What's the definitive story on Toslink 24/192?

[wappinghigh]

Q says it all really: and has arisen in searching for suitable DAC's which will do 24/192 via Toslink. (To work with the Bluesound Node)

 

Benchmark says they no longer do 24/192 since Toshiba stopped making 24/192 Toslink.

 

Yet there are a few so called 24/192 Toslink DACS out there.. Including new ones being released.

 

So what's the story ? Are they using different Non Toshiba Toslink connectors or is some other Toslink maker now doing 24/192?

 

Thanks!

[Miska]

So what's the story ? Are they using different Non Toshiba Toslink connectors or is some other Toslink maker now doing 24/192?

 

I would guess most are using non-Toshiba parts. There are many many companies producing Toslink transmitters and receivers.

 

My Mytek DAC works perfectly fine at 176.4 and 192 rates with Toslink from ASUS Xonar DX card... (on Linux, never tried it on Windows)

 

Neither one officially supports 176.4k, but that was the first route I played DoP through...

[fmak]

Q says it all really: and has arisen in searching for suitable DAC's which will do 24/192 via Toslink. (To work with the Bluesound Node)

 

Benchmark says they no longer do 24/192 since Toshiba stopped making 24/192 Toslink.

 

Yet there are a few so called 24/192 Toslink DACS out there.. Including new ones being released.

 

So what's the story ? Are they using different Non Toshiba Toslink connectors or is some other Toslink maker now doing 24/192?

 

Thanks!

 

It may work on the few, but generally the thing is lost cause.

 

If you montor waveform integrity at even 96k with Toshlink, it's absysmal.

[wappinghigh]

Sure. But can anybody answer my questions? Why did Toshiba ditch it? And who now makes them? Any why do Benchmark say what they do on their website?

[Miska]

It may work on the few, but generally the thing is lost cause.

 

If you montor waveform integrity at even 96k with Toshlink, it's absysmal.

 

Why do you think so? The cable and connector is not so bad (compared to ones used for Ethernet) and rest is about implementation of the module. The Toshiba module works up to 125 Mbps while signalling rate for 192 kHz is just 24.576 MHz which is pathetically slow compared to any optical cable engineering.

 

Waveform is after a Schmitt-trigger doesn't look bad.

 

Btw, works also perfectly fine from the ASUS Xonar DX to Fostex HP-A8C DAC/headphone amp.

 

[Miska]

Sure. But can anybody answer my questions? Why did Toshiba ditch it? And who now makes them? Any why do Benchmark say what they do on their website?

 

There are too many manufacturers to list. But I've been using modules from Sharp.

 

GP1FAV53RK0F | Plastic Optical Fiber | Sharp Microelectronics Europe

GP1FAV55TK0F | Plastic Optical Fiber | Sharp Microelectronics Europe

GP1FM55HTZ0F | Plastic Optical Fiber | Sharp Microelectronics Europe

 

Toslink was originally designed by Toshiba, but when it became generic standard as optical audio connector the modules are now manufactured probably by at least ten different companies. And the most common connector these days, the "mini toslink" which is combination of 3.5 mm electrical audio plug/jack and optical connector has nothing to do with Toshiba. These are used on Mac's and for example on the Xonar DX sound card (combined mic input / optical output).

 

So these days, there's no relation between "Toslink" and Toshiba.

[Sam Lord]

The low bandwidth kills it for future computing uses, and the connectors have poor protection. It's very easy to scratch the little buggers.

[Miska]

The low bandwidth kills it for future computing uses, and the connectors have poor protection. It's very easy to scratch the little buggers.

 

The other commonly used ST-connector is not any better in terms of protection. One just has to be careful to always use protective caps whenever the connector is not connected...

 

The dual LC- and SC- connectors used for 1G and 10G ethernet are similar plastic affairs and don't offer any better protection either.

 

 

Bandwidth is primarily determined by the electronics inside the module which is up to the manufacturer to decide...

[Sam Lord]

The other commonly used ST-connector is not any better in terms of protection. One just has to be careful to always use protective caps whenever the connector is not connected.... The dual LC- and SC- connectors used for 1G and 10G ethernet are similar plastic affairs and don't offer any better protection either.

 

Indeed ST is even worse, very fragile and exposed, you don't dare insert them blindly. I used an ST-terminated cable years ago on a Krell transport. I've ruined a couple of Toslink cables that way. I didn't realize SC were as bad, they looked more protected from other photos.

 

Bandwidth is primarily determined by the electronics inside the module which is up to the manufacturer to decide...

 

Yes, but gee the thought of any modern connection choking on 10Mbps feels pretty silly.

[Miska]

Yes, but gee the thought of any modern connection choking on 10Mbps feels pretty silly.

 

Sure it is, when that happens. There are bunch of Toslink optical modules capped at 16.6 Mbps which is enough for 96 kHz audio, but not for 176.4/192.

[fmak]

Why do you think so? The cable and connector is not so bad (compared to ones used for Ethernet) and rest is about implementation of the module. The Toshiba module works up to 125 Mbps while signalling rate for 192 kHz is just 24.576 MHz which is pathetically slow compared to any optical cable engineering.

 

Waveform is after a Schmitt-trigger doesn't look bad.

 

Btw, works also perfectly fine from the ASUS Xonar DX to Fostex HP-A8C DAC/headphone amp.

 

[ATTACH=CONFIG]15257[/ATTACH]

 

Why, all the ones I tried, including a test module from Twisted Pear don't work at 176.4k.

 

More reasons:

 

Toshlink to me sounds bad compared to electrical. I have only found one cable that sounds good and that was from Japan. All this glass fibre stuff is just so so.

 

Waveform - convert direct from Toshlink to electrical without relock and you don't even get anything that looks 'square'.

 

Modules - many of them don't fit well. The termination on the cables is often somewhat conical.

 

This site is still playing up. I logged in earlier but couldn't post. I shut it down and sometime later found that I was still logged in and could post!

[Miska]

Why, all the ones I tried, including a test module from Twisted Pear don't work at 176.4k.

 

Because you have slow transmitter or receiver module, one of those 16 Mbps things, on the way. Or some design mistake.

 

Toshlink to me sounds bad compared to electrical. I have only found one cable that sounds good and that was from Japan. All this glass fibre stuff is just so so.

 

The 5€ plastic fiber sounds (and measures) the same as coax S/PDIF if the implementation is not bad. I have never tried any "audiophile digital cables". My coax S/PDIF cables are standard RG-59 (MIL-STD compliant one).

 

I have one Supra USB-cable, but I haven't got interest yet to try it out. I prefer to stick to cables with official USB certification.

 

Waveform - convert direct from Toshlink to electrical without relock and you don't even get anything that looks 'square'.

 

It looks completely square after a Schmitt-trigger. And the looks doesn't matter as long as the bits are translated correctly.

 

Same as with a plain clock crystal without Schmitt-trigger, it looks something completely different than a square. Nothing special about that.

 

In many cases, the electrical isolation you get from optical cable is much more important than how the "waveform looks like". There are way too many cases where coaxial S/PDIF is implemented even without proper (mandatory) transformer isolation!

[fmak]

Sure it is, when that happens. There are bunch of Toslink optical modules capped at 16.6 Mbps which is enough for 96 kHz audio, but not for 176.4/192.

 

that posters who do not agree with you are simplistic and make statements based on one set of

experience and not that gained over many years with different high end systems.

 

I can tell you that even though RME claims 192k operation in Toshlink, they don't, and not just with one receiver.

 

Frankly, if you can't hear differences between Toshlink and Electrical, then you haven't got a very good system.

 

The Asus you have is a mid price box and I would have thought that you would have used something much better.

 

I have to post in response to an earlier post of yours because, for some reason, you current posts are preventing

responses by defaulting to 'not signed in'. This doesn't seem to apply to posts from others.

[Miska]

Frankly, if you can't hear differences between Toshlink and Electrical, then you haven't got a very good system.

 

I don't think so, although for me it is enough to hear all the minor details between different digital filters. At least I have good objective qualifications of my system performance.

 

I think a good system performs well even with crappiest cables you can find. Poor system is very sensitive to cables or types of data transfer.

 

The Asus you have is a mid price box and I would have thought that you would have used something much better.

 

What ASUS box? I don't have any ASUS box. I have quite extensive set of DACs and sound cards as well as measurement equipment.

 

For the optical input I tested with Mytek Stereo192-DSD DAC and Fostex HP-A8C. Optical cable is normal 5€ cheap plastic mini-Toslink (3.5mm plug) to Toslink thing, has been working perfectly for past five years. I've heard of people having lot of problems when they switched to expensive "audiophile" optical cables where cheap ordinary bulk cable was working fine. With audiophile gear, it is not unusual for quality to be inversely proportional to price.

 

I didn't try the optical input on exaSound e28 yet, but maybe I'll try it too at some point when I move it somewhere close enough to my Linux desktop machine.

[fmak]

that posters who do not agree with you are simplistic and make statements based on one set of

experience and not that gained over many years with different high end systems.

 

I can tell you that even though RME claims 192k operation in Toshlink, they don't, and not just with one receiver.

 

Frankly, if you can't hear differences between Toshlink and Electrical, then you haven't got a very good system.

 

The Asus you have is a mid price box and I would have thought that you would have used something much better.

 

I have to post in response to an earlier post of yours because, for some reason, you current posts are preventing

responses by defaulting to 'not signed in'. This doesn't seem to apply to posts from others.

 

 

Again, I cannot post a reply direct to you. Webapges ays I am not logged in when I am.

 

Response to your post:

 

''works also perfectly fine from the ASUS Xonar DX to Fostex HP-A8C DAC/headphone amp.''

 

AS for telling differences amongst digital filters, this is easy even on fairly low quality but good

functioning equipment.

 

Hearing cable, interface and relocker differences, one may need another level of equipment quality.

 

I can hear all these things on my dCS and W4S2DSD systems. The Mytek has too many compromises

built in ie.Low quality PS, cheap clock module to be used as a reference - I have one. The Mastering

version even has 50R bnc sockets instead of the proper 75R ones. It does well in terms of sonic

compromises and one pays for the many methods of connection only.

[Miska]

Again, I cannot post a reply direct to you. Webapges ays I am not logged in when I am.

 

Cannot say anything about that, you'd need to complain to Chris here:

http://www.computeraudiophile.com/f8-general-forum/current-computer-audiophile-site-issues-22207/

 

''works also perfectly fine from the ASUS Xonar DX to Fostex HP-A8C DAC/headphone amp.''

 

Xonar DX is a PCIexpress sound card, not any box. Used here as bit-perfect Toslink transmitter from computer to the DAC. Although it has pretty decent DAC based on CS4398 too, on Linux you can even switch the chip's digital filter. Obviously it is a very good Toslink transmitter since it works very well with my DACs up to 192k.

 

If your devices cannot do 192k with Toslink, they obviously have substandard Toslink implementation. Pretty simple.

 

Hearing cable, interface and relocker differences, one may need another level of equipment quality.

 

OK, I have zero interest on that area. When I hear differences between digital cables, I conclude that the DAC is poorly implemented.

 

When there are differences with analog cables, I conclude that the transmitter side doesn't have proper cable compensation.

 

 

P.S. 90% of my testing is with Sennheiser HD800 headphones plugged straight to the DAC.

[Miska]

I can hear all these things on my dCS and W4S2DSD systems. The Mytek has too many compromises

built in ie.Low quality PS, cheap clock module to be used as a reference - I have one. The Mastering

version even has 50R bnc sockets instead of the proper 75R ones. It does well in terms of sonic

compromises and one pays for the many methods of connection only.

 

Which dCS?

 

Do you have some objective measurement evidence that there would be something wrong with the Mytek implementation? I have no complaints about the performance, based on my measurement.

 

dCS Vivaldi is clearly having some issues:

 

Same goes for Debussy:

[fmak]

Cannot say anything about that, you'd need to complain to Chris here:

http://www.computeraudiophile.com/f8-general-forum/current-computer-audiophile-site-issues-22207/

 

 

 

Xonar DX is a PCIexpress sound card, not any box. Used here as bit-perfect Toslink transmitter from computer to the DAC. Although it has pretty decent DAC based on CS4398 too, on Linux you can even switch the chip's digital filter. Obviously it is a very good Toslink transmitter since it works very well with my DACs up to 192k.

 

If your devices cannot do 192k with Toslink, they obviously have substandard Toslink implementation. Pretty simple.

 

 

 

OK, I have zero interest on that area. When I hear differences between digital cables, I conclude that the DAC is poorly implemented.

 

When there are differences with analog cables, I conclude that the transmitter side doesn't have proper cable compensation.

 

 

P.S. 90% of my testing is with Sennheiser HD800 headphones plugged straight to the DAC.

 

 

I now understand where you come from, and arguing about whether the Xonar is a box or not

carries no interest to me.

 

The Xonar is a middle ranking card - I have one. It doesn't even sound very good.

 

You appear to be a software person and are not bothered about lower quality hardware. I am not

therefore interested in a continuing conversation.

[Miska]

The Xonar is a middle ranking card - I have one. It doesn't even sound very good.

 

It doesn't have "sound" as it is just sending bit-perfect data out over Toslink. And it is good at precisely that as you can see from my earlier messages.

 

You appear to be a software person and are not bothered about lower quality hardware. I am not

therefore interested in a continuing conversation.

 

I am both software and hardware designer, as you can see from my DSC1 DAC design etc. I've been designing building ADCs and DACs for 20+ years now.

 

My background is in passive sonar systems and industrial measurement systems. (plus of course DIY audio for myself since 80's)

 

 

I'm not interested to continue subjective hands waving about "sound" of digital transfer systems. I'm only interested on hard objective data.

[sandyk]

Quote Originally Posted by fmak View Post

 

The Xonar is a middle ranking card - I have one. It doesn't even sound very good.-fmak

 

It doesn't have "sound" as it is just sending bit-perfect data out over Toslink. And it is good at precisely that as you can see from my earlier messages.

 

Hi Miska

I am surprised to read that from you of all people. Through my Class A amplifier, and that of my friend's way better than average system , both via Xonar D2X or Xonar Essence cards, all of the Toslink cables that I have heard suffer in the area of 3Dimensionality . The same applies to coax vs Toslink with several different Oppo and other players.(Sony SA11) Even well made 75 ohm coax leads of the same length can sound a little different in this area. I was surprised to hear an improvement a while back with a $15 Concord cable over a well made RG213 cable that originally saw life as a patch cable in a Telstra 12MHZ carrier system before an original connector broke. Most readily available Toslink receivers and transmitters are mediocre, and the specs often vague. Some may even tend to oscillate, as Silicon Chip magazine found with one of their DAC designs.. IIRC, they fitted low value ceramic caps (33pF) across their output.

 

Regards

Alex

P.S.

I am surprised to see you appearing to push the "bits are bits" line.

[Miska]

I am surprised to see you appearing to push the "bits are bits" line.

 

Making sense out of digital transmission, even poor one, is job of the receiver. Better receivers can produce good results even with data from bad transmitters and cables. Just like with FM radio, better the radio receiver, better sound quality you get even in poor reception conditions.

 

A good DAC gives consistently the same best possible result regardless of how bad cable you connect to it.

 

In this case, the DX's job is to work as optical transmitter operating at 176.4/192 kHz sample rates. If you have better alternatives to offer for this particular job (preferably as PCIexpress card), I'm happy to hear about those. I can also use RME's card if necessary, but in this particular job I've been happy with performance I get from the DX and have not seen reason to change it.

 

What is important from testing perspective is consistency and reproducibility, you should change only one thing at a time in a system, otherwise knowing where a change came from becomes much harder.

 

This thread was about using Toslink for 176.4/192k rates and I just commented that it has been working perfectly fine for me with the stated hardware. I'm not interested to play with digital cables, unless someone can objectively rate those in some sensible order, not by means of which one "sounds better" for person X, because person Y could have completely different opinion, but by means of objective data.

 

I select my analog cables based on technical performance and been happy with Supra products so far. IOW, interconnects with low capacitance and speaker cables with low inductance.

 

If you think how many devices and cables there are on the market and how many combinations you could make out of these, attempting to iterate through all possible combinations by trying out is plain impossible. So some other way needs to be used.

 

I'm not going to accept that simply more expensive would mean better, I have seen too many cases where that's not the case.

 

 

I was surprised how well the TCAT JetPLL can actually perform for S/PDIF inputs (Toslink, coax or AES/EBU).

[sandyk]

I'm not going to accept that simply more expensive (cables) would mean better, I have seen too many cases where that's not the case.

 

 

Miska

Agreed. I am much like you in that respect. However, the mere act of another additional conversion to Optical and back again does introduce "Jitter" due to factors such as PSU excellence for both TX and RX and limitations in the optical devices themselves, and despite the best efforts of Schmidt triggers such as the 74HCU04, these differences may still be heard in a highly resolving system.

I am far from being the only member who reports hearing differences with well respected gear between Toslink and Coax SPDIF, even differences favouring glass Toslink over normal plastic Toslink are reported by quite a few members.

Perhaps Toslink using a low power laser diode instead would perform a little better too ?

This may also be another area where Peter St. and yourself may not be quite on the same page ?

You both get great results, but by very different methods

Regards

Alex

[Harpy]

I run 24/192 Toslink to my Gustard X-10 with no issues. It's newer Chinese DAC.

[Miska]

Agreed. I am much like you in that respect. However, the mere act of another additional conversion to Optical and back again does introduce "Jitter" due to factors such as PSU excellence for both TX and RX and limitations in the optical devices themselves, and despite the best efforts of Schmidt triggers such as the 74HCU04, these differences may still be heard in a highly resolving system.

 

I would say it is all up to the receiver side. Better the receiver, more it can filter out jitter. JetPLL has 100 dB jitter suppression. You can also run S/PDIF input into FIFO and clock data out of it by a second clock, like the Wolfson WM880x.

 

What I'd be careful about is the "differences may be heard by highly resolving system" part. Where it could simply mean that it has sub-standard receiver that emphasizes the differences between cables and transmitters or could be affected even by sunspots. For me, good highly resolving system can play out even the smallest differences that are included in the data, but will play the same data exactly same way regardless of how it got there.

 

I am far from being the only member who reports hearing differences with well respected gear between Toslink and Coax SPDIF, even differences favouring glass Toslink over normal plastic Toslink are reported by quite a few members.

 

I would say something to fix in the electronics, you should get a better receiver... Continue fixing it until the cable no longer makes a difference.

 

This may also be another area where Peter St. and yourself may not be quite on the same page ? You both get great results, but by very different methods

 

I think I recently heard him saying that his latest NOS1 DAC is not anymore sensitive to the computer, IOW sounds the same with same data regardless of computer optimizations...

[kdoot]

There's been a related discussion over at the PS Audio forums, in the DirectStream DAC section.

 

I've just bought a new DAC and one of the things I love the most about it is its Toslink input. PS Audio only guarantees up to 96kHz on that input, but some units just happen to be OK all the way to 192kHz. Happily, mine is one of those.

 

As for the sound quality of toslink, I agree with Miska *in principle* that an ideal DAC should care about nothing but the bits. I also agree with others that *in practice* many DACs have struggled to sound good with Toslink - which I believe is due to the difficulty of extracting a clean clock from a toslink signal. While toslink is totally immune to electrical noise, the jitter is severe, and most DAC designs fare much better with the sharp transitions possible on coax yielding a more consistent clock.

 

I'm trying not to sound too much like a sales pitch, so I'll just add that my new DAC sounds awesome via optical and leave it at that.