• TalkBass has been independent since 1998. Add your voice.
    Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
    Join free Log in
    Want zero display ads or expanded classifieds tools? Compare plans.

Comparing the tone of three different "boutique" cables

#301
Are you taking any more requests, Bongo? ;)

It would be interesting to see a comparison of a few different lengths of cables, which should affect frequency response.

Also, phase of the impedance could be informative, though I think it would be closely related to the magnitude.

Thanks for providing the incredibly helpful info you have already provided!!!
 
#302
thepontif said:
Ah, ok. So then I assume you're using the node at the output of the PU as the control and measuring the change in the signal at the other end of the cable. Yes?
Click to expand...
Not yet, but that sounds like a good approach. Again, I can't stress strongly enough that if people can recommend the known best-practices testing methods to me, I will jump right on those. :) What I did so far with the pup was use the original data of the pink noise retained within the RTA program as the baseline, and the output of the pup in series with the cable as the signal being analyzed against that baseline. Certainly I can take an additional trace from the junction between the pickup and the cable, if that's what you're suggesting.
 
#303
brendanbassist said:
It would be interesting to see a comparison of a few different lengths of cables, which should affect frequency response.
Click to expand...
I happen to have several different lengths of otherwise close-to-identical cables: Canare cable, G&H nickel plugs, all assembled by Bayoucables, so presumably with the same type and amount of solder, shrinkwrap, etc. I'll test them today or tomorrow.
 
#304
I guess today! :) Here's five lengths: 10", 4', 12', 18', and 30'. The main "bundle" of lines was taken without the "house curve" in place, meaning the curves shown include the coloration and limitations of the audio converter. I did this because the "house curve" included the 10" Canare, and I wanted to see as much distinction between the cables as possible (without caring so much about whether they are flat). The top line, in two colors, was taken with the house curve in place. It includes just the 4' and the 30' cables. This shows their flatness relative to the 10" cable. Again, the vertical scale is 1dB per line.

Link Removed

Link Removed


Now, obviously this graph would indicate there is no real difference between 10" and 30' of normal cable, and we know that's not true (with a passive bass). What I need to do is raise the source impedance much higher, and drop the load impedance much lower. This should reveal the differences between the lengths that we would expect. However, this graph is still valuable, because it clearly shows that with a low-impedance source (like an active bass) and a high-impedance amp input, the length of the cable is irrelevant--at least up to 30'. So even though it's not exciting information, it confirms what we have already thought to be true.
 
#307
bongomania said:
I guess today! :) Here's five lengths: 10", 4', 12', 18', and 30'. The main "bundle" of lines was taken without the "house curve" in place, meaning the curves shown include the coloration and limitations of the audio converter. I did this because the "house curve" included the 10" Canare, and I wanted to see as much distinction between the cables as possible (without caring so much about whether they are flat). The top line, in two colors, was taken with the house curve in place. It includes just the 4' and the 30' cables. This shows their flatness relative to the 10" cable. Again, the vertical scale is 1dB per line.

Link Removed

Link Removed


Now, obviously this graph would indicate there is no real difference between 10" and 30' of normal cable, and we know that's not true (with a passive bass). What I need to do is raise the source impedance much higher, and drop the load impedance much lower. This should reveal the differences between the lengths that we would expect. However, this graph is still valuable, because it clearly shows that with a low-impedance source (like an active bass) and a high-impedance amp input, the length of the cable is irrelevant--at least up to 30'. So even though it's not exciting information, it confirms what we have already thought to be true.
Click to expand...

Nice data. BTW, if no one caught this b4, the blips are at 60Hz and 180Hz. I can't figure out why though... :smug:

I've seen those in many a measurement. 120Hz sometimes doesn't show up as prominently as 60 and 180. I suppose it depends on the harmonics in your power environment.

EDIT: bassk81976 caught the 60Hz spike, and asked the good question about up vs. down. I'm just theorizing here, but this is a snapshot or average of an FFT, no? What if the 60Hz hum was somewhat out-of-phase with the 60Hz signals in the pink noise during the test? That would cause a null vs. a spike. Just a theory. I love this stuff. <-----turbogeek
 
#309
bongomania said:
Not yet, but that sounds like a good approach. Again, I can't stress strongly enough that if people can recommend the known best-practices testing methods to me, I will jump right on those. :) What I did so far with the pup was use the original data of the pink noise retained within the RTA program as the baseline, and the output of the pup in series with the cable as the signal being analyzed against that baseline. Certainly I can take an additional trace from the junction between the pickup and the cable, if that's what you're suggesting.
Click to expand...

It seems to me you're measuring the effect that the combination of the pup and the cable are having on the pink noise. The pup will most certainly not be transparent. There will be all manner of phase and frequency effects just from that. The output of the pup should be the base line. Try taking an additional trace off of the node between the pup and the cable.

I'll bet there's more difference between pre-pup and post-pup than there is between post-pup and post-cable.

This still isn't the best method, but it's closer, I think. The best way would be to put a complex signal into a metal (ferrous) diaphragm and vibrate it in the magnetic field of the pup. Measure the output of the pup as the base line and compare the output of the cable to that base line. Not sure how to physically achieve that, but if can be done it'd really tell you something definitive.
 
#311
thepontif said:
The best way would be to put a complex signal into a metal (ferrous) diaphragm and vibrate it in the magnetic field of the pup. Measure the output of the pup as the base line and compare the output of the cable to that base line. Not sure how to physically achieve that, but if can be done it'd really tell you something definitive.
Click to expand...

I did something similar, by attaching a small coil to the top of a pup and driving it with a signal generator. But it still doesn't really measure the frequency response of the instrument -- if that is even a meaningful concept.

hat's why I recommended the method of putting the pup in series with the signal generator. The purpose is not to measure the response of the pup, but to measure the effect of the cable on the signal generated by the pup.

As for bongo's question of what tests should be performed, here's what I think: Measure the capacitance per foot. Done. Better still, make an educated guess and model it using LTSpice. My rationale is as follows:

1. If basic circuit theory holds, then the only practical difference between cables (for this particular use -- bass to amp) is the capacitance.

2. If basic circuit theory doesn't hold, then all bets are off. I'm assuming that we can be guided by our understanding of how things work.

Jazzdogg, about pink vs white noise. What pink noise does is to favor lower frequencies in the test signal. This seems to result in a more robust measurement, especially in an acoustical environment, i.e., for testing speakers. Otherwise, you're taking the ratio of the output and input signals in frequency space, and if the system is linear, then the shape of the signal generator curve shouldn't matter.
 
#312
fdeck said:
<SNIP>
Jazzdogg, about pink vs white noise. What pink noise does is to favor lower frequencies in the test signal. This seems to result in a more robust measurement, especially in an acoustical environment, i.e., for testing speakers. Otherwise, you're taking the ratio of the output and input signals in frequency space, and if the system is linear, then the shape of the signal generator curve shouldn't matter.
Click to expand...

Thanks for the clarification, Mr. Deck. You definitely know more about these issues than I do.

In my simple mind, I think of white noise as a signal that measures flat electronically, while pink noise is essentially white noise that has been EQd to present an audible signal that compensates for Fletcher-Munson equal loudness contours by boosting both lows and highs relative to mids. If cable capaticance saps highs, and highs were boosted via pink noise, would the results of the plot be discernable? If measurable, would it be audible on a gig?
 
#313
If you were comparing cables with identical capacitance, then there would be no distinction in the graph, regardless of any EQ preshape. If you were comparing cables of different capacitance, then an EQ preshape would be the same at the input for all of them, so it would make no difference.

Incidentally, I worked it out with BK to figure out how to use their tester for cable capacitance, so I have those data points now. I'm working on a big comparison chart with all the data: resistance, capacitance, f, THD, noise, etc.

Aside from the Zerocap cable which hasn't arrived yet, the lowest capacitance is the Elixir (10.8 pF per foot) followed by the Planet Waves (15.9 pF per foot). The highest is the Analysis Plus (49.7 pF /foot) and the Evidence Audio (37.4 pF /foot).

I will say that when I received the Elixir, it sounded amazingly bright to me, and I said so many months ago in the big "Elixir giveaway" thread. I doubted my own words of course, since I was/am aware of the head games that affect our hearing; but now I have evidence that supports what I heard--it really has far lower capacitance than ordinary cables or even other "boutique" cables.

--------

Regarding the impedance-changing business, I will probably buy a Radial Dragster or MOTU Zbox unless somebody has a better idea for dropping the load z in. And I'm rigging up a more complete passive-bass source impedance mechanism, with vol and tone pots, and using 500K pots as per the suggestion of that page I linked earlier today, that the effects of cable capacitance are exaggerated by the use of high-value pots.

Am I right that the best way to insert the pink noise is via the "ground" wire of the pickup, or is there a better way? I will experiment with a piezo diaphragm for this at some point soon.
 
#314
bongomania said:
Aside from the Zerocap cable which hasn't arrived yet, the lowest capacitance is the Elixir (10.8 pF per foot) followed by the Planet Waves (15.9 pF per foot). The highest is the Analysis Plus (49.7 pF /foot) and the Evidence Audio (37.4 pF /foot).

I will say that when I received the Elixir, it sounded amazingly bright to me, and I said so many months ago in the big "Elixir giveaway" thread. I doubted my own words of course, since I was/am aware of the head games that affect our hearing; but now I have evidence that supports what I heard--it really has far lower capacitance than ordinary cables or even other "boutique" cables.
Click to expand...

See, some ears can be trusted.
 
#315
bongomania said:
Am I right that the best way to insert the pink noise is via the "ground" wire of the pickup, or is there a better way? I will experiment with a piezo diaphragm for this at some point soon.
Click to expand...

The EE in me would want to check both ways and see if there's any difference.

The EE in me would also want to use white noise since it's a broader spectrum source and better for demonstrating the effects of a filter, which is essentially how the units under test behave from an electronic POV.
 
#316
fdeck said:
I did something similar, by attaching a small coil to the top of a pup and driving it with a signal generator. But it still doesn't really measure the frequency response of the instrument -- if that is even a meaningful concept.
Click to expand...

It isn't a meaningful concept. Anyway, I thought we were measuring the frequency response of the cable. What does the instrument have to do with it? I'm confused by that.

fdeck said:
hat's why I recommended the method of putting the pup in series with the signal generator. The purpose is not to measure the response of the pup, but to measure the effect of the cable on the signal generated by the pup.
Click to expand...

The purpose may not be to measure the pup, but in this test you are doing just that. The signal isn't generated by the pup. It's generated by the generator. The pup changes it's characteristics in an apparently unknown way. In this case you don't know what the signal coming out of the pup is. You only know what's going into the pup, and what's coming out of the cable. You're measuring the characteristics of both the pup and the cable. I thought we were trying to isolate the effect the cable had on the signal as a result of how it might load the pup.

If you want to know what the cable is doing, you need a precisely characterized input TO THE CABLE that you can compare to the output of the cable. If you want to see how the cable loads the pup...which we want to...then the impedance characteristics of the pup need to be part of the CONTROL. Anything else skews the results.

Even so, dumping a signal into a pickup with an active source leaves a situation where the impedance characteristics of the signal generator will reflect through the pup to some degree.

Why not fashion a ferrous extension onto a small speaker and place it in the magnetic field of the pup. It doesn't have to make perfect pink noise. Whatever the signal is just needs to be thoroughly characterized before the cable.

Even just holding the voice coil of a small speaker near it MIGHT get some kind of result. You'll have to measure your control first regardless...so you know what you're comparing.
 
#317
Jazzdogg said:
Thanks for the clarification, Mr. Deck. You definitely know more about these issues than I do.

In my simple mind, I think of white noise as a signal that measures flat electronically, while pink noise is essentially white noise that has been EQd to present an audible signal that compensates for Fletcher-Munson equal loudness contours by boosting both lows and highs relative to mids. If cable capaticance saps highs, and highs were boosted via pink noise, would the results of the plot be discernable? If measurable, would it be audible on a gig?
Click to expand...

White noise contains an equal amount of "stuff" per unit frequency. Pink noise is equal stuff per octave. Another thing I should have mentioned is that early spectrum analyzers were based on filters that broke up the spectrum into bands of constant width in octaves, such as 1/2 or 1/3 octave bands. Pink noise fed into such an analyzer would show a flat display.

Modern analyzers are based on FFT math, which is based on equal frequency increments. White noise fed through an FFT shows a flat spectrum. But from an audio standpoint, we hear in octaves and fractions of octaves, so it makes sense to display the FFT spectrum on an octave scale. When you do this, you have far fewer data points in the lowest octave than in the highest octave, so the low frequency portion of the curve is more affected by noise. Using pink noise can compensate for this.

So the difference between white and pink noise is purely mathematical, but it turns out that pink noise is closer to what we hear in terms of equal loudness than white noise.
 
#318
This all seems to me like you are testing the response of pink noise and not a bass though these cables, that seems to flaw the experiement.

In addition I have used a DBX RTA mic and drive rack system, to flatten the response of my PA. I have a set of EV speakers and Bag Ends (subs & tops) everything else is the same. So I set it up "flattened" the bagends, afterwards plugged in my bass through a DI and played. Then I did the same with th EVs, you know what my bass sounded different even though I used an RTA mic and software so that each system had a flat response at the location of the mic.

Why is that?
 
#319
No, I think the pink noise is fine. It's a matter of how you get the pink noise into the cable. You need to put the impedance characteristics of the pup on the pink noise without skewing the result in order to see how the cable will effect the signal. With the particular pup being used in the experiment, you would get a real-world result with pink noise and a broad scoped picture of the cables response. Same token, a spec like IM distortion doesn't mean nearly as much with a bass signal as it does with a complex wave.
 

Latest posts

Top Bottom
Back