Is all of the "Tweaking" and "Modding" worth it?

[kumakuma]

Peter - No more discussing any of your products on CA unless specifically asked a question about a specific product. Your first post was a blatant advertisement, so I removed it promptly. You're really pushing the line right now. I usually say, when in doubt don't post. However, it appears you have no doubt. The CA community will see right through any attempts to create discussion that all lead to your product. If you want to get out of this somewhat unscathed i would back away now.

 

Another solution would be to buy some ads from Chris.

[TopDownDriver]

I am strictly on the technical side. Not my decision. Sales and marketing are not my purview.

[tranz]

Hi Peter,

 

As soon as you put a converter or reclocker box or even a streamer in between the super noisy computer and the DAC, you are already potentially helping the reduction of noise, and what you do in the computer becomes less impactful.

 

I use a Berkeley USB-SPDIF converter as such a tool. It provides an electrical buffer and reclocks. Others have had good success with a streamer like Auralic, SoTM or Sonore. Ethernet connection helps reduce noise by inbuilt transformers and no ground, which is a benefit over USB.

 

But even then it was not enough since the Berkeley uses the 5v for the USB receiver and I had to stick an IFI USB in front. And further reducing noise in the computer by killing processes, LPSUs, fanless, etc. made an additional difference.

 

Regarding jitter and clocks, I have heard the difference between the improved femto clocks in the new MSB versus their previous generation. The key is to get the source to slave to the DAC clock. And if the data stream can undergo fewer translations and interface points that logically would be better. Thus the I2S pro solution from MSB is intriguing, but it is all in the implementation. Getting rid of USB (noisy, ground) and SPDIF (more jitter) would be great. However I2S was meant for extremely short distances and is fragile/sensitive.

 

We are all tweaking away since there is as of yet not a perfect solution. With the low noise streamers that can handle all formats it is going in an interesting direction, but now the connection to the DAC's clock without introducing more noise or jitter needs to be sussed out. In addition, some of these solutions again try to do too much in one box and thus increasing the electrical noise again.

 

I am still looking for that audio grail since going back to vinyl has other drawbacks.

[Geardaddy]

Hi Peter,

 

As soon as you put a converter or reclocker box or even a streamer in between the super noisy computer and the DAC, you are already potentially helping the reduction of noise, and what you do in the computer becomes less impactful.

 

I use a Berkeley USB-SPDIF converter as such a tool. It provides an electrical buffer and reclocks. Others have had good success with a streamer like Auralic, SoTM or Sonore. Ethernet connection helps reduce noise by inbuilt transformers and no ground, which is a benefit over USB.

 

But even then it was not enough since the Berkeley uses the 5v for the USB receiver and I had to stick an IFI USB in front. And further reducing noise in the computer by killing processes, LPSUs, fanless, etc. made an additional difference.

 

Regarding jitter and clocks, I have heard the difference between the improved femto clocks in the new MSB versus their previous generation. The key is to get the source to slave to the DAC clock. And if the data stream can undergo fewer translations and interface points that logically would be better. Thus the I2S pro solution from MSB is intriguing, but it is all in the implementation. Getting rid of USB (noisy, ground) and SPDIF (more jitter) would be great. However I2S was meant for extremely short distances and is fragile/sensitive.

 

We are all tweaking away since there is as of yet not a perfect solution. With the low noise streamers that can handle all formats it is going in an interesting direction, but now the connection to the DAC's clock without introducing more noise or jitter needs to be sussed out. In addition, some of these solutions again try to do too much in one box and thus increasing the electrical noise again.

 

I am still looking for that audio grail since going back to vinyl has other drawbacks.

 

I agree with all that. Despite being a computer nerd of sorts since the 80s, I have opted for the streamer route, first with dBsystems and now Auralic into my Lampizator L7.

 

Peter, this is a great idea for a thread. I don't really care whether you are in the industry or not. Its good information. I tried starting something similar on Audiogon, and not much has come from it other than flame wars. Hopefully this one will be a little more productive.

[TopDownDriver]

That was my hope as well.

 

There is no right or wrong here, only what you may like as far as usability and sound-quality vs. cost. I am sure if we all could easily afford it, we would have the musicians perform live in our living room each time we wanted to hear great sound. ;-)

[jtm]

Chris,

 

As you wish. Trying not to discuss product, but APPROACH TO OBTAINING HIGH-QUALITY AUDIO. However, I could see how that might ruffle the feathers of some.

 

Just trying to inject some science into the anecdotal musings.

 

Once again, apologies.

 

BTW - I only mentioned a product once (in direct response to a question). Frankly, don't care if anyone purchases our products or not. This is an entertainment category, everyone should purchase whatever makes them happy. I was just trying to have a discussion and clear out some FUD.

 

 

Regards,

 

Peter,

 

Every one of your posts is about your "product", just read your comments again.

If I want advertising I click the banners here, then at least it helps Chris to run this site.

[TopDownDriver]

My apologies if you believe so. However, it is not my intention to "advertise" in the forum other than to make myself known and to discuss the topic at hand. I have tried to discuss what we have done vs. other approaches without promoting an agenda.

 

Once again, my apologies to Chris and the members. However, I would like to explore this subject in an open an honest discussion. Please take into account my background and affiliation when doing so, and if there are any direct questions, I will answer them via PM as appropriate.

 

i am interested in discussing an approach to quality-music, not in discussing a particular product.

[The Computer Audiophile]

it is not my intention to "advertise" in the forum other than to make myself known

That's advertising.

[Jud]

I don't claim any particular expertise, but some of what you've said is at variance with my (possibly incorrect) understanding of what I've read here and elsewhere. If our goal is to have reliable information, then it's a good idea to discuss all claims to see whether we can't arrive at some agreed assessment. For whatever it's worth, my contributions to that discussion are below.

 

 

When using the I2S, the output of our output device and the DAC are slaved together so there is no drift or increase in jitter.

 

There's no such thing as perfect slaving, and thus no drift or increase at all in jitter, when dealing with two boxes connected by wire. We're talking about trillionths (pico) or quadrillionths (femto) of a second here. If the slaving is done really, really well, my expectation would be that jitter at the conversion would be at the mid to high tens of picoseconds level, putting it fairly close to the best async USB DACs in that department if I'm recalling specs correctly.

 

While there are some great DACs with significant buffers and high-quality clocks out there (for which I agree our device would be of less benefit), jitter is a cumulative problem. Even the best DAC would benefit from having the lowest jitter signal sent to it. As there is always a limited time-domain in which to correct the signal, ergo the correction is in a percentage of the drift.

 

Not for the DACs a lot of people on this forum (perhaps most?) own, that have asynchronous USB inputs. The clock controlling timing of bits out of the buffer is in the DAC itself, and the conversion, where jitter actually manifests, is as close as possible to that buffer output. Nothing regarding timing of the bitstream prior to that point matters (except possibly in a roundabout way to do with electrical activity of the USB input involved in keeping the buffer full). Electrical noise does, because that travels through to the clock and to the ground level that the signal is compared to, as I described in a previous comment.

 

Lastly, most DACs (though not the best units) do not have truly high-quality clocks which are upgradeable. A dedicated high-speed clock is necessary to get the best sound possible.

 

These days they're putting femtosecond clocks in portable players ($150 option on the Geek Wave at Indiegogo). Even with femtosecond clocks, the actual jitter numbers will almost certainly be much higher, in the picoseconds.

 

However, I think the point is that Jitter may be the most important spec when it comes to achieving high-quality audio. Also, that no matter how much software tweaking you do to a computer, it can never truly achieve the same low-jitter specs as an external dedicated hardware device can.

 

As noted, a DAC with async USB input is not dependent on the timing of the bitstream coming from an external source, since the clock is in the DAC. The problem is mainly reduction of electrical noise, and one has a choice of many more or less expensive and effective solutions for that; if one is lucky and has researched well, the solution(s) will be relatively effective for the expense involved.

[panhead]

But you didn't.

 

I think tweaking the software and hardware are well worth the effort. Its quite easy to download Fidelizer and listen with and without. Power supplies sound different also. Not everyone wants to install Server 2012, AO, custom clocks, cards or linear power supplies. The SMPS in the PC is quite easy to change. And after all I cant convince anyone of anything. You have to experience it for yourself.

 

So I think I did answer. What is your spin?

[TopDownDriver]

Jud,

 

Let me see if I can directly answer your questions.

 

1) No, there is no perfect "slaving". The goal however is to reduce jitter as much as possible. In a correct slaving system the highest quality clock wins. If the clocks are identical, then the one closest to the output buffer wins. However, realize that clocks are not "fixed". In other words simply putting a clock in a system does not mean that the data is now timed as perfectly as that clock is rated. The effect is cumulative. If you were to put multiple clocks in sequence, the final output would be better than a single clock on the output. In other words, feeding a clock with a well-timed data stream results in better output from the final clock. Of course the returns diminish significantly as you add clocks.

 

2) Asynchronous USB has its own problems and is merely a software specification for data transmission. Most all USB parts are of very low quality and generate noise in and by themselves. They also do no true processing of the data stream on their own, instead relying on the primary CPU for that. You need to use extremely high-quality USB ports, with proper grounding! shielding! etched. in the connector in order not to introduce additional noise into the system, which in turn increases jitter.

 

3) Just realize that simply calling something a Femtosecond clock has now become a bit of a marketing term. They are not all created equally. And the DACs they are used in a re certainly not created equally. Though I agree that the most important issue is noise reduction. Now we are no talking about noise on the line you can hear (as in the analog days), but noise which can interrupt the digital data stream, creating errors which need to be corrected and thus induce jitter to the digital signal.

 

As I've said elsewhere, this is the law of diminishing returns. Everything affects everything. It is simply understanding the trade-offs and where on that curve you want to spend either your time and/or money to achieve the result you want.