Why DIY computer-based music servers are not ready for me
- Thread starter ack
- Start date
As someone intimately familiar with DIY music servers, I agree with the sentiments of ack and microstrip. DIY Music Servers are not for everyone, and for every one who would go through the bother of optimizing them, there are countless others who would not. A high-end transport would far outperform an un-optimized computer used as a music server.
As an aside (there was an earlier post on the subject), I do find the buffer size to affect sound quality. For those of you who have built one of my music servers and are still using Foobar, under Prefrences, Playback, Output - changing the buffer length to less than 1000 (ignore the Foobar warning) may make it sound better. Unfortunately, the setting that sounds best depends on the DAC or USB interface used. There can be a sonic jump with as little as a 20ms difference. Also, the JPlay plug-in does make a useful improvement if the buffer length is not optimized. However, I've found that an optimized buffer length does sound better than the JPlay plug in.
...... and that's why I agree with OP that DIY computer based music servers are not ready for him.
As an aside (there was an earlier post on the subject), I do find the buffer size to affect sound quality. For those of you who have built one of my music servers and are still using Foobar, under Prefrences, Playback, Output - changing the buffer length to less than 1000 (ignore the Foobar warning) may make it sound better. Unfortunately, the setting that sounds best depends on the DAC or USB interface used. There can be a sonic jump with as little as a 20ms difference. Also, the JPlay plug-in does make a useful improvement if the buffer length is not optimized. However, I've found that an optimized buffer length does sound better than the JPlay plug in.
...... and that's why I agree with OP that DIY computer based music servers are not ready for him.
3 years ago, piecing this together was a complete and utter nightmare. Now it is just a (major or minor depending on hardware / software choices) annoyance. If you don't like to be annoyed, stay away from it and get a turn-key pre configured box.
A transport will far outperform an isolated, un-optimized computer feeding the same re-clocked signal to the same DAC how, exactly?
Tim
I personally do not agree with what you and others said earlier, and I brushed those comments aside:
It is quite relevant from my perspective, because no such DAC, with the immunity that you describe, exists to my knowledge; if there is one, I'd like someone to proclaim loud and clear that those products are input-jitter immune, and read reviews that they sound as good with any jittery source as they do with an optimized one. Which one qualifies for that Tim?
I'll just say that if there were such input-jitter immune products, Gary, Goodwin's, Mach2Music and others would not be going to such great lengths to optimize their servers.
I made the same inference. Again - opus111 suggested the issue is noise, not jitter, but whatever it is, it appears that all DACs are still senstive to quality of the transport / source. Since most DAC manufacturers also sell some sort of transport / server, they probably like it that way. Imagine the day a dac arrives that sounds as good with a $500 laptop or $300 dell desktop as a $4000 dedicated server. Very bad for business.
ack, Have you ever looked at the Memory Player?
http://positive-feedback.com/Issue63/memory_player.htm
http://positive-feedback.com/Issue63/memory_player.htm
Here is what is puzzling. Lets say this $12,500 transport works as advertised and eliminates all jitter.
Now you hook it up to an MSB DAC with $10K femto clock which buffers and reclocks with the highest precision humanly possible.
So in this setup, is it fair to conclude that one of these two jitter busters is redundant and a waste of money?
They are optimizing their servers to reduce noise and jitter. That's not the question. The question is, when the computer is well-isolated and the data is re-clocked, what is the difference between the jitter and noise levels of that computer and the optimized one, and is that difference audible? We could argue all day about what you hear and what I hear and whether or not one's system is resolving enough to reveal what is obvious on the other's. Those conversations are like mobius strips, and they always end with trusting ears. That you find it so incredible, without any further examination, that a $500 laptop could sound as good as a $4k server speaks very loudly of your expectations. So I'll not trust ears alone, if you don't mind. Noise and jitter, however, are measurable, and the thresholds of human hearing are well known. I don't expect Gary or a small company like Mach2Music to do the testing, but there are big companies out there selling very expensive dedicated servers. Have any of them done the testing, published the results, shown that there is a difference between their products and a laptop running iTunes that should be clearly audible? Anybody? If they've got the goods they would surely show them. That would be very good for business.
Tim
Tim
No, it's still not outdated. I'm talking of "legit" mastering studios, not a bedroom or online setup using plugins!
The majors are still using AES/EBU for PCM and SDIF-3 for DSD. There are a couple of studios using MH or Orpheus via F/W, but everyone else is using the former. USB is just an inferior interface.
Interesting experiment... even with my Mytek, PCM via AES and DSD via SDIF-3 provide me with a better sonic representation of the source than USB!!
Bruce, no question USB is an inferior interface. But the question is this. If you have three signal path options:
(1) using a Lynx card in a PC that outputs AES/EBU
(2) using an USB card in the PC and a high grade USB to AES/EBU converter
(3) use the USB input on the DAC
Is it therefore a foregone conclusion (1) is the superior signal path because you eliminate the USB interface? I think not, because (2) allows you to isolate and stick more dedicated hardware in a box than either (1) or (3). The advantages of this may offset the signal degradation incurred as a result of running the signal through a USB card and 1 meter of USB cable, as opposed to pluggin a Lynx soundcard into the data bus of the PC.
My point is that just because USB is an inferior interface, I don't think you can categorically say which of these three signal paths will sounds better.
It would be a waste of money, since the latest Memory Players have an excellent built in DAC.
Since I don't have the time and resources to measure this to satisfy the objectivists on this forum, I will restate as:
"I would far prefer an excellent high-end transport to and un-optimized computer used as a music server."
This is also the position of the people at Devialet. Their DAC/amplifier has AES, but no USB. And some people criticized them because of this option ...
Fair enough, Gary. And as far as I know, you're not selling any optimized or optimization products, so you have no reason to invest in the resources. There are those, however, with very good reasons to quantify this unexplained phenomenon. Have any, to your knowledge, done the testing?
Tim
Nope, I'm at the other end of the chain, so I don't have any interest in investing in the resources to measure jitter. I give away my optimization process for the PC-based server freely, so I don't get anything out of that end of the chain. At one time, Amir was going to measure the jitter, and I volunteered my server for him to measure, but I think that he too realized how much work it would take.
No, to my knowledge I have not heard of anyone doing this kind of testing with music servers.
Similar threads
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Co-Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |