How much does source impedance of an amp alter the sound of a headphone? - Page 3

You've completely misunderstood the point of a low-impedance amp. The whole point of having a low impedance amp is because I know nothing about what you like! 

The whole point to have an amp which colors the sound the least and is the most predictable in how it will change the sound of the headphone. If I have a 120 ohm source, it would ruin the sound of any balanced armature headphone I plugged in, because it would screw up the cross overs, it would tweak the frequency response curve by 8-12dB in certain areas. But, putting IEMs aside, let's focus on normal headphones.

With most, but not all, dynamic headphones, a high impedance source will color the sound by emphasizing the 70-120hz area. This is the area which adds subjective "warmth" to the headphone. Many people love this sound with many different headphones, hence the popularity of tube amps. The problem is just that you can perfectly imitate this sound change with equalization. Famously, the Carver Challenge had one man take on the audiophile magazines by making a solid state amp sound like the expensive tube amps. One thing about tube amps is that they normally have a very high output impedance, which generally colors dynamic speakers by coloring them in the 70-120hz area. Carver was able to make the solid state amp sound so close that nobody could tell it apart in testing.

The issue isn't that the coloration is bad, the issue is that the coloration is unpredictable without a ton of data about both the amp and the headphone you're hearing. Not every headphone will have the boost in the 70-120hz region. Not every headphone will have the same amount of boost in any region. The reason it is more important for lower impedance headphones is because they have a greater amount of coloration from a high impedance source.

The way I see it, I want amps to color the music the least. That way, we have an equal ground to talk their sound, and we can also have the blankest slate to equalize them to how we want to hear it. I'm not telling you what sounds better, I'm giving you the blankest, most predictable slate to chisel the headphone's sound into.

Quote:

Originally Posted by nikongod 

 

[sarcasm]

I'm sorry to have to be the one to tell you that you are doing it wrong. 

If you drove them properly, from an amp with a 0-ohm output impedance, you would find them much more to YOUR liking.

That's right, I know what you should like and I'm going to tell you how to get it!

Please pay no attention to the fact that I know absolutely nothing about what you like. 

[/sarcasm]

 

Glad to hear you found a sound you like. I think the mixed drinks analogy is perfect. 

Quote:

Originally Posted by nikongod 

 

I'd say I understand the point of a low-output impedance amplifier pretty well. I also understand the point of a high output impedance amp. 

 

Whether to run on a high or low output impedance amp has little to do with what the listener likes (OK, maybe a little) but more to do with how the headphones were originally designed. 

 

I would say quite the opposite WRT colorations on single-driver systems. 

If the headphones were designed with the 120ohm IEC standard in mind, rather than the 0ohm standard there is no way to predict how running them off of a 0ohm amplifier will color the sound. 

 

We can debate which interpretation of this is correct back and forth depending on whether or not the headphone was designed for a certain standard or not. You can look at almost every type of professional grade equipment and, even though you'll find large impedances for headphone outs, they won't be close to 120 ohms in almost all cases. There are exceptions, but this is just generally how they work. I just quickly did a search for "audio mixer" and grabbed a $1500 mixer Onyx 24.4's spec sheet and it had a phones out of 25 ohms. For professional grade 300-600 ohm impedance headphones, this is a perfectly acceptable number. The industry has moved toward an output impedance which is significantly lower than the expected headphone impedance. From what I've seen in audio mixers, generally speaking the headphone out has about a range of 0.33 to 100 ohms. You can certainly find examples to show that there are mixers outside of that range, but I just feel that the great majority fall under it. 

I can't with certitude that this is what the headphone output impedance from professional audio equipment says, but that is what I think it implies. Both on the consumer level with the low output impedance of portable devices and on the professional level I believe the standard is now for a lower output impedance. I don't believe the 120 ohm standard was ever really used by headphone companies, with a few exceptions like the DT48. 

Depending on our reference point, we can argue that both are colored. I just think the evidence suggests that lower impedance outputs are what is expected. 

I was talking more about the sound differences between tube amps and solid state, suggesting that there is nothing fundamentally different. That was the only point I was trying to make. Speaker amps are a whole different monster because they now have a standard of 8 ohm speakers and incredibly low impedance amps for likely specifically this reason. I probably should have clarified that a bit more before I brought up Carver's Challenge.

When I made that post, I really did not think anyone would bring up the 120 ohm standard and try to argue for it. It's a legitimate point, but we once again have the problem where it can be argued in reverse. If I have a high impedance amp and my friend has a headphone designed for low impedance outputs, it would sound different if I didn't have a profile made. The difference is that you can't make the high impedance amp play nicely with the crossovers of a multi-driver BA IEM and I still maintain that very few headphones were legitimately made to the 120 ohm standard. 

Furthermore, a 0 ohm amp can be changed to be a 120 ohm amp with a resistor in the line, as you said. If I have a cable with a resistor inside, I can make that 0 ohm amp appear to the headphone like a 120 ohm amp. I cannot make a 120 ohm amp look like a 0 ohm amp. In both situations, there is more flexibility from a 0 ohm amp. I don't like this solution because it's a DIY soldering kind of a solution, but I feel it is a point worth bringing up. The only way to change a higher impedance amp to a lower impedance one would be to open the amp itself and swap parts in the amp, which I find to be a much more risky option.

The 120 ohm standard was just not as widely adapted as the 8 ohm standard of speakers, and that is honestly kind of a shame. It would have been very cool to have a single standard so we wouldn't have these issues. 

This is actually why frequency response is now one of the less important characteristic I pay attention to when I'm looking at headphone reviews. Assuming no funky massive roll-offs, FR changes can be made. Characteristics like phase response, tightness of the bass, decay of the sound, resolution and distortion cannot be changed without opening up the headphone and either modding it like a T50RP or changing the housing. 

Coming back to what is colored versus what is not colored, I feel like I've explained my position in this post enough.

That would be optimal.

In almost every single case, you don't. As I said, the only headphone I know of, personally, that was actually designed for a 120 ohm output is the DT48. I have also heard that Sennheiser designs for 0 ohm output before, though. 

As for the HD800, a higher Z-out would result in a warmer sound, given the impedance curves on it. A quick glance at the HD800 reviews on this site show that people find the bass very detailed, but a little lacking. The word "anemic" when referring to bass shows up 4 times on the Head-Fi review page on this site. Most likely the neutral/possibly overdamped bass response of the HD800 is too weak for most listeners and that bump from the higher Z-out makes the headphone sound more warm and appealing. Of course, the same bass hump could be added with equalization on a low-Z output and, in this case, the low-Z output is closer to what Sennheiser intended -- a neutral, analytic sound. 
 

Quote:

Originally Posted by redwarrior191 

actually I agree more with nikondgd in this case.. however, it could raise another question, how could we know whether headphone X is designed for high or low output impedance?? AFAIK, sennheisser declared before that their headphones were all designed with 0ohm output in mind..and yet it seems that some ppl like to run the hd800 from higher Zout better..

That's a mighty, mighty sneaky way of presenting the information! Showing the benchmark in negative dB is a very underhanded way of showing the data. Let's put those in a bit more balanced numbers.

The Sennheiser measured in the article showed up as -54dB versus -100dB. -54dB is 0.2% of THD. -100dB is 0.001% of THD. Thus, it's basically a 0.199% shift in THD at the worst point and lower elsewhere. Now, this is also at the almost unaudible 20hz! I doubt someone would notice the difference between sound when there is a tiny, tiny shift in distortion on those low rumble frequencies. You feel those frequencies, not hear them. 

But what about at a more reasonable number like 100hz? That's -67dB versus -100 dB or 0.04% THD vs. 0.001% THD. That's not audible, I'm sorry. 

Do I doubt these measurements? Nope. The only problem is just that the levels of the distortion they're pointing at and saying, "Look how bad this is!" are really very minor issues. I highly doubt that it would make an audible difference, if any.

Next graph, please!

Oh my! Now, I take issue with this graph. A damping factor of 2 is NOT going to give you 100% THD+N. Do you know how bad 100% THD+N would be? How unbearable that would be? I'm sorry, but that's definitely a measuring artifact of some sort, there is NO WAY that you'd get that kind of distortion out of this. If a damping factor of 2 resulted in that much distortion at peaks, you can bet that there would be a ton of moaning about odd awful sounding artifacts from high impedance sources. 

Edit: I feel the need to clarify this a bit. Artifacts like that show up when doing THD+N tests all the time at the harmonics, but it does not really imply that there is 106dB of noise at that point under actual use.

Anyway, let's tackle the more normal numbers in this graph, shall we?  Look at those peaks of -71 dB compared to reference level. Wanna know how much distortion that is? 0.028%.  Once again, that's not audible. 

I don't doubt the measurements from Benchmark Media. I'm just saying that they're meaningless and actually prove that there is no audible distortion difference. 

When I first skimmed those articles, I didn't notice the y axis being dB. That's just misleading. 

Quote:

Originally Posted by kingpage 

http://www.benchmarkmedia.com/discuss/feedback/newsletter/2011/12/2/0-ohm-headphone-amplifier-sonic-advantages-low-impedance-headphone-amp

http://www.benchmarkmedia.com/discuss/feedback/newsletter/2011/12/1/do-specifications-lie-do-our-ears-lie-where-does-truth-lie-examination

 

A couple of articles I saw someone posted on you-know-who's blog are interesting and relevant.

 

It's not just about frequency response that's changed with not enough damping factor, but also THD+N.

It all makes sense. There's reasons why the low-Z standard became unofficially adopted. Looking beyond the major reason of the crossovers, the portable age brought battery power to the table, which is a major reason why a new standard had to be made. Portable devices didn't have the voltage to be able to handle the voltage drop across a higher impedance to the headphones, which wasn't an issue for previous audio equipment. When amps are pulling energy from the wall, they didn't care about this drop, but it's mission critical for a battery powered device. Basically, even if a high impedance standard was actively used it would have been obsoleted by the outbreak of this technology. 

This is not blind faith in a low-Z standard. There are functional reasons for it being used and why I bet almost every single modern headphone is made on that standard. In that sense, it is more "correct" than a high-Z source standard. I did find another headphone that followed the 1996 standard, though -- the 1997 DT931. I still haven't found a 21st century example besides the DT48E for headphones that want a 120 ohm source, though. I did read some people suggesting the DT880 was meant to be run off a high impedance source because they would otherwise become too bright, so I ran the calculations. 

DT880 250 ohm from 120 ohm source FR deviation at 1V:

Max impedance of 300 ohm at 70hz, min impedance of ~240 at 2khz.

300 / (300 + 120) = .714V

240 / (240 + 120) = .667V

20 * log(.714/.667) = .59dB

Not an audible shift. There is no audible frequency response difference in running the 250 ohm or the 600 ohm DT880 from a 0 ohm source and a 120 ohm source. I won't look at the DT880 32 ohm because that was intended for portable devices which have low-Z outputs. Regardless of what standard the DT880 was made for, the perceived brightness of the headphone won't change so it shouldn't even be in the discussion for Beyerdynamic intending for the headphone to be less bright with a higher impedance source.