iPod measurements

[antonik]

I did some measurements on my iPod using RMAA and my RMA Digi96/8 PAD , off-load and loading the iPod into different headphones (Grado SR60, Sennheisers HD25SP and HD650. Rather interesting picture... . It looks like the headphone output without load provides very decent performance as a line-out . However it also shows how much even iPod could benefit from a proper amplifier... . And why with any load below 100 Ohm there is a lack of bass. As a reminder - SR60 = 32 Ohm, HD25 - 80 Ohm, HD650 - 300 Ohm .

Have a look here:

http://www.ant-audio.co.uk/Data/iPod_test2.htm



Cheers

Alex

 

[aki]

This is really interesting, thanks for making these available.

So it seems that the well known bass roll-off from the iPod headphone amp only occurs with low-impedance headphones and can be minimized by using higher impedance phones.

Although it is disappointing to see that even with an 80 ohm headphone like the HD25, that the roll-off is still so significant.

 

[Spankypoo]

Interesting. What're the impedences of those 'phones?

Would testing it against incremental loads be useful? (E.g., 30 ohms, 50, 75, 100, 125, 150, etc.)

 

[sumguy_]

so how about this: if you hook up the headphone out to a portable amp (something like a cmoy) with a short cable, does this provide load on the ipod headphone out? i dont know much about how amps work, but i was just thinking, if you were to use it with an amp like this, the bass roll of will not exist... true?

a cheap cmoy, or a diy one then can be used to solve this issue for those who are bothered by the bass rolloff.

 

[antonik]

Quote:

Originally Posted by aki This is really interesting, thanks for making these available. So it seems that the well known bass roll-off from the iPod headphone amp only occurs with low-impedance headphones and can be minimized by using higher impedance phones. Although it is disappointing to see that even with an 80 ohm headphone like the HD25, that the roll-off is still so significant.

Yes, it is -3dB at 20 Hz with HD25, but with, say, SONY MDR-EX71 (16 Ohm)it would be -3dB at about 80-100 Hz... . It is due to the output capacitor size on the iPod.

Quote:

Originally Posted by Spankypoo Interesting. What're the impedences of those 'phones? Would testing it against incremental loads be useful? (E.g., 30 ohms, 50, 75, 100, 125, 150, etc.)

I did mention it in my first message - 32, 80 and 300 Ohm for the headphones I've used, and about 10K load without headphones.

Quote:

Originally Posted by sumguy_ so how about this: if you hook up the headphone out to a portable amp (something like a cmoy) with a short cable, does this provide load on the ipod headphone out? i dont know much about how amps work, but i was just thinking, if you were to use it with an amp like this, the bass roll of will not exist... true? a cheap cmoy, or a diy one then can be used to solve this issue for those who are bothered by the bass rolloff.

True. With a headphone amp connected to the output of iPod there is no bass roll-off.

For this very reason I'm making a class A portable amp for my own use now. Also an amp could make the output much louder - iPod outputs only about 300mV RMS maximum, and there is a problem - with low impedance phones you've got some power - i.e. on 30 Ohm about 3mW, but no bass. With 300 Ohm load (HD650) you've got all the bass but only 0.3 mW power.

Alex

 

[Iron_Dreamer]

What is even more disturbing to me than the bass roll off on low-z cans is the terrible THD and IMD results for those, not to mention channel seperation. Funny, considering the most commonly used cans with the iPod are low-Z.

 

[antonik]

Quote:

Originally Posted by Iron_Dreamer What is even more disturbing to me than the bass roll off on low-z cans is the terrible THD and IMD results for those, not to mention channel seperation. Funny, considering the most commonly used cans with the iPod are low-Z.

Yes, that is what I'll remember every time I connect my HD25SP to my iPod. And I've already retired my Sony MDR-EX71 till i'll have a proper amp for them... .

Alex

 

[Zoide]

Quote:

Originally Posted by Iron_Dreamer What is even more disturbing to me than the bass roll off on low-z cans is the terrible THD and IMD results for those, not to mention channel seperation. Funny, considering the most commonly used cans with the iPod are low-Z.

Yes, this is something that I truly don't understand. I really doubt that the Apple people were picturing their "hip", "trendy" customers going around with HD600s in the San Francisco subway

I wonder if this low-z problem is found in other DAPs?

At the same time, I'm surprised that with such an apparent bass roll-off there aren't more canalphone users advocating the use of an amp w/ the iPod. It's becoming more and more common to read the type of comment that goes: "amping the iPod doesn't make a big enough difference for me, so now I use it ampless for more portability". And then they talk about "Team Minimalism" and stuff like that...

 

[Wodgy]

Pretty much all DAPs use 220 ohm coupling capacitors and have the same bass roll off. Due to space issues, engineers are limited in the size of the capacitors they can use. However, it's not as severe perceptually as the measurements seem to indicate.

One of the reasons Etymotic designed the ER-6i was to counteract this rolloff by building in a natural bass emphasis.

 

[Wodgy]

By the way, here are Stereophile's measurements of the iPod. They're unloaded, but much more accurate as they're done with Audio Precision gear:
http://www.stereophile.com/budgetcom...34/index5.html

(As an aside, you can compare these measurements against the RMAA results to see just how poor RMAA is. It misses most of the interesting stuff. It's a shame there aren't better free software analysis tools out there.)

 

[antonik]

Quote:

Originally Posted by Wodgy By the way, here are Stereophile's measurements of the iPod. They're unloaded, but much more accurate as they're done with Audio Precision gear: http://www.stereophile.com/budgetcom...34/index5.html (As an aside, you can compare these measurements against the RMAA results to see just how poor RMAA is. It misses most of the interesting stuff. It's a shame there aren't better free software analysis tools out there.)

I've obviously read that article at the time

. However good the results on a proper measurement system are, it looks like they missed "the most interesting stuff", namely huge LF roll-off and enormous distortion and intermodulations into low-impedance load - typical for most headphones used with iPod. RMAA is not "poor" , it does what it does very well and for free

. I've got my RME Digi96/8 PAD calibrated by "real" measurement equipment and I know how far I can trust the results produced by RMAA. BTW, the results of my measurements "off-load" are fairly close to Stereophile measurements. On-load results do look ugly, thought, but it is not RMAA fault... .

Alex

 

[Wodgy]

You need to compare oranges and oranges, not apples and oranges. Stereophile should have done measurements under simulated or standardized load(s), as they do with power amplifiers, but their results are much more accurate for the unloaded case.

RMAA is indeed poor (though tolerable for the price). Where is the DAC's rising noise floor with frequency in the RMAA measurements? The RMAA noise floor actually decreases with frequency! It's totally inaccurate. RMAA's THD+N graph in the unloaded case misses the second harmonic by nearly 60dB! This is very poor accuracy.

It's not just this one test. As Mr_Radar's RMAA tests on his Toshiba's digital out showed in another thread, RMAA shows a flat frequency response, when in fact there's an IIR filter running causing a significant dip in his Toshiba's frequency response.

I wish RMAA was a better tool, but it will never improve if people don't point out its weaknesses. Your claim that "I've got my RME Digi96/8 PAD calibrated by 'real' measurement equipment and I know how far I can trust the results produced by RMAA" is just pure BS. You don't have any idea how far you can trust RMAA. It's a black box which generates unpredictable measurements.

As for your remarks about "huge LF roll-off" and "enormous distortion", people need to remember you're in the business of selling headphone amplifiers. The LF roll-off, with the worst case cans, is down -3dB at 32 Hz. While this is significant in some cases, it will be hard to detect with Grados and 32 ohm earbuds since they have practically no low bass anyway. As for "enormous distortion", give me a break. Worst case, at 20kHz, the distortion is -80dB down. Enormous? That's very decent measured performance for an op-amp output stage, on par with what you'd expect from the RA-1.

By the way, Antonik, please follow the rules of this forum and contact Jude so that your profile will be identified as a commercial vendor. If you're going to go around being alarmist and misrepresenting a few measurements, people should be able to clearly see you're in the business of marketing headphone amps.

 

[antonik]

Wodgy,

I'll try to answer point by point.

Quote:

Originally Posted by Wodgy You need to compare oranges and oranges, not apples and oranges. Stereophile should have done measurements under simulated or standardized load(s), as they do with power amplifiers, but their results are much more accurate for the unloaded case. RMAA is indeed poor (though tolerable for the price). Where is the DAC's rising noise floor with frequency in the RMAA measurements? The RMAA noise floor actually decreases with frequency! It's totally inaccurate. RMAA's THD+N graph in the unloaded case misses the second harmonic by nearly 60dB! This is very poor accuracy.

That is not "very poor accuracy" but just a difference in the measurement's conditions. That makes a direct comparision near impossible. For instance the distortion graph in Stereophile for 1 kHz into 100 Ohm load, not off-load, so you need to compare it with (roughly) HD25SP graph in my measurements (green line) and you may see that the distortion picture is nearly identical . If anything this is a proof of RMAA quality.

Quote:

Originally Posted by Wodgy It's not just this one test. As Mr_Radar's RMAA tests on his Toshiba's digital out showed in another thread, RMAA shows a flat frequency response, when in fact there's an IIR filter running causing a significant dip in his Toshiba's frequency response. I wish RMAA was a better tool, but it will never improve if people don't point out its weaknesses. Your claim that "I've got my RME Digi96/8 PAD calibrated by 'real' measurement equipment and I know how far I can trust the results produced by RMAA" is just pure BS. You don't have any idea how far you can trust RMAA. It's a black box which generates unpredictable measurements.

RMAA in my experience is a good and reliable tool providing you know how to use it and what limitations are inherent in the measurements by audioconvertors of a sound card. There is no BS in what I've said. There are some limits and there are some bags in RMAA as in any software and you need to be aware of that. However I did run RMAA in a controlled test comparing the results with the results of a proper audio analyser, quality THD meter and true RMS voltmeter . As long as you know the limitations you can say that the results are fairly reliable - but this is true of "any" measurement equipment.

Quote:

Originally Posted by Wodgy As for your remarks about "huge LF roll-off" and "enormous distortion", people need to remember you're in the business of selling headphone amplifiers. The LF roll-off, with the worst case cans, is down -3dB at 32 Hz. While this is significant in some cases, it will be hard to detect with Grados and 32 ohm earbuds since they have practically no low bass anyway. As for "enormous distortion", give me a break. Worst case, at 20kHz, the distortion is -80dB down. Enormous? That's very decent measured performance for an op-amp output stage, on par with what you'd expect from the RA-1.

Distortions on-load do look ugly with very extended harmonics structure. What is enormous is the distortion increase by 50 times with the 80 Ohm load connected.

Quote:

Originally Posted by Wodgy By the way, Antonik, please follow the rules of this forum and contact Jude so that your profile will be identified as a commercial vendor. If you're going to go around being alarmist and misrepresenting a few measurements, people should be able to clearly see you're in the business of marketing headphone amps.

Jude had my PM about my status change two month ago (and responded to it then, saying he'll do it). I suppose he is a busy man and can not attend to everything immediately. You can see what I do clearly from my profile. However I should note that at the moment I make amplifiers more for pleasure than for business. And please don't use silly words like "alarmist and misrepresenting" - it is not correct in this case. Moreover, I'm quite sure that Paul Miller's system measurements under the same conditions would produce very similar results. I may do these one day.

Alex

P.S. - just as I've posted it I see that somebody already took care about my status.

 

[breez]

Quote:

Originally Posted by Wodgy Pretty much all DAPs use 220 ohm coupling capacitors and have the same bass roll off.

Did a quick test with H120 using the lowly ADC of chaintech av710. I guess the card is ok for testing frequency response (results quite flat response from the sound card's DAC which has been measured quite flat with better ADCs). Seems that iRiver doesn't suffer from bass roll-off with headphone load.

 

[yeeyy]

Quote:

Originally Posted by antonik I did some measurements on my iPod using RMAA and my RMA Digi96/8 PAD , off-load and loading the iPod into different headphones (Grado SR60, Sennheisers HD25SP and HD650. Rather interesting picture... . It looks like the headphone output without load provides very decent performance as a line-out . However it also shows how much even iPod could benefit from a proper amplifier... . And why with any load below 100 Ohm there is a lack of bass. As a reminder - SR60 = 32 Ohm, HD25 - 80 Ohm, HD650 - 300 Ohm . Have a look here: http://www.ant-audio.co.uk/Data/iPod_test2.htm Cheers Alex

Looking at the Crosstalk figure, it is now confirmed that my previous finding on iPod poor performance on headphone output.

I did posted before at HERE and HERE (5th Post).