Simple Streamer with 192/24 SPDIF output?

MDS? Do you mean Linn Magik DS - if so I understood that would do 24/192 via SPDIF. Not sure about any other device which supports 24/192. I think the PS Audio (stand alone) Bridge will but thats not yet available.

I would ask yourself the question if you really need 24/192 at the moment?

Eloise

Hu Eloise,

you´re right the MDS can output 192/24 if you disable the upsample option in Konfig. If enabled, it always sends 96 or 88. My fault.

I agree that the requirement of 192/24 is not absolutely necessary, but there are more and more highres downloads available every day. Additionally there are also a bunch of DVD-As with that rate. If I buy a streamer today, I just want a device which is not already outdated.

thx n brgds...TC

quote:

Originally posted by yeti.fro:
I agree that the requirement of 192/24 is not absolutely necessary, but there are more and more highres downloads available every day. Additionally there are also a bunch of DVD-As with that rate. If I buy a streamer today, I just want a device which is not already outdated.

That's cool - I think some people see "24/192" and think they automatically need it - you've obviously thought about it.

Eloise

quote:

Originally posted by yeti.fro:
Hi guys,

are there any standalone UPNP streamers out there that do offer 192/24 via their SPDIF digital out?

UnitiQute and T+A 1260 doing 96/24 only. Same is told about MDS (not tested yet), some rumours say that a MDS can do 44/16 only. Olive doesnt specify. Sonos is 48/16. Hifdelio is 44/16.

I dont want to use a PC and definitively no Mac because of noise and complexity, but is there any alternative to a Linn DS to play 192/24?

thx n brgds...TC

If you forgo on the UPnP requirement there are plenty reliable solutions

The UnitiQute can do 192/24 when it is received over its SPDIF inputs.

The UnitiServe will do 192/24 over its SPDIF-output which can than be connected to a DAC.

Why would you want to use a UPnP-server as source, if you could also connect to the network directly.

I use the Popcorn NMT, it can do UPnP-rendering, but all these UPnP implementations are rather unreliable IMO, and you need to find a combination that happens to work together.

I also have an Oppo BDP-83 which can also act as UPnP-renderer and is capable of playing 192/24.


-
aleg

The Linn Sneaky DS fits your requirements

Aleg,

what makes you think that UPnP is unreliable? I have/had various UPnP renderers (Linn, Denon, T+A etc) for more than 2 years and all are working fine. Installed at least a dozen setups for friends and never had any problem. All are nicely controllable by Kinsky, so you have the same structure for all players at home.

I just find if a bit strange that all DACs today accept 24/192 signals but there are hardly sources... even Naim does not have one (yet).

The UnitiServe is again a simple PC in a nice black box and it shares (almost) all flaws of the HDX. It makes no sense to have a server in the music room creating noise, heat and vibrations and requiring another backup strategy. But the SSD versions dont make sense either, because they still act as a server and are not designed to work properly as a client.

However, the most perverse thing is to corrupt the jitterfree ethernet transfer by sending the data through an ancient SPDIF interface again. This is like routing the HD videos from your Blueray player through an VCR to an analog CRT-TV... still trying to find a nice setup with an nDac though...

thx n brgds...TC

quote:

Originally posted by yeti.fro:
Aleg,

what makes you think that UPnP is unreliable?

Bad experience
I have tried in the past to setup UPnP with servers like Twonky, Asset, TVersity, Popcorn; with control points like Kinsky, Foobar, PlugPlayer,Oppian, Popcorn, EEZUPNP, Cidero, EyeCon; with renderers like PC with Foobar, Popcorn.
And I have not succeeded to get a single one combination to work properly, and I'm no noob with computers I can assure you.

That's when I gave up on UPnP and it isn't a necessary thing either IMO.

quote:

...
The UnitiServe is again a simple PC in a nice black box and it shares (almost) all flaws of the HDX. It makes no sense to have a server in the music room creating noise, heat and vibrations and requiring another backup strategy.

You should think about the SSD-version which doesn't have the flaws you mention, even though I think Naim will have done their best to minimise all the things you mention.

quote:

But the SSD versions dont make sense either, because they still act as a server and are not designed to work properly as a client.

IMO it is specifically designed to be a client as well and I think it does so nicely as well looking at the user interface on both desktop and iPod.

quote:

However, the most perverse thing is to corrupt the jitterfree ethernet transfer by sending the data through an ancient SPDIF interface again.

You do throw around some harsh value statements. I wonder if you can actually hear the jitter produced by the UnitiServe over an 80 cm interconnect into the Naim DAC.

quote:

This is like routing the HD videos from your Blueray player through an VCR to an analog CRT-TV...

I think this is BS. SPDIF delivers 192/24 quite nicely and I don't think for one sec, that you can hear your 'bad' quality of SPDIF.

-
aleg

quote:

Originally posted by yeti.fro:
However, the most perverse thing is to corrupt the jitterfree ethernet transfer by sending the data through an ancient SPDIF interface again.

Regardless of any actual judgment on how bad jitter from a UnitiServe into DAC via SPDIF actually is ... how does this tie in with your original post/question...

quote:

are there any standalone UPNP streamers out there that do offer 192/24 via their SPDIF digital out?

Eloise

quote:

Originally posted by yeti.fro:


However, the most perverse thing is to corrupt the jitterfree ethernet transfer by sending the data through an ancient SPDIF interface again. This is like routing the HD videos from your Blueray player through an VCR to an analog CRT-TV... still trying to find a nice setup with an nDac though...

thx n brgds...TC

Not true. Where did you read Ethernet is jitter free?

There is a packetizing on the transmitter end and there is a receiving end receiving the packets as well with power supply even in case of the Ethernet transmission. Generating and receiveing the Ethernet data packets requires CPU, and there is a power supply and jitter where there is a CPU. So no, Ethernet transmission with its IP-based packetizing and different IP-based protocols are not jitter free and not to cure all of the problems of the SPDIF/AES transmission without introducing other problems.

SPDIF (and AES) are well known and trusted. In the studios where the recordings made there are no other devices, only AES and SPDIF, no Ethernet devices there. There are different Ethernet and IP-based transmissions for Live and PA, where you need to move lots of channels in real-time and the distances are much bigger than in home. In this case it makes sense to use a cat5 or 6 cable and a kind of IP based transmission, but even in thsi case not the jitter free nature is the reason why it is used. If I remember well, Stereophile measured something like hundredsomething ps jitter on the KDS while the CD player part of Uniti is something like 70 ps. Please correct if do not remember well.

Probably the most modern (and one of the most interesting) implementation of using cat 5 cabling (without IP packetizing) is the MSB Network which transmits I2S signals with clock in akind of serial fashion. It can transmit 384 kHz/32 bit PCM audio and digital video as well, using 295 Mbps serial LVD signal. But even in this case the manufacturer is not brave enough to mention jitter free transmission, they say: "excellent jitter and noise rejection with low radiated emission".

Jitterfree. Noisefree too? I want this perfect world.

quote:

However, the most perverse thing is to corrupt the jitterfree ethernet transfer by sending the data through an ancient SPDIF interface again.

Total misinformation - Ethernet is prehistoric just like TCP/IP - you'll be saying Windoze is good product next - it isn't all three Ethernet, TCP/IP and Windoze are badly flawed (though I understand v7 is not as awful as previous MS atrocities). Designers have to make the most of these arcane technologies when forced to used them.

S/PDIF done the Naim way is about as good as it gets.

UPnP is awlful and as Aleg says redundant (a solution waiting for a problem) - surely if you use iTunes on a Mac and want to use a NAS you could just use NFS.

Excellent post by Ferenc too - MSB sounds interesting.

quote:

I just find if a bit strange that all DACs today accept 24/192 signals but there are hardly sources... even Naim does not have one (yet).

HDX, UnitiServe...? Both platforms play 192/24 files without a problem.

Hi guys,

I´m sorry if my statements sounded harsh. They were not meant to be, pls blame it on the foreign English language.

But wasnt it Naim itself telling us for years that the SPDIF connection can only decrease sound quality? That´s why their CDPs didnt have a dig out and they had no DAC?. With a lot of technical effort they now managed to minimize this sound decrease, but it´s still there. Just try different digital interconnects and hear it for yourself. If Jitter etc. wouldnt be an issues, all ICs would sound the same. A lot of DACs do have the option to have additional clock links between source and DAC. In all tests I´ve done, those links improved sound quality significantly, so there´s definitively a benefit. I dont say that SPDIF cant produce excellent results, but IMHO a packet based link, while not perfect either, should be able to improve things big time (it surely improves the clocking situation). 44/16 can sound great, but nobody argues that 96/24 has much more potential.

Eloise,

although I believe SPDIF is the wrong way, I really would like to have a NAIM source. Since the HDX is not for me and there´s no Naim streamer, I need at least something to feed the nDac. My devotion to Naim and my technical mind are fighting a bit

Aleg,

the SSD versions are IMHO crooked. They only try to cure some symptoms but not the root cause. The HDX SSD is still a "simple" HomeServer PC with big problems and a noisy fan. The Flash userinterface is almost unusable with WinMobile or Droid. Marketing sayz it copies the touchscreen interface, but this is an arrant lie. Try to search for an Artist and you know what I mean.

ROTF,

so what´s your opinion of Naim selling Windows PCs (HDX, UnitiServe)? You´re moving to Linn because of that? Apple IMHO is one of the most aggressive and sneaky companies around and I definitively will not buy anything from them. They always had much more lawyers and designers then engineers and that´s clearly visible in any product, not only since Antennagate. And Steve J proves every day that he doesnt care at all for his customers.

Dave,

HDX and UnitiServe are servers. I´m looking for a client, because I already do have a proper server running. The UnityQute would be one, but it does not have 192/24 output.

Maybe in a couple of month Naim will present its reference class NDX, streaming music directly into a built in DAC and by avoiding the shivering SPDIF it will become the best sounding digital source ever beating the hell out of 555 and KDS And when that happens I hope you remember this nice little thread

thx n brgds...TC

quote:

you'll be saying Windoze is good product next

Its good enough for Naim to use in the HDX

I have tried 192 from various sources and failed to be impressed. I certainly have 44/16 content from CD:s with much more impressive qualities.

Dan Lavry has some interesting views on sampling rates - higher means less accuracy. He suggest 60kHz as the optimum tradeoff between bandwidth and accuracy.

quote:

Originally posted by jlarsson:
I have tried 192 from various sources and failed to be impressed. I certainly have 44/16 content from CD:s with much more impressive qualities.

Dan Lavry has some interesting views on sampling rates - higher means less accuracy. He suggest 60kHz as the optimum tradeoff between bandwidth and accuracy.

IIUC it is very mu h dependent on the quality of the clock of the product, higher sampling frequencies require much more stable clocking than lower frequencies. There is less margin for errors.

-
Aleg

quote:

Originally posted by ferenc:

Where did you read Ethernet is jitter free?

Where did you read that data transimitted over a network has jitter added to it? How can it? There's no embedded clock to jitter. Does that mean downloads from the NAIM store have ethernet induced jitter? Better not download them then.

quote:

Originally posted by ferenc:
.. and there is a power supply and jitter where there is a CPU.

There are two CPUs and a power-supply already in the nDAC. Better not use it then either.

quote:

Originally posted by ferenc
So no, Ethernet transmission with its IP-based packetizing and different IP-based protocols are not jitter free.

They are absolutely jitter free, since there is no embedded (jittery) timing signal. The construct just doesn't apply. No wow and flutter either. Or rumble even.

quote:

Originally posted by ferenc:


Not true. Where did you read Ethernet is jitter free?

Where the hell is Andy when you need him????

quote:

Originally posted by yeti.fro:
there´s no Naim streamer, I need at least something to feed the nDac. My devotion to Naim and my technical mind are fighting a bit

YES there is!!! The Unitiserve. As "standalone" as a streamer gets.

Can anyone tell me of an "streamer" that is a true standalone streamer?

The DS include a DAC.
The Serve includes a ripping drive.
The Squeezeboxes include a DAC and/or touch screen.
The HDX includes DAC and touch screen.

It seems that almost EVERY option out there has one or two add-ons that put people off. The biggest offender in my book is the DS player which limits inputs to one and can ONLY be used for computer audio.

The Serve+Dac seems ideal when you think about it. It can be used with ANYTHING with a digital output, and also anything streaming from a computer.

-Patrick

You still have to pull the clock and make a stream from packets after data transmition. plenty of places after transmition to add noise and jitter.

quote:

ROTF,

so what´s your opinion of Naim selling Windows PCs (HDX, UnitiServe)? You´re moving to Linn because of that? Apple IMHO is one of the most aggressive and sneaky companies around and I definitively will not buy anything from them. They always had much more lawyers and designers then engineers and that´s clearly visible in any product, not only since Antennagate. And Steve J proves every day that he doesnt care at all for his customers.

More misinformed rhetoric I'm afraid; I can tell from you what you say that you haven't dealt personally with either Apple or MS. I guess you don't work in IT or technology, which is probably no bad thing as I wouldn't chose that profession if I had my time over. Very few computer companies are customer focused in the way Linn, Naim and Rega are. However, Apple are rather more pleasant to deal with than MS and do make better products because of the quality of their designers and engineers: I know this because I've worked with some though never as an Apple employee. If you find Apple aggressive then heaven help you if you ever cross swords with the other lot. Still, I'm sure I won't convince you; so I'll stop trying.

IMHO, Naim should avoided Windows, but I still might buy a UnitiServe because I know I'll get great service from Naim and never ever have to have any dealings with you know who. Of course, I'd much prefer they used a TripOS system (Amiga) or Inferno (Lucent) or even Linux, but I understand that might increase the R&D budget and the cost to you and me.

BTW I like Linn as a company, I have a Linn CMS and LP12SE/Ekos SE/Radikal, but the ADS didn't do it for me.

I hope you find what you're looking for.

ATB Rotf

Ferenc is right in what he posts, Ethernet does not eliminate jitter. That is the inherent problem with Voice over IP networks - it is why concepts like IP Class of Service were added to modern networks to prioritise streaming traffic over carriers like Ethernet, which are prone to random fluctuations caused by packet collisions. Ethernet has a jitter characteristic written in to its specification. To overcome jitter, network designs use de-jitter buffers to ensure a continuous audio. Now, if DAC designers were clever they'd use a buffer - oh hold a minute, the two best sounding DACs I've heard do ....

It is completely different with the Naim store as you don't use UPnP or any other real time streaming protocol to download music. It is http as far as I recall.

Bring back Andy.

ROTF,

I´m sorry, I have to disappoint you, I´m working in IT for almost 20 years and dealt with many companies. That´s were I have my experiences from. But there is a hugh difference if you have an Apple workstation for €10k or an iRubbish. Apples has created almost a monopoly with iTunes and this for me is not acceptable, but offtopic.

You are totally mixing up the jitter terms for VoIP and SPDIF. In VoIP the packets jitter, in SPDIF the bits themselves.

For VoIp over a packet or cell oriented network, jitter is indeed the timedifference between two packets. Not a big deal as long as the difference doesnt get two big because you have a buffer and can realign. This jitter does not affect the payload of the packets as long as the packet is not discarded.

This has nothing to do with the jitter in SPDIF. This is caused because there is (usually) no separate clock sync and the clock is hidden in the signal. There the jitter directly influences the content and hence the output.

You cant compare a complex WAN with 3m of LAN cable either. There´s surely no relevant packet jitter in a simple not overloaded LAN.

Every DAC jitters, more or less, but it´s obvious that jitter inside a DAC is much more controllable then outside where you have a big unknown component with the IC, connections not designed for a digital signal (Cinch) and the signal is much more prone to other influences. In many CDPs the clock is put as close as possible to the dac to eliminate exactly those effects. If 5cm of component placement inside a box make a difference, it doesnt get better by creating several dm of distance? Yes, of course the external DAC has it´s own clock, but this cant impact the source.

But maybe jitter is not the main flaw.... what about wander?

brgds..TC

Re: likesmusic:

data packets and frames plus (word)clock transmitted through Ethernet in packetized form. In case for example the most used so called Cobranet or A-Net, pro Ethernet and iP-based multi-channel audio transmission, the (word) clock is transmitted in a so called broadcast packet.

It means timing information is sent along with other switched packets on the network, and it is question of implementation how low the overall jitter of the transmission chain will be, but there is no jitterfree transmission in audio be it pro or high-end, for sure.


Even if you dumb down your IP-based audio transmission to loose some of the bandwidth hungry overheads of IP protocols (which is UPnP surely not - it is kind of bandwidth hungry implementation of IP for audio with quite high latency) you can make the latency and jitter lower, but you can not get rid of it completely.

In case of KDS for example there is a large buffer and fixed frequency reclocking in the close proximity of to the Wolfson chip and the Ethernet receiver and even in this case the overall jitter of KDS is bigger than the jitter of Uniti playing CD or a jitter of other well known cd players. Funny isn't it?

Few more random notes regarding Ethernet, SPDIF and jitter.

SPDIF (and AES) with all of its problem, buffering, clock recovering well known and researched. With all the fast DSPs, buffers, precise clock generators its jitter can be kept in a good shape quite easily.

Ethernet and mainly software/OS/protocol induced jitter are not that well known yet in high-end audio, it is knwon mainly from the telco/compuet industry and from pro audio, where the priority of research is a little bit different.

The UPnP communication is not even optimized for jitter, latency so for high-end audio even if it is using Quality of Service. Because of the wide variation of connectable hardware devices and (sometimes embedded) software players, the interoperability and compatibility is more important than low jitter and latency and the overall sound quality. Because of compatibility and video handling, it surely includes layers of elements which are just making the overall overhead bigger and more cpu power and memory hungry. In case of the new HDX for example it is a good step from Naim to provide its own UPnP server(even if we do not know how it differs), but it is only a first step. I am not sure however high-end audio companies have enuogh resoruces, budget to optimize UPnP for high-end audio ONLY purposes to get rid of the unnecessary elements of the protocol. This is the reason why I feel a solution like MSB Network is a good idea. It is optimized for multi-channel, high quality, high-res audio with low latency, tranmits the clock separetely and can easily handle multi-channel 384 kHz/32 bit streeams without using lots of CPU power and buffering.

I do think there is no way to put all the different transmission methods above each other. You can get very good, musically satisfying solutions using SPDiF, USB, AES, Firewire, CobraNet, MADI, and God knows how many other ways. So be open and do not base a decision only on the technical information or the actual hype.