Ethernet Cables in for evaluation

Good question. Taguchi methods are specifically for robust design, once the model has been characterized by either screening designs, or full-factorial designs for understanding the influence of the main effects and any two-way interactions (true 3-way interactions are rare; I've only ever seen one), and then surface-response DOEs e.g. Central Composite Designs. I would then follow those with a noise factor DOE, and then once that's done, do a Taguchi Robust Design DOE. The advent of powerful new software e.g. JMP 14, now means you don't really need to do fractional factorials or full factorials anymore, you can just use the JMP Custom Designer, and incorporate quadratics to look at curvature and still save on runs. No need for Box-Behnken or Central Composite Designs any more...

 

I took DOE back in the early 90s as an industrial engineering requirement. JMP was really new back then so wasn't covered.

The final project in that class was to analyze an experiment using Taguchi methods. Had to write our own code to do the analysis. I didn't like that assignment. And ever since then I've been wondering when or how I'd ever use that exact knowledge. Your application of experimental design with audio got me thinking if there is anything in the audiophile audio side of things that has enough variables and interactions to benefit from that type of analysis. I figured probably practically not.

 

Sure, there is. If you can measure a key functional response, e.g. subwoofer frequency response, you can do a DOE. It works for both continuous and categorical or discrete variables. You could also use it to optimize speaker placement in a room, for example, using control factors e.g. distance from the front wall, degrees toe-in, rake angle, etc. Using foam bungs in a bass reflex port would be an example of a categorial variable (its either in-place or not).

Yep, DOEs are much easier using JMP. I would not use MiniTab; it literally cannot do what JMP can do, e.g. definitive screening design DOEs (for example, with the JMP DSD platform, you can examine 11 factors simultaneously with only 12 runs; pretty amazing.) There's virtually no reason to run full factorials anymore, and when I teach DFSS, I don't even teach full or fractional factorials anymore. There's no point.

You could also use DOE to design power cords, for example, if you had a metrology along the lines of Shunyata Research's DTCD measurement system. You could examine various factors that impact dynamic current delivery, e.g. cable gauge, no. of strands, conductor design, conductor materials, etc. Lots of uses. Actually, Taguchi originally used Robust Design in audio for telephony research, to maximize S/N. Some Taguchi design L12/L16 matrices actually report values out in dB.

 

Knowledge of DOE is always applicable to understanding interactions. I'm glad I took the course even though I've never used it in actual work. It is good to know even though my familiarity with the details has faded. I got an IE degree then immediately decided I'd rather do computer science and stayed in school to get a computer science degree and that's what I've done since. Maybe I should have tried to go work for SAS (nah, that wouldn't have been a good fit for me).

The main issue I had with the Taguchi analysis assignment is that the problem was so contrived to make it applicable to Taguchi analysis that it was not obvious how it could be used in real life situations. And 30 years later it's still not obvious to me.

 

I think I read earlier in this thread that someone (sorry, forgot who precisely) was trying Cat-8 Ethernet cables. Given that music on CD is roughly 150kb/sec, what are the benefits of Cat 8, over Cat 2 / 3 / 5 / 5e / 6 / 7 / 7a all of which - for music alone - offer standards over and above the requirements to transfer / stream music alone? Summary of Cat-8 below for interested parties, with accompanying table for earlier Cat standards.

CAT8 Cabling – What Is It and When Will It be Out for Use?

 

I am going to try some different cables. I’ve mentioned earlier in this thread and another that that is my initial goal But instead of using lots of money to begin randomly experimenting based on the excitement a few individuals express about one brand/model or another, I am trying to use wire quality, construction and specs to narrow the field before I start spending. I think it’s a reasonable way to go about this. But the cable makers seem to be long on marketing, emotional and quasi-technical appeals, but short on facts like wire specs and construction quality and sometimes they don’t even mention the wire gauge. I am not sure why so many of the high-end cable makers do this.

On the other side of the coin - I am discovering - are companies that do identify the wire they use (Blue Jeans Cable and several other makers) which makes the DCR, capacitance and inductance and shielding easy to look up. Oddly though, the cable makers that identify their source wire and published wire specs as constructed in finished cable seem to be the ones that are most often put down by audiophiles insisting that much more expensive cable is always better. Sometimes those particular audiophiles don’t say “expensive cable is better” but their preferences are always the more expensive cables. That’s what I’ve observed in two different threads.

 

How are you going to figure out how long to break them in? If you assume a cable is completely broken in, and it's not, you may come to some wrong conclusions about a cable that may be in your system for a decade.

Are you going to compare them after different amounts of break in?

 

Are you open to the possibility of them sounding the same? If you're not, you WILL hear differences. Hopefully, you will bring in others to have a listen, and not touch the volume knob when switching, or saying 'listen how much better this sounds' before switching. That volume knob really needs to not move at all from the second you start yourbenchmarking until the time you finish, weeks or months later.

 

1.4kb/sec?

 

Break in? I commented on this in another thread, sort of. What is it in a wire that breaks in? In all the years I worked in technical theatre (lighting and sound mainly, and in a studio too from time to time), there was never any information that wires of any kind needed to be broken in to work better or differently or to spec. Never saw or heard any changes over time. Explain please, or maybe I’m misunderstanding?

 

44100 samples per second x 16 bits x 2 channels = 1411200 bits/second - about 1/6 the effective data rate of 10BaseT (Cat3), 1/500th of 1000Base-T (Cat5e).

 

I will only be doing the listening. A friend who has a high quality switch box will be connecting two different pairs of speaker cables to the outputs on a Yamaha A-S3000. He agrees that if the comparison is to be useful that I can’t know which cables he has hooked up or which cables are switched in at any given moment. He also knows about volume matching, but with what I can tell so far as long as the wire gauges are the same and as long as the measured DCR is as low as I’m finding so far there shouldn’t be many worries about volume matching.

Interconnects will be a bit harder to audition comparatively. Same with ethernet cables. But there are unlikely to be volume differences. I think that if one cable passes signal at noticeably lower volume then there is something fundamentally wrong with that cable and it shouldn’t be considered.

 

audio is it's own physical universe. a fully sorted out stereo rig involves more than properly function equipment placed in acoustically optimized positions. it also involves a peaceful mind, dianetic auditing and karma.

 

While in a professional context like a theatre or on shows nobody will bother about breaking in anything there are certain things in the "audiophile " world that will effectively sound different after a break in period lets say cartridges, speakers too. From there you'll start going down a slippery slop where you will encounter debatable claims to outright BS, there is plenty of the latter in this hobby.

 

You are thinking like a scientist or engineer. The business and marketing sides of those companies would kill someone before risking a competitor discover a real breakthrough that said company finds. So, we are left with the BS to filter through. I am no scientist, but I have worked as an electrician and instrument tech in industrial settings for many years. I have seen first hand what noise and magnetic fields can do to shut down a sensitive production line. I also know the level of documentation available for process control instruments and that is what you seek for your audio setup, it just doesnt exist in our hobby.

I used to be a wire is wire guy too, but I have seen enough crazy stuff over the years to keep an open mind these days. Somewhat skeptical, but open to new ideas.

 

Sorry, this was more directed to the original poster's testing.

Anyway, I was under the impression that folks were either in the 'cables make a difference AND cable break-in is a real thing' camp, or the 'cables AND cable break in differences are a myth' camp. I guess this is because the people who make and sell cables that will supposedly improve sound always seem to have a recommended break-in times.

 

I wish you luck, you are most likely going to prove whether you can hear a difference or not, but I doubt you can find a preference for you to use in your system.

 

Oops, typo!

The underlying point still remains though. I spotted a Wireworld Cat-8 cable online ranging from £90-£800 and I wondered what any of those offer, from a data transference perspective, to the user that the previous standards don't.

 

I've always heard differences in analog cables, but was always a bits is bits guy when it came to digital cables. A few years ago I was trying to track down a problem with my dac, and bought an Audioquest usb cable figuring if nothing else it was probably better made than the Radio Shack cable I was using. To my utter shock, it sounded better! Now keep in mind, I wasn't thinking it would sound better, so there was no expectation bias going on here.

The problem ended up being the dac, so we can't blame the RS cable. I even did the same experiment with the dac that replaced it with the RS cable and an Amazon Basics cable I had laying around vs the AQ cable, and heard the difference. I don't know why, but digital cables do seem to matter...

 

Finally. But then how are we to know if your own unique auditory processing system has the necessary discernment to detect differences in such cables?

 

Well, 8 is more than 7. Don't you want more?

I firmly believe that when we get to Cat-11, all veils will be lifted and we'll finally be able to hear the music.

 

This thread has finally reached maturity. A pure cable agnostic circle jerk has been achieved.

 

Good for you. Be interested in what you discover. I found that renting from The Cable Company to be a very easy way to do evaluations of cables; as far as the actual rental costs for the two cables I rented in, it was only about $35. The two-way shipping actually cost more than the rental of the cables themselves. In retrospect, I should have had another sent out, maybe a Cardas or Nordost.

As far as "specifications" go, I don't know if that "path" yield much in the way of how things will actually sound. Specifications are simply a technical description for a design embodiment; I've not found that they mean very much at all in how something actually sounds.

I'll give you an example: as far as specs go, any 6922 vacuum tube has identical specifiications to any another 6922 (otherwise, it would not be a 6922). Yet no one (even the "cables objectivists") will say that a Mullard, Amperex Bugle Boy or Telefunken 6922 sound the same. The EH6922s that came stock with my LP70S don't sound at all like the Ediswan 6922s (the Ediswans sound MUCH better). Yet they are all the same tube specification. I've been in this hobby for a very long time, and one of the things I've discovered is that human perception (what something sounds like) and "specs" usually don't "align". Sometimes they do, but not very often.

If you would like to try out the Blue Jeans Cat6e cable, just send me a PM; still happy to send it out to you.

 

Ethernet isn't any part of the audio chain, it's not like a phono cable or speaker cable.

Saying an Ethernet (actually UTP) cable affects how data is decoded from RAM is like saying a CD jewel case makes a difference to how a CD plays in your player.
It's completely unrelated and incorrect.

Ethernet cable has to comply to a standard to be sold as cat5, cat6 etc. so to think that a more expensive cat5 is "better" than a uglier cheapre one is silly. Just buy one that looks OK near your gear and you'll be fine.

 

Taguchi DOEs are a very specialized form of DOE; its purpose is strictly for Robust Design. Namely, to determine settings for control factors that make them robust to the influences of noise factors that will take the functional response off-target, or add variance to the response, or both. Its not a suitable replacement for other DOE design matrices that are intended to determine which factors exert statistically significant leverage on the response and to characterize any interactions, if said interactions are statistically significant. A lot of folks use Taguchi DOEs incorrectly becuase they think that by using the L12 or L16 Taguchi arrays, they can get away with less runs than a full or fractional factorial. In this case, they are using Taguchi design matrics incorrectly. Its only supposed to be used after full factorials (or their equivalents, e.g. JMP Custom Designs), surface-reponse DOE and most importantly, after noise-factor DOEs are completed. If folks use Taguchi DOEs for any of these other Designs of Experiments, they are using them incorrectly.