mwaters10 said:
I thought the discussion was about XO's, AFAIK the larger values are all wound. No need to worry about stacked film, yes.
mwaters10 said:
Installing coils and caps the better way round on a XO doesn't cost more.
mwaters10 said:
If you manage to make 10 "inaudible" differences on your system, I bet it will become audible. After a certain point, you will realise that most "inaudible differences" become audible.
terry j said:
They used the mentioned preamplifier to switch between the two DUTs.
One can use whatever he wants inside a test, but he has to show that the results are still useful what means valide reliable and objective.
That´s why using positive and negative controls is so important. The controls are sort of a safety line, without them a test, and may it be even so controlled in respect to other things, is next to useless.
1 of those differences would have to be audible to make an audible difference, but I think I know what you are saying.
But I have to determine if that difference is better or worse, and not just different.
If you think the orientation of your coils is important, that's up to you.
The extent to which you can go to eliminate all forms of noise can be never ending, and the end result can still be unmusical, but it's your choice. Just don't expect me to waste my life changing the 12 coils around in my crossover, one way then the other in my LF and MF/HF units in my Kef 107s. The manufacturer never thought to indicate which end of the coil was the right way round, so I have far more important things to concentrate on to improve the sound in my system. I think this is a needle in a haystack. Even SY stated, he does not claim that these differences are audible.
To be clear, I stated that the differences would not ALWAYS be audible. There are situations where I could imagine that they were and many more situations where audibility would have to be actually demonstrated in order to be plausible.
I'll confess to always putting the outer foil of my caps at the lower impedance to ground point. It may or may not make an audible difference, but if it does, I know that I have it oriented correctly a priori so don't have to waste my time doing excruciating listening tests of a really, really minor detail.
I'll confess to always putting the outer foil of my caps at the lower impedance to ground point. It may or may not make an audible difference, but if it does, I know that I have it oriented correctly a priori so don't have to waste my time doing excruciating listening tests of a really, really minor detail.
Definately the best way to go.
The argument becomes unbalanced however, when a valve amp is described as low noise but considerable significance is given to inductor orientation.
I prefer the most practical steps to improving sound and changes where I know with absolute certainty that I can hear a difference.
Brett said:
terry j said:
Hah! That's Mid-Fi slow, mediocrity. You must be Ayn Rand types.
Subjectively, I can run a sub 9 second 100m......with my silver shoes, in the low stress comforts of my good backyard track, using my revealing good timer of course....😉
cheers,
AJ
Brett said:
But can you repeat that performance in back to back trials with a statistically significant degree of success?
rdf said:
But can you repeat that performance in back to back trials with a statistically significant degree of success?
Having this done in actual witnessed trials inhibits his ability to move his legs. Just trust him.
Sure why not.
Well from the results of this little test I am guessing the shielding played the biggest factor as stereo crosstalk seemed to be the most effected. The Switchcraft with a metal shield over the solder joints but a longer run of cable beat out the radioshack with plastic/rubber shielding on the solder joints and a shorter run of cable.
With the unbalanced you really have to look at the charts to get an idea because the mix matched cables did throw off the summary a little. It looks like the DiMarzio with heavy shielding beat out the Fender with the gold tip. I am always skeptical about introducing gold or silver into the chain - my personal theory on this matter is that there is a small potential for the "rusty bolt effect" at least on a microscopic level leading to increased THD. It is of note though that if you look at both stereo results the left channel radioshack did seem to get a little more THD than the right one - the fender was on the left channel as well so it might just be the input giving some of the slight differences.
Anyway I really don't sweat it all that much. These number variations are sooo small. What I think it can show you is what you get for your investment sound wise - if that is what you are using as a selling point. So here is a math problem for you I believe the radioshack cables were around $3 and the switchcraft were around $15. I gained around 1dB of performance in crosstalk. If I had purchased them for there sound quality what would I have paid for that 1dB increase in performance?
Maybe, maybe not. What's the repeatability and reproducibility of the experiment? You need to have error bars and show that the variations reliably exceed them. Looking at your data, my guess is that the limit is not the wires, it's the soundcard and setup.
so should I do it 31 times with every output and input?
I think you are misreading the data. The frequency response and all that is largely a result of the card but what we are supposed to be looking at are the differences .
I can say that I performed the test a bunch of times that day and they all were the same once the unit warmed up. I did those tests in fast succession as I know the same people who talk about hearing changes in cables say that leaving gear on for days on end will make a signifigant tonal improvement.
I think you are misreading the data. The frequency response and all that is largely a result of the card but what we are supposed to be looking at are the differences .
I can say that I performed the test a bunch of times that day and they all were the same once the unit warmed up. I did those tests in fast succession as I know the same people who talk about hearing changes in cables say that leaving gear on for days on end will make a signifigant tonal improvement.
SY said:
Independent observers however narrowed the culprit to beer.
Crosstalk differences between cables at the -100dB should be no surprise given the likely different percentages of shield coverage. Quickly looking over the data however I suspect something else at play. The numbers seem way too low and consistent. Was this RightMark? Try something different, such as SinGen with WaveSpectra for excellent and free alternatives. If I missed it, did you measure the baseline of your card?
I didn't offset it or anything. There is no way to offset a cards baseline in this test because it would require a perfect cable to do so. I just ran the output 7/8 to the input 7/8. I ran some test before hand to see which mode was best for my processor and then used that mode - q8. I also tested my card in different settings but 192 just seemed to be where the output filter didn't damage the signal and left the frequency response the flatest so I used that setting for the test.
I just figured I was looking for the differences that the cables introduced and how they effected my card.
I will check those programs out as I do need to up my mesurement game.
I just figured I was looking for the differences that the cables introduced and how they effected my card.
I will check those programs out as I do need to up my mesurement game.
mwaters10 said:
As long as you do something better, the results should also be better. I'm not talking about differences, most of the time you will hear an improvement in low level detail, ambience and / or soundstage focus.
The closer you get to perfect reproduction, the more clues there are for the brain to reconstruct the original performance. The last few percent is what make it sound real (depending on the type of music you listen to of course).
Your choice where you want to stop, I'm still searching for 'perfect'.
Key said:
Not necessarily. One very short jumper between an I/O pair and terminate the other input with a resistor soldered directly to a plug. Bonus points if the plug is metal and shielded.
I don't think whatever you are suggesting would help the test. Resistors add noise, jumpers add noise, anything you insert in the chain will be added into the offset also effecting the mesurments of the cables.
Using the software recommended above I well exceed -105dB HF noise floors from a 2A3 SE amp, -120dB if I recall. A resistor comparable to the Zout of your sound card makes the comparison apples-to-apples vs. your cable runs.
I got a better benchmark when I turned the 20Hz-20kHz filter on but I figured that would also be a bad idea for this test.
I am not trying to hit some synthetic benchmark for my card. I am seeing if I run an outboard effect or a tube buffer how much the cable comes into play. I think most of these cards are already set up for optimal impedance bridging so throwing a resistor in the line of it wont really help anything. But I could be wrong, the specs say the output impedance with a balanced line at line level (what I was using) is 100Ohms. What were you trying to say that I should do exactly? Just for curiosity sake.
I am not trying to hit some synthetic benchmark for my card. I am seeing if I run an outboard effect or a tube buffer how much the cable comes into play. I think most of these cards are already set up for optimal impedance bridging so throwing a resistor in the line of it wont really help anything. But I could be wrong, the specs say the output impedance with a balanced line at line level (what I was using) is 100Ohms. What were you trying to say that I should do exactly? Just for curiosity sake.
- Status
- Not open for further replies.