I2S Computer Build

[4est]

If there is, it is not by Amanero. AFAIK, they have licensed their circuitry and others are producing it and isolaters for it. Maybe one of them is?

There's more than 1 version of the Amanero though, isn't there? I was pretty sure I saw an LVDS version in the Melodious MX-U8 thread over on head-fi. The OP arbatels said he is planning on building a single I2s output USB DDC, and I gathered it was LVDS style. I'm not basing that on anything though.

[Maldur]

Yes, as far I can remember, Domenico gives licence to two more companies.

[Ben-M]

The Pink Faun card offers some possible advantages by being more closely coupled to the processor via PCIe rather than being a further step removed via USB. However remember that we have gone from SPDIF via PCI(e) cards partly for the very valid reason of removing the controller from the electrically noisy environment of a PC case. Most people have found USB to SPDIF devices superior for this reason.

 

A very interesting statement. I'd say it has to do with the quality of the card, not the concept. There's also PCIe to AES cards, I think from a company called RME, that're supposed to be the absolute bomb. The one I'm thinking of is about $900-1600USD, IIRC, and it's used in pro audio.

 

In my belief, if the Pink Faun card is engineered and implemented correctly, it could be totally amazing. But like you said, there are other advantages to separating the I2s stream from the PC clocks/power/ground/environment too. You've certainly made me think more deeply about their offering now and the possible better'ness of a USB->I2s DDC.

[Audio_ELF]

A very interesting statement. I'd say it has to do with the quality of the card, not the concept. There's also PCIe to AES cards, I think from a company called RME, that're supposed to be the absolute bomb. The one I'm thinking of is about $900-1600USD, IIRC, and it's used in pro audio.

As I say, I'm not trying to preach one true way, but just pointing out each has their advantages and disadvantages.

 

On the topic of PCIe to AES cards... I k ow in the past the RME and Lynx cards have been promoted as the "best" method of connecting computer to AES DACs such as Berkerley Alpha and Reference DACs, however the current advise is tending towards devices such as Berkerley Alpha USB being better. It's comments like this (from people who's comments I feel are worth listening to) which guides me towards suggestion the a PCI(e) card is not the best method - though as you say something like the Pink Faun card is different to such AES cards...

[toddrhodes]

My custom PF I2S card is on the plane, headed to the states. It's an RJ45 out pinned specifically for my Aqua I2S in. Also, my HDPlex LPSU and ATX DC-DC converter should be here tomorrow. Going to be a busy, and hopefully glorious week I just spent the last week and change dialing in my room with REW and I have to say, that by itself - a free flippin' tweak minus the cost of a mic, preamp, and stand - is a great tool. The funny thing is I'm back to exactly where I started with my room treatments (nailed it on the first try apparently) but I found that extreme toe in, crossing in front of my listening position yields by far the best experience in terms of stage depth (no surprise) and width is still pretty good, not as good as it could be but I have to work within my room's space and my own preferences, but reflections are at a bare minimum this way.

 

Off-topic, but I also observed some really odd behavior when adding in JRiver room correction so, for now, I'm just using the PEQ settings based on my in-room response curves from REW. Turning those off for a quick A/B test is frankly astonishing. Should have done this a long time ago.

[Ben-M]

As I say, I'm not trying to preach one true way, but just pointing out each has their advantages and disadvantages./QUOTE]

Certainly, and it's important to point out when we're looking at what might be best. Definitely got me wondering.

 

On the topic of PCIe to AES cards... I k ow in the past the RME and Lynx cards have been promoted as the "best" method of connecting computer to AES DACs such as Berkerley Alpha and Reference DACs, however the current advise is tending towards devices such as Berkerley Alpha USB being better. It's comments like this (from people who's comments I feel are worth listening to) which guides me towards suggestion the a PCI(e) card is not the best method - though as you say something like the Pink Faun card is different to such AES cards...

As for this, I haven't heard from anyone reputable or whose opinion I'm familiar with, so I can only take this with a grain of salt. I also don't have a means or venue to test for myself :-( Statements like that do make me wonder, though, as the others are mainstays in professional audio and geared towards being the best. And their prices are that way too. The Berkeley is in audiophile territory and can definitely ride the hype wave, whereas the others ride the "used for doing real work" wave. I'm no pro audio only preacher, I'm:') ist afraid of getting caught by the hype train that follows audiophilia.

 

Thoughts? And links :-)?

[Superdad]

At the end of the day there are many ways to achieve computer to DAC interface; and in the real world all require compromises to some extent.

 

The trouble with external I2S (even the LVDS variety) and S/PDIF (although it has a host of other implementation issues of its own) is that the clocking is virtually always in the wrong place. The best DACs always have one (pair of) master clock that everything is slaved to--otherwise they use some form of ASRC. But with external USB (or Ethernet)>I2S interfaces, the clock is in the box and gets sent TO the DAC. Almost nobody runs a clock line FROM the DAC to the interface.

Heck, even a lot of the available USB>I2S don't have a clock input line (though the Amanero, WaveIO, and JLSounds boards all now do allow it--but with some weakness in the area of isolators).

 

Anyway, I2S/DSD is obviously an end goal into the DAC chip, but I think the interface from the computer (for USB, Ethernet, etc.) to I2S belongs inside the DAC for ultimate performance.

 

That is not meant to take anything away from some of the very fine D-D converter boxes out there. Just recognize that they sound good because they may sound better than the DAC's other input implementations.

[Ben-M]

Just recognize that they sound good because they may sound better than the DAC's other input implementations.

 

Get it to sound as good as you can with what you have. In the end you're always going to be picking the best input your DAC has, not the best technical version of feeding something. I suppose I'm assuming that he has an I2s LVDS DAC, though.

 

Either way, the OP did say he didn't want to get too weighed down by theory and arguing about technical implementations and opinions, which is what that other I2s thread is full of.

 

So, who else has first had experience I2s and can help?

[vortecjr]

Kilroy, IMO these posts have done you a disservice. You had an idea and you where looking for group consensus. I don't fault you for that. However, I fault the people responding to you for clever wording in an attempt to push you in a different direction then the one you first proposed. You could have found gold if properly instructed, but you can't explore things if your playing follow the leader. I'll respond to each of the post with some comments in case its helpful to you.

 

Jesus R

[vortecjr]

I'm going down this path as well.

 

I'm building a CAPS Pipeline to replace my CAPS Zuma (with SOtM LPS + JCAT BPS).

The CAPS Zuma will then be changed to use the Pink Faun I2S board.

 

It'll likely be a few weeks before I have both done. Sound quality is my main goal too.

 

Pepi and Kilroy, I have made my share or i2s based music server and my advice here is that you use a good power supply to feed the computer and Pink Faun I2S board. At minimum use a good power supply to feed the Pink Faun I2S board.

 

Jesus R

[vortecjr]

Pink Faun has no NA distributors but they will sell direct for €227.27 excluding shipping.

 

I'm also seeing that this unit will not pass DSD, which may throw a wrench into my plans as I prefer to upconvert to DSD in the software player before passing to the DAC.

 

The Pink Faun I2S board should not have an issues streaming DSD64 via DoP. However, the Wyred4Sound DAC does not except DoP via the HDMI LVDS i2s input. For for that you would need to look at something like a PS Audio DS DAC. Having said that you could up sample to 192 or 176.4 as a alternate.

 

Jesus R

[4est]

I would be careful over thinking this. Your DAC should accept i2s or DSD w/o issues, and is likely the best input for your DAC.

The Pink Faun site has a page discussing some of the technical aspects. I also exchanged a few emails with Jord who seems to be a Pink Faun principal, here is an unedited excerpt from one email to me. I had explained how I currently upsample PCM to DSD in JRiver with USB out to my DSD se, and his response:

 

"I2S can only handle PCM signals, because I2S is a multi bit signal. The up-sampling in JRiver to DSD can definitely sound better if you use an USB connection to the DAC. I noticed this by my self, when I was using an USB connection.

Before we launched the I2S-bridge, I did some serious testing with PCM and DSD files over USB. Because I know, if we choose for the I2S-bridge, there's no room for DSD. The sound quality of DSD is absolutely not bad, and I think in some situations DSD will sound better than PCM. However, when playing PCM files over the I2S-bridge the quality is by far better than upsampling the files to DSD."

 

Intriguing.

 

 

[vortecjr]

I contacted Wyred4Sound and asked about it since my DAC has that input available. They said that most of the users that they were aware of were using it to connect to their W4S M2 music server.

 

My guess is that there are very few folks using it so not much anecdotal reporting or comparisons...

 

There are more people using i2s than you think....they just don't post about it for whatever reason. Probably because when they do they are told it's obsolete, warned about it's intended spec, and how superior USB should be;)

 

Jesus R

[vortecjr]

Very intriguing. Although my DAC with handle DSD natively I find that 99% of the time I am listening to Redbook or hi-res 24 bit downloads. I have tried upsampling to DSD but my i3 processor which I chose for its low heat in my fan-less case does not seem up to the task. I get dropouts or garbled output. I am tempted to try the i2S card.

 

To be honest this is just fun play as I am very happy with my current system even playing 16\44.

 

This is a great example of someone with common sense....

 

Jesus R

[vortecjr]

I'm using the I2S interface on my W4S DAC2 DSD-se via an Empirical Audio OR5. I vastly prefer the I2S input but I have to admit that I'm not sure if it's the interface or the OR5 that's making the difference. Probably both.

 

Probably both indeed.

 

Jesus R

[vortecjr]

Great to get some feedback. In looking up the Black Dragon I noticed they have the Hydra Z with is a sort of swiss army knife adapter for USB to different digital outputs, one of which is i2s via HDMI. Not too many reviews about it online though. Also the Sonore Rendu has an i2s output option although it is a network renderer and so would replace my local USB server altogether or at least use it as a streamer instead.

 

It is hard to evaluate a purchase of an i2s interface as there are few folks out there comparing them to a somewhat tricked out CAPS box which is what I use now.

 

I have tricked out boxed for lunch...LOL. I know you guys are looking for group support on this, but you just going to have to venture out on your own. I don't think you will regret regardless of what product you buy.

 

Jesus R

[kilroy]

Kilroy, IMO these posts have done you a disservice. You had an idea and you where looking for group consensus. I don't fault you for that. However, I fault the people responding to you for clever wording in an attempt to push you in a different direction then the one you first proposed. You could have found gold if properly instructed, but you can't explore things if your playing follow the leader. I'll respond to each of the post with some comments in case its helpful to you.

 

Jesus R

 

Jesus

 

Your inputs and clarifications are gratefully accepted.

 

Fact is I don't want to be a techie to enjoy the music, albeit some fundamentals are needed. It is a learning experience to a point, then starts to get too muddy. This is why I avoid the other I2S post.

 

Thank you.

[vortecjr]

This one even came with a warning:)

 

What you have to understand about i2s is that it is the native interface to the DAC chip inside your DAC. It was designed by Phillips (IIRC) in the mid 80s (and later developed in the mid 90s) to standardise interfacing between digital audio ICs (i2s or iis stands for Inter-IC Sound) and was designed for a few mm between devices on a circuit board. At some point various manufacturers extended this outside the box to connect transports to DACs.

You will be hard pressed to find a device that uses i2s over a few mm. I know of one, but I don't know anyone who uses it. Most devices break this requirement right on their own boards. A lot has been accomplished with i2s since that spec was published.

 

When PS Audio created their PWD DAC and PWT Transport (in the late 2000s) they developed an LVDS (low voltage differential signalling) version of i2s which utilised HDMI cabling (please note this is not a HDMI protocol just uses the cabling). If you are not using an LVDS version of i2s; please be aware that the device is using i2s in a way it is not designed for, and which may be of lesser quality even if it measures well.

This is classic fear mongering. Steve Nugent has done extensive testing on his DAC with standard i2s via RJ45 and he seems to prefer it over HDMI LVDS i2s. The manufacturers that support standard i2s via RJ45 do not seem concerned about the spec and neither should you.

 

Now in most cases your USB interface on your DAC will convert to i2s internally, the same for inputs such as Ethernet.

 

Right now thats cleared up you have to look at how to implement i2s for connection to a computer in the real world.

 

A computer does not output i2s. That is that if you wish utilise the i2s connection to your DAC you are going to have to use some form of interface. That may connect via USB (such as the OffRoad or they Hydra) or may be a PCI card such as the Pink Faun. This in itself is no guarantee of improvement over the USB input of a DAC; it is simply moving the interface from one location (inside the DAC) to another (in a separate box or into the PC). There is potential for improvement in quality because of a higher quality USB interface, better isolation, etc.

There are some computers that output i2s. For example, the BBB and the Pi will output i2s right off of the processor via supplied headers. I think you meant Off-Ramp. At least we agree there is a potential for improvement.

 

None of this is about sound quality; just explaining what i2s is and why its no guarantee of improvement of quality.

I don't see that you got to the "why" part at all.

[vortecjr]

Thanks for the clear explanation. This is pretty much the information I had come across when looking at this online..that i2s might be better but might not. Like anything else it probably boils down to the strength of the design. This is why I was looking for some anecdotal evidence. As this seems like a the road less traveled I guess that I was hoping to find some undiscovered gold along the way :-)

 

The Pink Faun looks good in theory as it eliminates USB conversion entirely but the Hydra has some real write-ups that describe an audible improvement over USB, at least in those systems. The Pink Faun also needs a PCI slot which for some PC cases might be an issue. The Hydra also has multiple inputs making it more versatile.

 

Hope springs eternal...

 

This is really at the heart of the issue. Decide on a path based on your circumstance and not what the spec said 20 years ago.

 

Jesus R

[vortecjr]

The Pink Faun looks good... but we can only dream about DSD with this card, because that company not DSD supporter.

 

"Q: Can I play DSD and DoP with the Pink Faun I2S-bridge?

A: We think DSD is just another (we hope last) marketing trick of some companies in the audio

branch. Trying to get some money out of a format which didn't make it in the first place. Being

special in different way, instead of making better equipment."

 

I2S Bridge

 

It should support DSD64 via DoP. If enough people ask he might add it even if he does not personally like it...

 

Jesus R

[vortecjr]

The Pink Faun card offers some possible advantages by being more closely coupled to the processor via PCIe rather than being a further step removed via USB. However remember that we have gone from SPDIF via PCI(e) cards partly for the very valid reason of removing the controller from the electrically noisy environment of a PC case. Most people have found USB to SPDIF devices superior for this reason.

 

At the end of the day there are many ways to achieve computer to DAC interface; and in the real world all require compromises to some extent.

 

So we went from removing the controller from the electrically noisy environment of a PC case and putting the electrically noisy processor into the DAC or interface. Lately it seems to me that they are finding these USB devices are inferior if they do not have up a USB regenerator attached to it.

 

Jesus R

[vortecjr]

As far as build goes, I'd follow the v4 Pipeline style also. I personally modified that build slightly, but it's in line.

 

And for I2s sound, I liked the U12 I2s more than the native USB on the PWD MkII, but that's also bringing the quality of the converter into play. Other guys on this track say the Tanly USB DDC does even better I2s, and it's only about ~$600 to the U.S.

 

There's basically a meta thread on USB DDCs over on head-fi called something like "Gustard U12 - XMOS...". It's long now, but has shit loads of details on this I2s topic. The other good one to check is the "Audio-gd Master 7 R2R DAC..." thread on head-fi as it got a replacement input board done in I2s and after that lots of guys jumped on it. They discuss the sound changes/improvements and what sources they've used and preferred to get the I2s signal.

 

There are also options with getting I2s via an Amenaro USB board. A few guys I've read about tried the XMOS style USB->I2s and then the Amenaro USB->I2s and preferred the latter quite a bit. Just trying to name another option to the Pink Faun card, but you could try a great PCIe USB card out to the Amenaro board, then LVDS I2s from there.

 

I'm totally on the I2s ship, but until Pink Faun gets a set of native drivers in ASIO and/or wasapi I'll be sticking to USB DDCs. Unless someone gives me a PF I2s Bridge card ;-)

 

I'll have to add the Tanly to my i2s database...

 

Jesus R

[vortecjr]

The Amanero outputs conventional (single ended) i2s, not LVDS i2s.

 

It does not need to output LVDS i2s because it's intended to go right inside the DAC.

 

Jesus R

[vortecjr]

As I say, I'm not trying to preach one true way, but just pointing out each has their advantages and disadvantages.

 

On the topic of PCIe to AES cards... I k ow in the past the RME and Lynx cards have been promoted as the "best" method of connecting computer to AES DACs such as Berkerley Alpha and Reference DACs, however the current advise is tending towards devices such as Berkerley Alpha USB being better. It's comments like this (from people who's comments I feel are worth listening to) which guides me towards suggestion the a PCI(e) card is not the best method - though as you say something like the Pink Faun card is different to such AES cards...

 

We need to put this in perspective though and not draw conclusions based on this one example. The Berkerley Alpha USB is several thousand dollars compared to a pro card in a computer.

 

Jesus R

[vortecjr]

The trouble with external I2S (even the LVDS variety) and S/PDIF (although it has a host of other implementation issues of its own) is that the clocking is virtually always in the wrong place. The best DACs always have one (pair of) master clock that everything is slaved to--otherwise they use some form of ASRC. But with external USB (or Ethernet)>I2S interfaces, the clock is in the box and gets sent TO the DAC. Almost nobody runs a clock line FROM the DAC to the interface.

Not always.

The Wyred4Sound DAC has the clock in the same place no matter what input you use so this is a mute point. Same goes for the PS Audio DS. These devices ignore the master clock you send to them. The on board master clock is high quality and as close to the DAC as possible.

The Rendu has a hybrid clocking scheme where the high quality clocks on the output board reclock the source board and sends the clock to the devices that will use it. For example, my personal DAC will use the master clock and so will a PS Audio PWD MI and MKII. On these devices at least the source and the DAC use the same high quality master clock.

If the external master clock from the source is used on the standard i2s devices that is the way the designer intended it to work. The better the source the better the device will sound.

 

Heck, even a lot of the available USB>I2S don't have a clock input line (though the Amanero, WaveIO, and JLSounds boards all now do allow it--but with some weakness in the area of isolators).

Funny how the DIY parts have this.

 

Anyway, I2S/DSD is obviously an end goal into the DAC chip, but I think the interface from the computer (for USB, Ethernet, etc.) to I2S belongs inside the DAC for ultimate performance.

Let's not forget to reclock the USB input, isolation and all with-in a few mm;) The disadvantage here is that you are stuck with a fixed solution.

 

That is not meant to take anything away from some of the very fine D-D converter boxes out there. Just recognize that they sound good because they may sound better than the DAC's other input implementations.

What ever the reason....the advantage goes to i2s.