxial, if that would make it better, would be nice. I am ready to waste my money.

.LOL.
If you think USB ground is poluting your DAC inside, you are probably right (most cases). USB reclockers do galvanic isolation well, it is due to the complicated protocol involving DC signalling. A simple and inexpensive self-powered USB 2.0 hub (but not 3.x) can do excellent job redirecting ground loop from PC to the hub's power supply. PSU usually has 2-prong plug, you can switch polarisation and chose one that works better for you. Hub does regenerating packets with its own crystal oscilator, so it also allows to exend cable lenght beyond 5m if your HiFi is in a distant position. If you don't plug more devices just only one, any USB 2.0 hub works well, otherwise you would need a hub with multiple transaction translator (MTT) that cost more and may be false advertised. You can chain few hubs to extent a cable, but the last one must be powered from the same power outlet as a DAC on the short USB cable. This is a mandatory requirement. You can add LPS for the last hub to tune a setup better. Finally make sure that USB host is negotiating with your DAC asynchronous mode, so a clock in your DAC receiver dictates a speed at which new packets arrive. It is a very special case where no reclocking is required. Ares will do reclocking anyway, but it is easy job as clock source is in a DAC receiver. Firmware update can improve reclocking.

For a long cable a better alternative is network streamer with USB port. Few problems: It is difficult to find out whether streamer USB port works in the best asynchronous mode. Secondly, avoid plugging network cable to the home Ehternet switch, it will accumulate ground loops from all devices connected to this hub. Use a dedicated one (or even a pair), the one conneting to your streamer must be plugged to the same power outlet as usual. A better solution is to use a dedicated WiFi extender with a short Ethernet cable to your streamer. A built-in WiFi can pollute inside the streamer, it is why a WiFi extender working in adapter mode (hit!). You can use RPi as a network streamer.

All S/PDIF solutions (except a separate external clock input) require to recover clock from the data stream that carry a jitter, it require PLL for cleaning up. There are inexpensive converters from USB and these work well. Probably it is the best option for the Ares that do not have I2S port. There is typically a galvanic isolator inside a socket, but there are leaky due to the large capacitance against socket's ground. Better devices have larger transformers, chose a quality device like Mutec $1k as advised, but I am sure you can find something less expensive.

 

[]eep]

Hans is one the few people that have NOT sold out. He's very neutral and honest which is welcoming. Thomas and Stereo is another channel that is quite honest too.

Interesting you discuss "clean power". That is a topic I didn't want to get into too much as it's all dependent on the cleanliness of your local power grid, dedicated power line, power conditioning and connection of termination. Also a whole different topic on linear power supplies and manipulating frequency of the sine wave to modify tone........

As far as front end digital, it's rather simple. Eliminate EMI with galvanic isolation and take care of the x and y axis of the square wave signal. Stable amplitude of the wave form as well as taking care of the x axis (precision timing) so it will deliver music as intended by the artist. Also hitting the re-clocked signal with a redundant process to even do more cleanes. I know this works as I've performed daisychain re-clocking and so have others I know.

I'm not agains I2s. My concern is that Ares II DAC owners may get the impression that they are left out due to lack of I2S. Once I2S becomes a standard with identical pinouts I am not too excited about it. I've spent under 6 grand on DDC and audiograde ethernet switches. My DDC does not have I2S. When the time comes I'll dive into I2S when I find a worth DAC. I cannot limit myself in the Denafrips eco system to use proprietary pinouts and specific Denafrips DDC for I2S implementation.

I demand a world standard for the I2s pinout so I can use a re-clocking ddc for any dac with I2S. This moment in time there is no such thing.

75 ohm is not a limitation. I'll take a Tchernov Special coax and be floored in the cost/performance ratio. That cable alone runs lateral if not surpasses extremely expensive coax cables.

I'm a person that looks at specs but I use my ears more. Look at a Berkely audio design alpha 3 dac. Looks wonderful and costs an arm and a leg. Let your ears tell you if a cheaper Ares II is more musical and has a flow of holographic enjoyment.

If you have ever heard of a delta/sigma dac with cascaded re-clocking DDC, please inform us about the performance. I would take a 16/44.1 or 24/192 with re-clocked digital source than 32 bit. I2S has so much more bandwidth. If you buy into a Denafrips DDC and go with a Aqua dac or other I2S dacs the pinouts can be off and the Denafrips DDC is useless, however you can use the AES/toslink/coax iirc.

People are all different. Many high end system users will prefer 24 bit or even 16. If you hear the natural state of analog flow with insane holographic imaging the extra upsample will not give you more. Infact, I am testing Audirvana studio and the analog flow and ease is much more engaging 16 bit than upsampled 24 bit. Try testing Diana Krall's live "a case you".

One thing to consider is that you will not own a DAC forever. Limiting on yourself buying a $$$$ dac or DDC hoping I2S is the best is not versatile. I'd invest in a world standard or even older coax before I2S. That's my ears and experience with this stuff. In due time I'll spend a lot of money on I2S. That will be a no brainer but for now due to no standard.....no thank you

I really appreciate your input and perspective. You already convinced me to get a ddc (probably the Iris because it sort of stacks with the Ares and it can feed my other R2R with usb>coax). So what I'm saying is to build on or nuance, not criticize.

I too appreciate HB for his direct, honest and professional opinions. What bothers me is his presentation and that despite all his expertise he is still firmly stuck in the old reference framework of western made, expensive high end audio. I truly believe that he is honest only honest people can get duped too. Like in the case of MQA where he is still fully backing his old stance on the new hope for high res audio. But thats an entirely different topic.

Also my remark about your focus on the frontend (ddc). Im just saying that you shouldn't let the pendulum completely swing to the other side and forget about the backend (I/V output stage) nor the middle, central part (actual dac). They are all equally important. They form a chain too, just like your entire rig. The chain is as strong as the weakest link.

Your analogy with the x- and y-axis is what I often point out too. Only in the digital domain is is simply less important because the dac only has to decide on a certain tact if its a yes or no. If you can clearly read the text there's no reason anymore for increasing the font size or screen resolution, is there? In the analog domain it's not even an alphabet of letters you need to discriminate but much more values. And on the other axis, the timedomain, smearing has a lot more impact than on the digital tact which is very high and precise by default.

I agree that an actual i²s standard would be nice, but like others already mentioned, we're already halfway there since PS-Audio pretty much set the standard. Once you reach a certain critical mass most outliers will follow suit eventually. And after all, the incompatibility doesn't lie in protocol but merely the strand numbering of the cable.

One other thing that I'm really in agreement on is the (lack of) importance of bitdepth. 16bits is really enough, 20 bits is all you ever need. 24 bits is overkill (unless you want the dynamic range between dropping a needle and a jackhammer directly tied to your skull), 32 bits is completely absurd. Useful for mixing 64 tracks in a studio but not at home. Unless you insist on using digital volume control the wrong way.
96kHz is really helping improve the quality of filtering and keeping the full frequency range. Anything more is overkill. >192kHz is overkill for R2R dacs, usefull for DS (because the chip is stupid and needs lots of filtering).
In 3 words; 16-96 is enough. I've tested it with a panel of friends on my little 4xTDA1543 R2R that I'm still using and actually prefer over the Ares. The Ares has a better frontend, mine has a better backend. If I can bypass the simple dir9001 s/p-dif input chip for i²s and/or feed it with the Iris... That would be hilarious. A $50 dac on a $500 DDC beating other $5000 dacs.

Maybe I won't own this dac forever but the 'giant slayer' Ares (hmm, Ares or Mars was a giant himself...) didn't really win tbe fight with my little David dac. I've got 2 spares and a big brother plus several spare chips plus the ability to build one myself so I reckon I won't run out anytime soon. It's only 24-96 max but I can send i²s to its legs direct. So I agree, better a proper low-ish high-res signal than all-out 32 floating point 1536kHz rubbish. Once that is settled I suggest just enjoying the music or upgrading elsewhere.

I'm now enjoying Maria Callas in Bellini's Norma from 1960 in la Scala de Milano. Such a beautiful recording. So spacious. It's well remastered and I'm playing it in 24-96 But a recording quality and such a performance that most still can't equal these days despite DSD or insane high pcm rates nowadays.

 

[sabloke]

Maria Callas in Bellini's Norma from 1960 in la Scala de Milano

Marvelous recording! Listening sitting at my desk on speakers and if I close my eyes it feels like I'm there in the concert hall. No more DACs, cables and protocols, just music.

 

[alvin1118]

Hey guys,

The latest licensed Thesycon Driver (For Windows PC) v5.12.0 is available to download here: link.
https://www.denafrips.com/support

TUSBAudio - Thesycon USB Audio 2.0 Class Driver for Windows

Revision History
-----------------------------------------------------
V5.12.0 (May 18, 2021)
-----------------------------------------------------
* New: MIDI pipe statistics in the Spy utility
* New: several registry parameters for MIDI added
* Fix: switch preferred ASIO buffer size with driver package containing MIDI only and audio devices
* Chg: MIDI RX now uses USB flow control
* Chg: improved ISO packet error check
* Chg: scripts use Python 3.9.1 now
* Chg: DCK needs Windows 10 now
* Chg: documentation
* Chg: one channel can be part of more than 4 sound devices now

http://www.thesycon.de/usbaudio/TUSBAudio_history.txt

 

[sajunky]

The Ares has a better frontend, mine has a better backend. If I can bypass the simple dir9001 s/p-dif input chip for i²s and/or feed it with the Iris... That would be hilarious. A $50 dac on a $500 DDC beating other $5000 dacs.

Maybe I won't own this dac forever but the 'giant slayer' Ares (hmm, Ares or Mars was a giant himself...) didn't really win tbe fight with my little David dac. I've got 2 spares and a big brother plus several spare chips plus the ability to build one myself so I reckon I won't run out anytime soon. It's only 24-96 max but I can send i²s to its legs direct.

Not directly, you have to convert logic levels from LVDS to TTL (or LVTTL) to get it working. And then you can use any DDC or USB to I2S converter device with HDMI output. Otherwise TDA1543 is fully compliant with I2S protocol in the current form. A single SN65LVDS348 chip can do it. I have a schematic of a board HDMI upgrade for the Audio GD Singularity DACs, you can order a board or make yourself, PM for details.

Correction: Iris have RJ45 connector which use a similar logic levels, but HDMI connection carries less jitter, better make such board.

 

[Lolito]

.LOL.
If you think USB ground is poluting your DAC inside, you are probably right (most cases). USB reclockers do galvanic isolation well, it is due to the complicated protocol involving DC signalling. A simple and inexpensive self-powered USB 2.0 hub (but not 3.x) can do excellent job redirecting ground loop from PC to the hub's power supply. PSU usually has 2-prong plug, you can switch polarisation and chose one that works better for you. Hub does regenerating packets with its own crystal oscilator, so it also allows to exend cable lenght beyond 5m if your HiFi is in a distant position. If you don't plug more devices just only one, any USB 2.0 hub works well, otherwise you would need a hub with multiple transaction translator (MTT) that cost more and may be false advertised. You can chain few hubs to extent a cable, but the last one must be powered from the same power outlet as a DAC on the short USB cable. This is a mandatory requirement. You can add LPS for the last hub to tune a setup better. Finally make sure that USB host is negotiating with your DAC asynchronous mode, so a clock in your DAC receiver dictates a speed at which new packets arrive. It is a very special case where no reclocking is required. Ares will do reclocking anyway, but it is easy job as clock source is in a DAC receiver. Firmware update can improve reclocking.

For a long cable a better alternative is network streamer with USB port. Few problems: It is difficult to find out whether streamer USB port works in the best asynchronous mode. Secondly, avoid plugging network cable to the home Ehternet switch, it will accumulate ground loops from all devices connected to this hub. Use a dedicated one (or even a pair), the one conneting to your streamer must be plugged to the same power outlet as usual. A better solution is to use a dedicated WiFi extender with a short Ethernet cable to your streamer. A built-in WiFi can pollute inside the streamer, it is why a WiFi extender working in adapter mode (hit!). You can use RPi as a network streamer.

All S/PDIF solutions (except a separate external clock input) require to recover clock from the data stream that carry a jitter, it require PLL for cleaning up. There are inexpensive converters from USB and these work well. Probably it is the best option for the Ares that do not have I2S port. There is typically a galvanic isolator inside a socket, but there are leaky due to the large capacitance against socket's ground. Better devices have larger transformers, chose a quality device like Mutec $1k as advised, but I am sure you can find something less expensive.

Even as the nerd that I am, I am not going to make any of that, because I am not that sick yet. I am very bad, but not that much. I appreciate of course the time to write this, I will read it now slowly a few times, I think I can get some easy and fast test to do, with that USB 2.0, but man, that is a lot of talk there. I rather pay for a box that do all that, or maybe connect it directly to laptop and test. I need to liberate a laptop port though, not an easy thing... anyhow, hehe, will buy a proper dac before that, lol, ares 2 convinced me, any for sale??

 

[joseG86]

Hey guys,

The latest licensed Thesycon Driver (For Windows PC) v5.12.0 is available to download here: link.
https://www.denafrips.com/support

TUSBAudio - Thesycon USB Audio 2.0 Class Driver for Windows

Revision History
-----------------------------------------------------
V5.12.0 (May 18, 2021)
-----------------------------------------------------
* New: MIDI pipe statistics in the Spy utility
* New: several registry parameters for MIDI added
* Fix: switch preferred ASIO buffer size with driver package containing MIDI only and audio devices
* Chg: MIDI RX now uses USB flow control
* Chg: improved ISO packet error check
* Chg: scripts use Python 3.9.1 now
* Chg: DCK needs Windows 10 now
* Chg: documentation
* Chg: one channel can be part of more than 4 sound devices now

http://www.thesycon.de/usbaudio/TUSBAudio_history.txt

Thank you so much

 

[sajunky]

Even as the nerd that I am, I am not going to make any of that, because I am not that sick yet. I am very bad, but not that much. I appreciate of course the time to write this, I will read it now slowly a few times, I think I can get some easy and fast test to do, with that USB 2.0, but man, that is a lot of talk there. I rather pay for a box that do all that, or maybe connect it directly to laptop and test.

I give you tips to avoid spending a fortune. Even you don't hear, you need to do some cable management after all.

 

[Lolito]

I really appreciate your input and perspective. You already convinced me to get a ddc (probably the Iris because it sort of stacks with the Ares and it can feed my other R2R with usb>coax). So what I'm saying is to build on or nuance, not criticize.

I too appreciate HB for his direct, honest and professional opinions. What bothers me is his presentation and that despite all his expertise he is still firmly stuck in the old reference framework of western made, expensive high end audio. I truly believe that he is honest only honest people can get duped too. Like in the case of MQA where he is still fully backing his old stance on the new hope for high res audio. But thats an entirely different topic.

Also my remark about your focus on the frontend (ddc). Im just saying that you shouldn't let the pendulum completely swing to the other side and forget about the backend (I/V output stage) nor the middle, central part (actual dac). They are all equally important. They form a chain too, just like your entire rig. The chain is as strong as the weakest link.

Your analogy with the x- and y-axis is what I often point out too. Only in the digital domain is is simply less important because the dac only has to decide on a certain tact if its a yes or no. If you can clearly read the text there's no reason anymore for increasing the font size or screen resolution, is there? In the analog domain it's not even an alphabet of letters you need to discriminate but much more values. And on the other axis, the timedomain, smearing has a lot more impact than on the digital tact which is very high and precise by default.

I agree that an actual i²s standard would be nice, but like others already mentioned, we're already halfway there since PS-Audio pretty much set the standard. Once you reach a certain critical mass most outliers will follow suit eventually. And after all, the incompatibility doesn't lie in protocol but merely the strand numbering of the cable.

One other thing that I'm really in agreement on is the (lack of) importance of bitdepth. 16bits is really enough, 20 bits is all you ever need. 24 bits is overkill (unless you want the dynamic range between dropping a needle and a jackhammer directly tied to your skull), 32 bits is completely absurd. Useful for mixing 64 tracks in a studio but not at home. Unless you insist on using digital volume control the wrong way.
96kHz is really helping improve the quality of filtering and keeping the full frequency range. Anything more is overkill. >192kHz is overkill for R2R dacs, usefull for DS (because the chip is stupid and needs lots of filtering).
In 3 words; 16-96 is enough. I've tested it with a panel of friends on my little 4xTDA1543 R2R that I'm still using and actually prefer over the Ares. The Ares has a better frontend, mine has a better backend. If I can bypass the simple dir9001 s/p-dif input chip for i²s and/or feed it with the Iris... That would be hilarious. A $50 dac on a $500 DDC beating other $5000 dacs.

Maybe I won't own this dac forever but the 'giant slayer' Ares (hmm, Ares or Mars was a giant himself...) didn't really win tbe fight with my little David dac. I've got 2 spares and a big brother plus several spare chips plus the ability to build one myself so I reckon I won't run out anytime soon. It's only 24-96 max but I can send i²s to its legs direct. So I agree, better a proper low-ish high-res signal than all-out 32 floating point 1536kHz rubbish. Once that is settled I suggest just enjoying the music or upgrading elsewhere.

I'm now enjoying Maria Callas in Bellini's Norma from 1960 in la Scala de Milano. Such a beautiful recording. So spacious. It's well remastered and I'm playing it in 24-96 But a recording quality and such a performance that most still can't equal these days despite DSD or insane high pcm rates nowadays.

those sample rate are very useful and in my experience is exactly like that, although 16-96 is a strange combination to get, much more common 24-96. And yes, with delta sigma, chip is stupid indeed, if you feed it extra the DS errors are less apparent in a way... same if you play DSD, with the inconvenience of not able to EQ under unconverted DSD playback. Very often, 24-96 music is not really that much higher resolution, you can hear it is almost cd quality and cd sound (example rage against the machine in high res). In other recordings, 24-96 is so much better (depeche mode remastered). Some cd quality already awesome (like dire straits) and the hd is a bit weird really...

 

[sajunky]

those sample rate are very useful and in my experience is exactly like that, although 16-96 is a strange combination to get, much more common 24-96.

It is a limitation of the chip TDA1543. I own a Nobsound 8xTDA1387, very useful on the road. It has the same limitation. But no, it is not better than Audio GD R2R-11, maybe because of a vintage type oversampling interface chip CM108.

 

[]eep]

I don't consider 24-96 a limitation really (for the TDA1543). I see a lot of little dacs based on the 1543 that are (stupidly) limited to 16-48 by the input chip. But that isn't really why I mentioned the samplerates. I use sd-cards and I want some music on my phone too. So sometimes I downsample stuff that I don't really play that often or care about all that much. There is a lot of music that just wasn't recorded all that well.

Also, when I'm looking for high-res, I come across the 'me too' candidates that are presented in their full 24-48 glory. I am saying: that is useless. 24 bit is useless in real life conditions. If you are wearing headphones and you want to hear the full 16 bits from 0dB to full blast you wil have tinnitus in no time. And 48kHz is NOT high res either. Offering 24-48 as high-res is marketing lies. Upsampling Redbook to 24-96 and selling it as high-res is even worse. But easy to check. Just like MQA.

On the other hand there are really old recordings that are carefully remastered from 35mm tapes etc. that were recorded with just 3 microphones that sound stunning. That I don't mind using some space for the DSD for. And that's why I'm glad to have the Ares II. In the workshop I can output D2P (DSD to PCM) in 96 max. So I totally agree that it's better to have a good signal at lower bitrates than a messy data tsunami.

I am comparing a recording of Rimsky-Korsakov now, in 24-192 it is rather large at 2,36 GB. In 24-96 it is almost exactly half that at 1.19 GB. But now the strange thing... since you would expect most data would come from the 16 msb, in 16-96 the data volume is only just 434 MB. So 2/3 of the bitdepth is only 40% of the size. So if I resample 24-192 to 16-96 I save 2GB of space! That is only 18%, a saving of 82%! And still usefull high res in that it makes filtering a lot easier for the dac.
numbers:
24-192 2416 MB 100%
24-96 1218 MB 50%
24-48 707 MB 29%
16-96 434 MB 18%
16-44 290 MB 12%