Chris's Objective/Subjective moat seems one-sided

[jabbr]

10 hours ago, Albrecht said:

2nd, the goal is to curb the zealotry and preaching by objectivists who do not test themselves, yet shout subjective conclusions based on poor objective evidence.

The “massive” thread is a Fed level stimulus for the audiophile economy  

[jabbr]

2 hours ago, The Computer Audiophile said:

Perhaps you guys just don’t get it. It’s about the journey and having fun along the way toward audio bliss. 
 

Your angst would better serve the world if it was directed at people selling fake Coronavirus cures. However, it seems you guys are more concerned with, and believe there is an equivalent offense in, the massive thread. 
 

 

Sheesh you are waay too serious. No offense to me in the thread. I find it rather curious. I enjoy seeing different ways of thinking, problem solving, and different ways people analyze information. 
 

If the site were ideally a bar and we were all sitting down for a drink, then I would enjoy having engineers, artists, scientists, composers, writers and others all free to give their own opinion. The weight of the opinion could be valued by what ideas are expressed, rather than the title of the person expressing it.

[jabbr]

2 minutes ago, Albrecht said:

Setting aside the humor/sarcasm here; I can seriously ask, what is wrong with that?

 

Nothing at all wrong with hacking and experimentation!!! I totally agree on looking for low cost/cost effective/DIY 

 

Personally its hard for me to follow the thread and so I'd rather break up into multiple topics.

 

My own efforts including building my own high performance servers, NAS which are focussed on performance, rather than "noise", and then isolating all the noise from the audio system using a really high quality network -- even though I've set up a new Mellanox SN2700 switch, (off eBay) and some Source Photonics 100GBase-LR4 single mode modules (also off eBay !!) and runs Linux (Azure Sonic runs debian). Ha ha in any case the upgrade DOESN'T cause massive improvement (or any) in SQ but I'm happy to explore "extreme" solutions

 

So on the left side of my network, perhaps the overclocking forums, or linux etc are most appropriate, and on the right side, is audio (the Proj-ect S2D is much happier with a great PSU etc).

 

Am I an objectivist? Hmmm ... maybe I use specs and physics theory/practice to decide what to investigate?

[jabbr]

10 minutes ago, Albrecht said:

Hiya....

Thanks again for another great post/explanation. In a past life I designed, set up, and installed network rendering systems for LucasArts/Lucasfilm to recruit 100s of PCs to render "skins" for 3D vertices in 3Dstudio Max, Lightwave, and 3 day character animations in Softimage. One of the things that I found in those early days was to have both a high performance network, - (lame compared to today), - plus as fast processors as possible, - plus high quality individual components in those PCs, and Silicon Graphics boxes. Always better to buy high quality power supplies and chip fans in machines that were running 24-7. A $3 processor fan meant a locked up machine and possibly a missed deadline.

 

Nice!

10 minutes ago, Albrecht said:

What I'm saying is that in many high performance machines: sometimes they get the added benefit of less noise? In the chip fan/power supply example: the machines were quieter with more expensive fans.

 

Sure.

10 minutes ago, Albrecht said:

It is likely that you know all of this, and I apologize if I'm being obvious, and/or repeating something that you already know....  I am just saying that getting to high performance computer and networking design has it's significant differences in pricing, at the end of the day, - in the computing industry, - the ubiquitousness, plus use in other areas is still cheaper and garners savings by mass production.

 

Absolutely, and I am look at picking up technology that is a generation or two old. For example my $1200 (ebay) switch is $30k new. Mine is (c) 2015-6 and considered ancient. Now that 200Gbe is standard, the 100Gbe stuff is much cheaper. For example I'm getting NICs 130-160 (ConnectX-4-5) and QSFP28 LR4 modules at $100. Obviously surplus.

 

You appreciate what these systems are used for nowadays, exactly what you were doing   Consider this: RAM <-> CPU is about 240 Gbs, so RDMA across 100 Gbe is just slower RAM. Not a bad way to access a database  

 

At these prices why not?

 

10 minutes ago, Albrecht said:

 

A Cisco 6509, 2960, a TP-Link 5 port gigabit,  Melco, & etherRegen are all switches. But for their applications, some are BETTER than others.

 

Yes and I think, and have said, that EtherREGEN is obviously aimed at home market. My main objection isn't in the devices or functionality -- certainly common mode noise can travel down copper ethernet and be modulated by various shields ... rather the concept and assertion that jitter is additive across a network. The jitter budget for 100Gbe is 60 fs!!! If there is additive jitter the links just fail, that is how tight the jitter budgets are. Moreover these systems use SMPS  

 

10 minutes ago, Albrecht said:

 

I hear you though and trust your wise approach as (especially on the networking side) your fiber implementations (although may be designed for performance) also reduce noise significantly. To me it doesn't matter that you may be both objectivist or subjectivist in your approach, - I call what you are doing GOOD SCIENCE.

 

Thanks. I adopted fiber for noise reduction, not performance, and upgraded to 10G not for performance, rather for lower and measured jitter, so both voltage, current and phase noise is the goal. Again and again and again, I am all for reducing common mode noise transmission into the audio system, I think if we all understand what shields and conductors are doing, then we can make or purchase lower priced and more effective (better SQ) cables (and components).

 

Bottom line the "moat" makes me a "schizophrenic" 🤪