[luvdunhill]

Quote:

Originally Posted by nikongod /img/forum/go_quote.gif in multi-driver crossovers, the phase shifts &turnover points of a crossover are essential to get right. in reality though, it does not matter if BOTH are at exactly 3khz, or both are at EXACTLY 2.9khz. As long as they mesh properly they will work. the problems come when things DONT mesh up properly. the sounds from the 2 drivers can cancel or partially cancel (I agree) which is all sorts of screwy. In a single driver system this whole issue is less important. there is nothing else to reference the mis-phased sound to, and you just go with the flow.

good examples, but I was thinking more along the lines of padding down a tweeter, not crossover points.

Point being that human hearing has much greater resolution than 3dB.

 

[tomb]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif good examples, but I was thinking more along the lines of padding down a tweeter, not crossover points. Point being that human hearing has much greater resolution than 3dB.

... but as an industry standard, 3dB is the convention used for "barely perceptible"

 

[luvdunhill]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif ... but as an industry standard, 3dB is the convention used for "barely perceptible"

I'm not trying to be a donkey here, but I'd like to see a reference to this claim.. Half-power bandwidth has history that preceeds audio.

It's easy to test... see if you can distinguish +- 3dB, I'd wager that you'd find the difference much more than "barely perceptible".

 

[tomb]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif I'm not trying to be a donkey here, but I'd like to see a reference to this claim.. Half-power bandwidth has history that preceeds audio. It's easy to test... see if you can distinguish +- 3dB, I'd wager that you'd find the difference much more than "barely perceptible".

Well, I'm not trying to be persnickety here, either, but the North Avenue Trade School* down the street might have something to say about using your reference of Georgia State's HyperPhysics data in a licensed engineering practice setting.

Regal may have stated it a bit strongly, but as stated, 3dB is the recognized industry standard intensity level change for a "Barely Perceptible" subjective loudness change (else why the focus on 3dB for "cutoff?"). My own reference may be dated (I've only used it since 1980), but the 6th Edition of "Mechanical and Electrical Equipment for Buildings" by McGuinness, Stein, and Reynolds has the following table in their chapter on "Fundamentals of Architectural Acoustics". It's "Table 26.3" (facsimile-reconstructed here to save me from scanning):

At the same time, you are no doubt correct. It's no stretch to say that someone in the audiophile community building, testing, and critically listening to audio equipment can tell the difference at a much smaller interval than 3dB - especially at certain frequencies. Your reference for the HyperPhysics confirms this - stating that variable frequency sensitivity by the human ear must be factored into the sound intensity if one is looking for an absolute difference. Nevertheless, industry has to pick a convention for a common frame of reference outside of the niche audiophile community. That table summarizes it for many cases.


*Georgia Tech

 

[tomb]

Quote:

Originally Posted by FAQ /img/forum/go_quote.gif May I ask a question not related directly to the thread topic? I have a DAC with a .15uF output capacitor (I had to keep the value small for it to be high quality), and an amp with a 50K input impedance. I have calculated that the high-pass corner frequency (-3 dB) is at 21 Hz. So where can I expect the response curve to "start to fall", like for instance with a 1 dB attenuation (which I suppose to be barely audible)? Thanks in advance!

You can model it from the curve I published a few posts back - it's probably going to be somewhere between 90 and 100Hz where the curve starts. I think most of us size an ouput cap for connection to a 50K input pot on an amp at ~ 0.47uf to 1.0 uf. You don't want to lose anything off of a DAC. It gets a little less critical at the other end at the connection to the headphone.

 

[dsavitsk]

Quote:

Originally Posted by FAQ /img/forum/go_quote.gif hear the difference in low bass performance

A lot of people would be surprised at what actually constitutes "low bass". It is worth your (and everyone's) while to find a signal generator, or download some test tones, and get a feel for what 20Hz, or 60Hz, or 120Hz actually sounds like. Oh, and for the record, I don't buy this new claim that phase distortion caused by a high pass filter are inaudible, or that they aren't distortion. True, it is probably a bigger issue with crossovers where it can lead to waves canceling, and it is not harmonic distortion as such, but I'll stand by it being an issue with the 3dB point.

Oh, also, as to whether 3dB is audible, like everything, this is more complicated and less understood, than it should be. A 3dB drop caused by a inductor attenuator, while it measures the same, seems to sound smaller than a 3dB drop caused by a resistive attenuator. I don't think anyone has a clue why, but the perception is different.

 

[holland]

Hmm, interesting post dsavitsk. FWIW, I can't actually hear 20Hz, but I sure as hell can feel it...a bit hard with headphones though as most roll off before that. You do have a good point. Some people perceive low bass as a higher frequency than they think it is.

Regarding caps and inductors. Does that imply that one is subjectively better than another (i.e., inductive filter vs the typical RC)? I know many people are against such networks, but this leads somewhat towards zobel RLC networks at the outputs of amps. I view them as a necessary evil since I'm not constantly monitoring if I'm causing adverse effects or not with various loads.

@ FAQ, hehe, yeah, uAmp107, one of my favorites for sure...not. Some people seem to like them, though. I recall it ranked fairly high in Skylab's initial viewing of it, but it's slunk down towards the bottom over time.

 

[regal]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif A. Oh, and for the record, I don't buy this new claim that phase distortion caused by a high pass filter are inaudible, or that they aren't distortion. Oh, also, as to whether 3dB is audible, like everything, this is more complicated and less understood, than it should be. A 3dB drop caused by a inductor attenuator, while it measures the same, seems to sound smaller than a 3dB drop caused by a resistive attenuator. .

Great discussion.

This stuff is all very easy to test with a set of phones, a tone generator and a VST like Ozone.

With my rig I cannot hear any difference between a 40 hz tone at 78hz and 81 hz. Also no difference in music played with up to 180degrees phase distortion (a cap doesn't get near this amount of phase distortion.)

I think dsavitsk has high end equipment (I think even a DC coupled tube amp with no output transformers or caps) so he may be able to pick up on small nuances like this that those of us with more common equipment wouldn't. Most of us have bigger issues with our equipment like opamps and S-D chips in our DAC's.

 

[philodox]

I think that the bass rolloff with output caps is a little over stressed. For instance, my amp has 50uF caps on each of the 4 outputs [balanced] and my RS-1's still sound lovely. From what I understand, if you have a cap on all 4 channels of a balanced amp [rather than one on each side of the amp] you require twice the capacitance... so effectively that is only 25uF that the RS-1's are seeing. I plan on raising the value of these caps in the future, but it is definately not an urgent change.

Any idea what would be an ideal value for my amp given that it will primarily be used for K340's and RS-1's?

I've been looking at the Mundorf M-Cap 250V and 400V MKP lines as well as the Mundorf M-TubeCap line since they have values ranging from 100uF to 330uF. Would these be good choices?

EDIT: Looks like Solen has some caps ranging from 100uF to 220uF as well. I think they are the same motor run style that I am currently using, so I wouldn't need to change much. Havent been able to find pics of teh Mundorf caps yet, so I'm unsure if they are radial or axial...

 

[dsavitsk]

Quote:

Originally Posted by regal /img/forum/go_quote.gif Also no difference in music played with up to 180degrees phase distortion (a cap doesn't get near this amount of phase distortion.)

Even with phase distortion on only part of the signal? I would think that 180 degrees out of phase for the hole thing is fine, but on not hearing it when just the bass is this much out of phase, and where the change is gradual, would be surprising.

Quote:

Originally Posted by regal /img/forum/go_quote.gif I think dsavitsk has high end equipment

I like to think so, but then, it is only high end to me. Someone else might not like it at all. I don't make any claims on actual measured performance, only my subjective preference.

Quote:

Originally Posted by regal /img/forum/go_quote.gif (I think even a DC coupled tube amp with no output transformers or caps)

Actually, everything is pretty much transformer coupled -- two digital pulse transformers (on each end of the spdif cable), transformer on the output of the DAC, and transformer on the input and output of some of the amps (just output on others). Some of these are parafeed, too, which means cap coupling in a sense, though it is a little different than RC cap coupling. But, no opamps for whatever that's worth.

Quote:

Originally Posted by philodox /img/forum/go_quote.gif I think that the bass rolloff with output caps is a little over stressed. For instance, my amp has 50uF caps on each of the 4 outputs [balanced] and my RS-1's still sound lovely. From what I understand, if you have a cap on all 4 channels of a balanced amp [rather than one on each side of the amp] you require twice the capacitance... so effectively that is only 25uF that the RS-1's are seeing. I plan on raising the value of these caps in the future, but it is definately not an urgent change.

Most pop music doesn't go as low as people think, and most stereos don't really produce much true bass -- more like lower midrange, so we aren't really used to hearing it. Also, higher harmonics can be interpreted as bass by the ear, so you'll sometimes hear what isn't actually there. That said, you are probably in that range where even just doubling up on the caps would make a difference. Shooting for 2Hz as a 3dB point is anal retentive, but shooting for 40Hz to 60Hz is probably minimal -- especially with your phones. Years ago, I used my original Millett with the stock 100uF output caps, and I thought everything was fine. Moving to bigger caps did give a little deeper bass, but it wasn't a huge change.

 

[nikongod]

Quote:

Originally Posted by philodox /img/forum/go_quote.gif I think that the bass rolloff with output caps is a little over stressed. For instance, my amp has 50uF caps on each of the 4 outputs [balanced] and my RS-1's still sound lovely. From what I understand, if you have a cap on all 4 channels of a balanced amp [rather than one on each side of the amp] you require twice the capacitance... so effectively that is only 25uF that the RS-1's are seeing. I plan on raising the value of these caps in the future, but it is definately not an urgent change.

With some OTL tube output stages you can "get away with" an undersized output cap because of the fact that the amp puts more POWER into a higher impedance load. As the load impedance as seen by the amp RISES (the impedance of the cap at LF plus the headphones) the power goes up too. most calculations for bass rolloff through a cap assume that power decreases as load impedance increases.

Although you can get power to the headphones through an undersized cap, im not as sure about the effects on phase.

There is another school of thought that DELIBERATELY rolls off deep bass so that an amp which may be slightly underpowered (or one which has transformers that can only play to 40hz) in this area can perform better everywhere else.

 

[tomb]

Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif <snip>Shooting for 2Hz as a 3dB point is anal retentive, but shooting for 40Hz to 60Hz is probably minimal -- especially with your phones. Years ago, I used my original Millett with the stock 100uF output caps, and I thought everything was fine. Moving to bigger caps did give a little deeper bass, but it wasn't a huge change.

Anal retentive? OK, but maybe it's anal retentive to sweat about the difference between a 100uf cap and a 470uf one. 470uf is all we're talking about to assure flat bass down below 20Hz for almost all major cans. The difference is in pennies for many caps. Spend a few extra of those pennies and then just forget about even having the discussion.

Spend that effort on worrying about which brand of cap or how to bypass it for best sound - starting with reading your Notes on Output Coupling Caps.

 

[regal]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif the difference between a 100uf cap and a 470uf one. 470uf is all we're talking about to assure flat bass down below 20Hz for almost all major cans. The difference is in pennies for many caps. Spend a few extra of those pennies and then just forget about even having the discussion. [/URL].

Negative, the idea is to avoid electrolytics. A 120uF Solen is $35 , while a 330 uF is $95, so the difference in price for a pair is $120. And Ill bet money a $35 120uF Solen will sound better than a $15 470uF Blackgate on the end of a Maz or Bijou. Well I'm going to bet as I plan on ordering a pair today.

 

[tomb]

Quote:

Originally Posted by regal /img/forum/go_quote.gif Negative, the idea is to avoid electrolytics. A 120uF Solen is $35 , while a 330 uF is $95, so the difference in price for a pair is $120. And Ill bet money a $35 120uF Solen will sound better than a $15 470uF Blackgate on the end of a Maz or Bijou. Well I'm going to bet as I plan on ordering a pair today.

Personally, I would take that bet. By the time you deal with the size, the Solen's bad rep, and the price difference that you incur, I don't think it would be a good choice - but admittedly, that's just my opinion.

Work the numbers - bypassing with a high quality film cap on an electrolytic will have the film cap seeing most frequencies down to the last few octaves where the electrolytic doesn't sound that bad anyway.

 

[regal]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Personally, I would take that bet. By the time you deal with the size, the Solen's bad rep, and the price difference that you incur, I don't think it would be a good choice - but admittedly, that's just my opinion. Work the numbers - bypassing with a high quality film cap on an electrolytic will have the film cap seeing most frequencies down to the last few octaves where the electrolytic doesn't sound that bad anyway.

It is an expensive experiement, several of the european Bijou builders are reporting good results with the Solen's. I know you like by-pass caps but many don't believe in it. I think Runelight is one. Have you done any A-B comparisons w/ vs w/out a bypass?