I would suggest that an Ethernet-Ethernet renderer would be redundant. I do not usually advocate reducing noise by increasing components (but sometimes it is necessary). But i would caution folks on using DACs with built in Ethernet interfaces. Most DACs with built in Ethernet interfaces do not have the best implementation of their Ethernet interface. I will site the Linn Klimax DS as being an exception here, and I am not against built in Ethernet interfaces, but most DAC manufacturers are far behind when it comes to Ethernet interface development. Putting an Ethernet interface in a DAC is tricky to do right, it requires extensive shielding and power supply isolation to work right and not interfere with the DAC/Analog circuitry. Remember, the noise coming in over Ethernet is not the only potential problem, the interface itself is a small computer and this needs lots of attention to implementation. Also a lot of built in interfaces have sample rate limitations which may not be acceptable to those who prefer to oversample/filter in software.
Help me to better understand the digital terms I2S and Ethernet Renderer
- Thread starter steve williams
- Start date
I would suggest that an Ethernet-Ethernet renderer would be redundant. I do not usually advocate reducing noise by increasing components (but sometimes it is necessary). But i would caution folks on using DACs with built in Ethernet interfaces. Most DACs with built in Ethernet interfaces do not have the best implementation of their Ethernet interface. I will site the Linn Klimax DS as being an exception here, and I am not against built in Ethernet interfaces, but most DAC manufacturers are far behind when it comes to Ethernet interface development. Putting an Ethernet interface in a DAC is tricky to do right, it requires extensive shielding and power supply isolation to work right and not interfere with the DAC/Analog circuitry. Remember, the noise coming in over Ethernet is not the only potential problem, the interface itself is a small computer and this needs lots of attention to implementation. Also a lot of built in interfaces have sample rate limitations which may not be acceptable to those who prefer to oversample/filter in software.
Empirical Audio
Industry Expert
Here are some jitter measurements on my renderer and some DIY renderers:
http://www.audiocircle.com/index.php?topic=154299.0
This renderer interface is also in my new DAC, the Overdrive SE. The review is due out in the next 2 weeks from Audiostream. Jesus helped me get started with the renderer BTW. Much appreciated.
Steve N.
Empirical Audio
Industry Expert
BTW, Barrows, do you know if your gear is compatible with Ravenna?
http://www.merging.com/products/pyramix/ravenna
Does this require a new interface for a different protocol?
Thanks,
Steve N.
Empirical Audio
http://www.merging.com/products/pyramix/ravenna
Does this require a new interface for a different protocol?
Thanks,
Steve N.
Empirical Audio
Properly executed as I'm sure Steve N has done, Ethernet has to be the winner. My next question is, why does the Ethernet Render have to be in a different chassis than a DAC?
I never suggested the Renderer has to be in a different chassis, I just suggested that MOST current Ethernet inputs incorporated in DACs are sub-optimal. And many also are quite limited in terms of sample rates.
Details really matter here, as we are talking about placing a small computer inside a DAC chassis, to keep this from being a problem will require great attention to detail, just making a great Ethernet interface is difficult enough that most DAC manufacturers use a third party solution (and often a very affordable one at that).
As for Ravenna, my understanding is that for DAC manufacturers Merging sells the actual interface (hardware) so that will be another option: Ravenna is completely different protocol than what Ethernet currently uses, and appears to have benefits for running many channels (24+) of audio in the pro environment, but there appears to be no benefit over current methods when they are done correctly. I doubt we will se the Merging/Ravenna interface inside a lot of affordable DACs, as I expect it will cost around 10x that which is most commonly used now.
Details really matter here, as we are talking about placing a small computer inside a DAC chassis, to keep this from being a problem will require great attention to detail, just making a great Ethernet interface is difficult enough that most DAC manufacturers use a third party solution (and often a very affordable one at that).
As for Ravenna, my understanding is that for DAC manufacturers Merging sells the actual interface (hardware) so that will be another option: Ravenna is completely different protocol than what Ethernet currently uses, and appears to have benefits for running many channels (24+) of audio in the pro environment, but there appears to be no benefit over current methods when they are done correctly. I doubt we will se the Merging/Ravenna interface inside a lot of affordable DACs, as I expect it will cost around 10x that which is most commonly used now.
I would feel tempted to say the same. Theoretically almost all digital devices would be redundant ... Bits are bits as they said ...
However, considering that people have reported that just replacing the Ethernet card of their PC by a better card, such as the JCAT NET Card FEMTO, improved the sound quality of their systems, I feel we can be sure.
Empirical Audio
Industry Expert
Empirical Audio
Industry Expert
I could say the same for USB. Not very many good implementations in DACs.
Steve N.
Empirical Audio
Yeah Steve, at least with Ethernet interfaces most designers have an excuse of not knowing how to do it! With USB implementation it is pretty "known" how to do it right now, and still many, many DAC manufacturers take short cuts with their USB implementation-no excuse for this really at this point.
I agree on all points. One example is they are still using USB 2.0 when 3.0 has higher bandwidth. Whether there are "cheap" USB 3.0 subsystems ready to drop in their implementations, its not my turn to watch but I think there is an inexpensive USB 3.0 based interface gadget that claims to help, so the subsystems can't be that bad $. All I know it is easier to scale a system down than up. Properly implemented Ethernet has no jitter by design. Steve N. points out some of the issues he has run into with grounds and so on.
I sit back and look at people spend megabucks on multiple chassis systems with streamers, renderers, DACs, switches, NAS, etc. for digital. Sad. Used to Spotify, Tidal SiriusXM, iTunes and other MP3 sources, many are led to believe 44.1k/16b signals are "high resolution." MQA will save the day. Rubbish. Why not take a look at primephonic?
Read Kal's column in "Music in the Round" in either the Jan or Feb. 18 issue of Stereophile. Roughly paraphrasing he describes a multichannel studio recording at DSD256 or higher perhaps mixing down to 7 channel DSD, monitoring with 7 Sound Lab M-645 panels. I wish he would have taken a picture - pin up of the year. Now that is a transparent, high resolution system. I have only two U-645s, and no more room.
More's the pity.
Last edited:
Empirical Audio
Industry Expert
That would be nice, but there is always jitter. My latest Ethernet renderer measures 18psec.
The technology treadmill is what a lot of audiophiles feel will lead them to nirvana. It's actually the NOS R2R chips that will only do 24/96 that are really shining today, as well as other DACs that have good implementations and did not chase the DSD and MQA treadmill.
I would have to hear it. I hear a lot of these demos at trade shows. I am almost never impressed. I had rooms with megabux dCS Vivaldi systems on both sides of my conference room at one RMAF that I listened-to. I would rather listen in my room, thank you. Most people think I'm playing hi-res or DSD when it's actually 16/44.1 CD rip. It's all about implementation and being clever in the design. The fact is that most audiophiles love the music from the 70's or just play classical etc., so much of the music that is dear to them is not available on DSD or MQA. These people need good sounding systems at 16/44.1.
Steve N.
Empirical Audio
Fortunately you prefer your own room - otherwise something would be wrong! FYI the Vivaldi system plays CD's excellently - one of the reasons I got it was just because I listen mainly to CD's - although a SACD plays at a different level in this player.
MELCOs available in NA and UK, EU have a similar topology. I'll have to check the FiData.
Although I don't always agree with Steve N. I respect his point of view. He's an honest engineer and business person - he's making the best solution he knows, and keeps improving it! - for a given market segment.
BTW I went back and read Kal's article and I did get some of the details incorrect, albeit in haste. I was just "in a plethora" to have someone at Stereophile mention (the new) Sound Labs electrostatic speakers after all these years.
BTW I went back and read Kal's article and I did get some of the details incorrect, albeit in haste. I was just "in a plethora" to have someone at Stereophile mention (the new) Sound Labs electrostatic speakers after all these years.
Empirical Audio
Industry Expert
Thanks for the vote of confidence. Here is the latest improvement BTW:
https://www.audiostream.com/content/empirical-audio-overdrive-sx-ethernet-dacpre
Steve N.
Empirical Audio
Similar threads
