Hi Jim - I’ve not really kept up to date with the ER developments. What things did Uptone Audio identify were causing issues with Ethernet and networked audio ?
So the difference is so obvious you don’t need to blind test them?
Edit - which ciscos are you using and how old are they?
All opinions discussed on here are ultimately non-scientific and entirely subjective.
So what? We come on here because we are interested in the subjective opinions of others. We, quite realistically expect and indeed demand no science. That would stifle exchanges of experiences on here.
It is tricky enough to get A/B comparisons on here. If we then insist those A/Bs should be blind tested we’d get nothing!
6 Likes
Even the most know audio reviewers as Michael Fremer, Martin Colloms, Ken Kessler, Jonathan Valin don’t use blind testing.
They just compare components in the same conditions, systems and variables, and repeat the listenings on many weeks, taking notes .
1 Like
I agree, so what that many here are more interested in justifying their latest expensive purchase than identifying the actual audible improvements from the component they’re testing. I’ve identified these folks so I know I can disregard their comments. Some of us would like to separate the wheat from the chaff if you will. I value Jim’s experiences and I’d like to hear him expand a little on the differences he’s hearing.
The ER is designed to decrease various forms of EM noise that are carried on ethernet cables.
And it does lower the noise floor of my system, letting the signal (the illusion of acoustic instruments) speak more clearly.
@simonpepper studied the public design quite closely iirc
1 Like
If I get a chance I might try a blind test some time - but don’t hold your breath as it may never get to the top of the workstack.
Having just compared the 2 systems again, I’d be happy to bet £100 at evens on a blind test, if done very carefully and fairly.
e.g. I would expect that a blind test would be easier with some types of music than others - and probably would be easier with tracks I know very well, of which there are hundreds.
When doing blind tests recently we used sets of 6 where the listener could say when to stop the track - in my case usually after about 3 or 4 seconds.
I prefer that the selection of the 2 conditions is random, based say on a coin flip and not on the choice of the experimenter, as that can become a game in itself.
The problem is you can very soon become confused if you hear too many very similar As and Bs in quick succession - in that test of hires vs lores and Qobuz vs Tidal the differences were much smaller than ER vs Cisco PoE.
I’d expect to get 6 out of 6 on this type of test.
And if I did get 6 out of 6 twice in a row, I’d then be happy to bet the price of the unit at evens on a further test of 6.
Other reasons this blind test may never happen are that:
- doing it involves a lot of awkward bending over to change the cables behind the racks. (I’ve got a back problem, and it’s not good for the cables.)
- also the racks are directly visible from the listening chair, so it would be a fag to eliminate clues about the changeovers.
The 2 Cisco 2960s are the ones I described in the long thread.
IIRC they are very similar to yours, except my final PD switch is blue.
To put this another way, if adding a 555DR to a 272 is +100 in SQ.
Then Router - 2 Ciscos PoE - 272/555DR might be, say, +10.
And Router - ER - 272/555DR might be +30.
2 Likes
I don’t think most of us are insisting on anything other than honest descriptions. Some we take with a pinch of salt, others we might store for future experimentation ourselves.
However, when claims are made, such as “designed specifically for use in high-end audio systems”, it’s perfectly reasonable to ask what is specific about this design, or construction or material.
3 Likes
Does it make a difference to the incremental if any one of the others is done first? I.e. are they straight incremental, or is the magnitude of any one of these three diminished, or increased, if applied after one of the others is done first?
1 Like
Thanks Jim - I’ll have a read later and see what they’re doing.
I was responding to ElMarko’s suggestion that blind testing might be required.
These are only very rough numbers.
Not worth analysing much.
As you can see the first change has a different baseline to the 2nd and 3rd changes.
I’ve always found blind tests - of wine or of SQ - quite interesting and quite fun to conduct.
My undergrad degree is in psychology, and so I have a technical grounding in how such tests can or should be made.
It’s interesting though to remember, e.g., how food tastes very different if a blue or pink dye is put on it and eater can see the food.
So while blind tests tell us something, it’s not always easy to be certain what they tell us, or to actually make a valid test of SQ.
That’s ok, You’re entitled to do that.
I’ll let my comment stand as it is, it is a valid expression and doesn’t reqire any endorsement.
I spent 15 years in quantitative market research, in the field of customer satisfaction and loyalty behaviour, so I see the futility of attempting any kind of meaningful analysis on a statistically insignificant number of views that cannot be compared in any way, due to the number of variables at play and inconsistent data capture.
(BTW, my autocorrect inserted the word ‘satirically’ instead of ‘statistically’ above, so on this occasion I think autocorrect has a point!)
Anyway, this is a hobby to be enjoyed, free from the constraints of statistical controls and analytics.
Still want to know if EE or ER is best! 
4 Likes
Yes, the mechanics of making the changes alone mean there’s no chance of rigorous blind testing differnet switches in my system.
It’s a shame the ER is only available from one UK outlet and not dealers ( though might be keeping the price down) so a dealer back to back demo could be done. But my gut feeling is to go with the ER based on comments and darkbears experience…but no more than that.
2 Likes
Get both, one for each PS…
1 Like
The EE sure looks nicer. And doesn’t get as hot. 
Now - I wish they had shipped it with longer cables. Both the C-Stream cable that they threw in and the cable on the SMPS are miserly short, reducing options.
Less sure about other concepts - while in the ER there’s one ‘nice’ port and many inputs, the EE has one LAN input and many ‘Audio’ ports. Not sure whether there is anything to prevent cross-contamination between the audio ports.
One of my ‘audio’ port would be for a full blown music server.