[megabigeye]

^^There are quite a few ‘headphone calculators’ out there that do the maths for you.
http://www.digizoid.com/headphones-power.html
Then again I’ve tried several amps that on paper seemed like a poor match for my HE500, yet in a blindtest up against a kilobuck monster amp, I could detect no differences outside of cranking the volume beyond what I find pleasing.

The online calculators are useful in a pinch, but knowing how to do the math yourself allows you to see and understand what's going on. It's the difference between having a ready answer and understanding that answer. Plus, writing out equations makes you look smart.

 

[Kammerat Rebekka]

I’d rather look stupid and use my time enjoying the music

 

[skhan007]

LOL! There is a good amount of self-validation that occurs when one writes out formulas and solves for unknown variables. That's were there calculator is handy.

I've done years of amp repair and modification on my guitar tube amps and it's fun. For fixed bias amps, I learned how to use a multimeter to see the mA draw on octal power tubes and set the bias, according to calculations for plate dissipation and DC voltage. For example, on my Marshall amps, EL34 tubes are generally biased between 60-70% dissipation, are 25 watts each, and DC voltage can be around 400-450 volts. Doing the math gives me range of 32-38 mA. Interestingly, math gives you the ballpark settings and then you use your ears to determine where it sounds best. Sometimes biasing cold preserves tube life (like in Mesa amps) but may not sound as good. You've gotta find that sweet spot! Sometimes the math and sweet spot don't reflect the same values. This is what makes tube amps so fun.

 

[bigshot]

Did a little poking around, and it looks like 85 dB is a fairly common reference level used in both mixing/mastering and also in home theaters. The 85 dB reference listening level is usually set with an approximately -20 dBFS signal though. Which means 0 dBFS should be 20 dB above that, at approximately 105 dB.

As I explained before, the 20dB headroom is protection in case a transient exceeds 80dB in a mix. It is *above* the sustained peak level, not measured from the average level. They're mixing at 24 bit, so they have a very low noise floor. They can afford to be safe.

But playing back music in the home is a totally different thing than recording and mixing it. A commercial CD doesn't have the sort of huge transient peaks you encounter while recording. Peak limiting has already been applied to allow the CD to play at a decent volume level and not shove all the signal down to the bottom of the dynamic range near the noise floor. Figure this... 20-30dB headroom, 55dB range for dynamic music, Now we are pushing -85dB which is close to the noise floor of a CD.

But this is all related to preventing digital clipping. The context we are talking about here is whether an amp is powerful enough to render all transients cleanly. If you turn the volume up to a playback level of 80 to 85dB, that is a LOUD listening level. If you have headroom in the amp to double that, you are absolutely safe. That is 10dB. 90dB is all anyone with normal hearing would need.

I'd recommend getting a SPL meter and measure some sound so you get a feeling for what dB figures refer to in actual sound. There is absolutely no reason to need 115-120dB. That is an insanely loud volume.

I’d rather look stupid and use my time enjoying the music

I'm with you. I only worry about it when I shop for equipment. When people ask in the forum for calculations, I don't bother because I know there are plenty of wannabe junior scientists who will jump at the chance to argue about the nuances of the math for five or ten posts. They enjoy it, and they can have it! I'm only concerned with what I can hear.

 

[ADUHF]

As I explained before, the 20dB headroom is protection in case a transient exceeds 80dB in a mix. It is *above* the sustained peak level, not measured from the average level. They're mixing at 24 bit, so they have a very low noise floor. They can afford to be safe.

But playing back music in the home is a totally different thing than recording and mixing it. A commercial CD doesn't have the sort of huge transient peaks you encounter while recording. Peak limiting has already been applied to allow the CD to play at a decent volume level and not shove all the signal down to the bottom of the dynamic range near the noise floor. Figure this... 20-30dB headroom, 55dB range for dynamic music, Now we are pushing -85dB which is close to the noise floor of a CD.

But this is all related to preventing digital clipping. The context we are talking about here is whether an amp is powerful enough to render all transients cleanly. If you turn the volume up to a playback level of 80 to 85dB, that is a LOUD listening level. If you have headroom in the amp to double that, you are absolutely safe. That is 10dB. 90dB is all anyone with normal hearing would need.

I'd recommend getting a SPL meter and measure some sound so you get a feeling for what dB figures refer to in actual sound. There is absolutely no reason to need 115-120dB. That is an insanely loud volume.

Fwiw, I hear and understand everything you're saying here. And always appreciate hearing your insights on this stuff, bigshot. And would agree that under normal listening, you probably need no more than the ranges you're discussing.

Would I ever use reference levels?... Probably not. But it's good to know if my gear could handle it, if I ever should. There's really not much more to it than that.

I have been to my share of noisy nightclubs and concerts though in my misguided youth. And probably acquired most of my current hearing damage from those experiences. So your comments are not falling on deaf ears (not totally deaf yet anyway).

I'm with you. I only worry about it when I shop for equipment. When people ask in the forum for calculations, I don't bother because I know there are plenty of wannabe junior scientists who will jump at the chance to argue about the nuances of the math for five or ten posts. They enjoy it, and they can have it! I'm only concerned with what I can hear.

The music is the thing. But yes, some of us are geeks. And also enjoy digging into and trying to better understand the numbers on this stuff, and exactly what they mean. So different strokes, as they say.

 

[bigshot]

The subject we were discussing was how much headroom do you need in a headphone amp to accommodate dynamic transients that give music punch. The answer is, if the amp can take your cans up to 95dB, it will be able to handle any transient in recorded music. I guess the subject drifted to something more theoretical with much bigger volume levels. Audio gear is a tool. The important thing is how that tool performs the job you need done. Nothing else matters.

The best way to understand the numbers is to experience them in the real world. But I wouldn’t recommend experiencing volume levels much above 80dB. It isn’t good for you.

 

[sander99]

As for me: I just happened to see an oversight in a calculation and always like to help if I can. I didn't have an opinion about what level someone would need, be it 85 or 95 or 120 dB, whatever it is it is nice to be able to calculate how to reach that level.

 

[bigshot]

You and I are a great team!

 

[ADUHF]

As for me: I just happened to see an oversight in a calculation and always like to help if I can.

Which is appreciated as well.

 

[skhan007]

So it looks like the ADI-2 does indeed have sufficient headroom to reach reference listening levels with a 300-ohm headphone, with power efficiency in the 97db/mW range. The Sennheiser HD800S is supposedly even more efficient at 100 dB/mW, so it should have enough headroom as well. Let's confirm it though, just for grins...

P= 10^[(115-100)/10] = 10^1.5 = 31.6mW

Looks pretty good.

The empirical validation is much appreciated here, as I'm learning. Great discussion.

While we were all doing our math homework, I had sent a message to Matthias Carstens, who is the head of RME product development, asking him if the ADI-2 can sufficiently drive 300 Ohm headphones (e.g. HD800, ZMF) and his reply: "Of course it can."

I'm going to go ahead and trust the math, the head of product development from RME, and my own ears. That being said, I will continue to test drive separate amps, listen for myself, and see if I can note a preference towards a separate amp or the RME.

 

[megabigeye]

As for me: I just happened to see an oversight in a calculation and always like to help if I can. I didn't have an opinion about what level someone would need, be it 85 or 95 or 120 dB, whatever it is it is nice to be able to calculate how to reach that level.

I always appreciate the help with the math and science. I'm really not a master or expert on that stuff, so it's always a little awkward for me to poke my nose into these conversations as if I know what I'm talking about. But, like you, I like to help where I can and it seemed like a question was asked that I (more or less) knew the answer to. I used to let people know the limits of my understanding straight off the bat, but that just made people assume the math was just another opinion on the Internet and therefore something to be argued with. Now I let people assume I'm an expert and then be embarrassed by it later.

I'm perfectly willing to admit that 120dB is completely ludicrous. I'd like to actually measure how loud I listen, but I really don't want to buy I measurement mic just so that I can do one measurement once, just so that I can look more like an expert on the Internet.

 

[bigshot]

I can tell you this, having measured my system. 80dB for more than a few minutes would be enough for the neighbors to complain and the dog to hide on the other end of the house!

 

[skhan007]

I used a simple spectrum analyzer and noted my preferred comfortable listening level is about 60 dB.

 

[megabigeye]

I once used my phone to measure how loud my speakers were, and if I recall it was something like 65dB when I figured my upstairs neighbors could hear it. I assumed that speakers sound loud at lower SPL than headphones do, so not really applicable.
I could try to measure my headphones with my phone, but I don't think it'd be remotely accurate.

 

[VNandor]

Here's some infographic on what average SPL the ears can tolerate for different time windows. According to this, a 120dB SPL transient peak doesn't necessarily cause hearing loss but from what I know about the average level of my music and how much time I spend on listening to music, 120dB would be well over the safe threshold for me.