manueljenkin
100+ Head-Fier
- Joined
- Sep 18, 2015
- Posts
- 410
- Likes
- 218
Hi Manuel,
The way I measured is that I stick an electret capsule mic into my ear canal and put the headphones on.
There could be all sorts of problems with that setup, but it's the closest to what would be the usage scenario, so whatever is happening with parasitic vibrations, it would probably happen in use.
Sure, I measured bursts and IMD, just for fun. It follows the falloff towards 500Hz just like THD does, so it doesn't give more usable data.
In transducers, the IMD will show no surprises as with amplifiers, as there is no increasing non-linearity as you go towards small signal (crossover distortion and such), so it is very strongly correlated with THD and it is mostly governed by the same mechanisms as THD.
Anyhow, there are several standards for measuring IMD and I still don't know how IMD reflects in what is audible as a problem with transducers, so I don't publish such stuff.
It's great for RF, though, when you need to directly express the number for selectivity of a radio receiver, for example, but I don't see much use for it with transducers. For Amps that depend on lots of feedback, maybe, as you can see how the noise floor looks like, if it's random noise or not, but in transducers, the THD is large enough that noise floor is never just a random noise, so it won't tell nothing new as THD mechanisms will govern the behavior if IMD.
Anyhow, digging deep with measurements is good when you need to solve audible problems with the transducer you're making. Because to my ears, the SR1a has the cleanest midrange I've ever heard on speakers or headphones, I consider that part of the work done. Having midrange THD better that 0.25% at 100dB, along with the cleanest impulse response in headphones ever, attests to the non-problematic midrange.
Simply, there was no reason to try to explain anything with IMD, as it follows THD, or burst envelopes, as they follow the Impulse response.
Speaking of audibility of distortion, I did a lot of testing for distortion in low frequency.
What I've learned correlates very well with what Axiomaudio published:
The explanation is here: https://www.axiomaudio.com/blog/distortion
There we see that below 40 Hz, the noise can be louder than the tone and we'll have trouble detecting it.
Also, up to 280Hz, the noise can be just 12-13dB lower than the tone and we'll have trouble detecting it.
Now, this is done with noise that doesn't correlate with music, so it's not THD. THD should be harder to detect.
In any case, I've managed not to do go that high with distortion, but had to stop reducing it somewhere to make a feasible product.
The goal was to get inobtrusive (if not inaudible) levels with 100W/8Ohm amplifier.
As you go higher with power, things will become progressively more audible until you reach the point of demanding more excursion than the mechanical limit is.
For Classical, acoustical Jazz, I'd say 150W is safe and modern stuff is safe up to 100W.
For driving the headphones, the requirement is simple. You have 0.4 ohms with the cable and you need to use EQ that will flatten the open baffle cancellation.
I think all that has been discussed before page 80 or so, of this thread.
Cheers,
Alex
The burst responses follow up with thd on good systems. But there are lots of bad systems where they just don't correlate. Happy to hear the raal is one of those good systems.
I personally own thx00 and I know the list of aberrations it produces which go unnoticed in FR and thd charts. I have a full on explanation of why thx00 has a sibilance with male vocals even though it's thd and fr says opposite. I have test tones which can measure the same, I'm currently writing out equations which describe the same.
Also thanks for the measurement advice, using an electret in the ear and measuring. I have to try the same for my srh1540 and if I could find, a srh1840. They seem to measure bad in thd but in reality it's actually the measurement rig rattling. That kind of thd is quite impossible mechanically considering their Csd is among the quickest.
Last edited:
