Do my Sundaras need more power or are they just faulty?

[bogart]

Folks, we are pretty far afield from the original discussion to help @nohomo_sapiens with their Sundara headphones and the missing soundstage. @hmscott may I suggest you start a new thread to share your experiences and link it here? All who want to follow it there can, and ask their pertinent questions.

This discussion is not furthering the original question at this point.

 

[jawbfl]

It is quite possible that you have a defective unit. I have used my Sundaras in conjunction with audio interfaces, MacBook Pro outputs, dongles and several dedicated amplifiers without issue. With the exception of limited volume and encountering clipping if certain sources are pushed with low frequency EQ, the Sundaras have sounded identical on everything.

I'm driving Sundaras from a Macbook, it gets loud but it feels weak and lacking the punch, I'm not sure what are the specs of the Macbook port but I think a powerful amplifier should provide more current to drive them properly?

 

[Weeb Labs]

I'm driving Sundaras from a Macbook, it gets loud but it feels weak and lacking the punch, I'm not sure what are the specs of the Macbook port but I think a powerful amplifier should provide more current to drive them properly?

What is your basis for comparison and is this purely a sighted observation? The 2017 MacBook Pro's output is capable of delivering about 40mW of power into a 37 Ohm load, which is more than adequate for the Sundaras.

Even with a -6dB pad for low frequency EQ, my MacBook Pro is perfectly capable of driving the Sundaras to uncomfortable levels without clipping.

 

[BartekG]

I'm driving Sundaras from a Macbook, it gets loud but it feels weak and lacking the punch, I'm not sure what are the specs of the Macbook port but I think a powerful amplifier should provide more current to drive them properly?

This looks similar to what I report here https://www.audiosciencereview.com/...dac-combo-to-power-sundaras.19230/post-641445. Is it possible that an increased demand for power is not satisfied by the amp instantly but after a delay during which the voltage drops?

 

[Weeb Labs]

This looks similar to what I report here https://www.audiosciencereview.com/...dac-combo-to-power-sundaras.19230/post-641445. Is it possible that an increased demand for power is not satisfied by the amp instantly but after a delay during which the voltage drops?

That would effectively be clipping, which is volume dependent.

Again, are all of these sighted comparisons? Does it occur at lower volume levels?

 

[hmscott]

Folks, we are pretty far afield from the original discussion to help @nohomo_sapiens with their Sundara headphones and the missing soundstage. @hmscott may I suggest you start a new thread to share your experiences and link it here? All who want to follow it there can, and ask their pertinent questions.

This discussion is not furthering the original question at this point.

When someone asks for a recommendation that includes listening with Hifiman Planar headphones, the Sundara in particular, and he wants an AMP/DAC recommendation, it is on-topic to share with him that there is a large percentage of Hifiman Planar listeners that enjoy the Tube sound of amplification when listening to Hifiman Planar headphones.

My posting to the OP was on topic, OP was looking for a recommendation about his Sundara Planar and whether his amp was the problem or his headphones.

I answered him directly and clearly with yes your amp is too weak, and gave him my experience with a range of Hifiman Planars, and suggestions for alternative amps, and where my own experience has taken me.

I never said tube equipment was "more accurate" or anything other than "I find the tube sound more enjoyable with my Hifiman Planars", a conclusion many in the Hifiman headphone community that includes the Sundara the OP's question is discussing.

Measurements inform the buyer apart from the manufacturer's published specifications and measurements.

There is no law that says we must purchase the top measuring equipment, or not buy the lower measuring equipment. We buy what think we might like, and keep what we do like. I've listened to a lot of hardware, so I know what I like.

I also like to make informed decisions, which is assisted by @amirm's testing, measuring, and posting those measurements on ASR. It is an invaluable actual hands-on confirmation of how the device is performing. Much thanks to Amir and ASR for this invaluable data.

We are not a slave to measurements. If I listen to equipment that provides me joy when listening, the measurements inform me, but the measurements don't rule my decision-making.

For my own purchases, the final deciding factor is how the device gives me joy when listening to music.

 

[solderdude]

Is it possible that an increased demand for power is not satisfied by the amp instantly but after a delay during which the voltage drops?

Again... there is no increased demand for power. Amps do not 'push' power nor 'grab' drivers and certainly not demand more power than dictated by the voltage present.
As Otaku+ already mentioned the output of a music source provides a voltage. That is not a constant voltage but varies. That voltage has a limit. That limit depends on the design of the amp.
The maximum voltage also depends on the current capabilities of the same amplifier.

So when either the voltage reaches its limit OR the current reaches its limit (both have a limit) and it depends on what limit is reached first the output voltage clips. Yes, the output voltage can also clip because a current limit is reached and that can happen before the voltage limit is reached.

The lower the voltage limit and the lower the current limit and the lower the impedance of the headphone and the lower the voltage efficiency of the headphone the earlier the device 'clips' meaning the output voltage clips meaning the SPL is limited.

When this happens and you turn up the volume beyond that point (depending on how 'hard' the clipping is) the loudness (volume) of the sound will still increase and bass levels will 'suffer' first as they have the largest voltage swing. The sound will turn 'nastier' but the volume of the sound keeps increasing but NOT the bass. That is clipped so stays 'behind'.

In these cases using an amplifier that has a higher voltage and or current limit moves the voltage clipping point upwards and so the clipping occurs at higher listening levels.

At some point, where peak SPL of 120dB is reached, the sound is so loud you won't go there simply because it is unpleasant for longer periods.
You can listen to 120dB SPL peaks for about a minute and this is impressively loud and because of equal loudness contours has tremendous bass power (makes Amirs earlobes shake). But you won't be doing that for long.

This means that when the amp chosen for driving your headphone can deliver the needed voltage (= drawn power, not delivered power) without clipping and the amplifier has enough gain and the music source has enough output voltage then having more available power is pointless.

Yes, having 3dB more headroom is desirable but not needed for normal listening levels.

When amplifiers (even a dongle has an amplifier inside) clips at SPL levels of say 90dB SPL in a specific headphone then it will sound as if it struggles with the bass. And it does. It clips the bass as that has the highest amplitude.
90dB SPL for bass is NOT loud at all. 90dB of bass is about 30dB above the hearing threshold.
60dB SPL of bass makes the fundamental of that bass inaudible already while the rest of the music is at a enjoyable soft background level.
90dBA is loud and for bass fundamentals to reach that level the actual SPL = 120dB !!

So in short... when you want to use your headphone at all kinds of levels it needs to reach voltage levels that can make it produce 120 to 123dB SPL. You don't need as much if you never play loud. Around 110dB is good for comfortable loud and 100dB ensures normal listening levels never clip.

That said.. efficiency levels are often given at 1kHz or 500Hz. Looking at the frequency response of headphones you will find most of them roll-off and some 'boost' in the lows. This too has an effect of how bass is perceived. When you have to boost it the voltage limit is reached much, much, much earlier as well.

It is a complicated 'play' between headphones, voltage efficiency and voltage levels that can be reached which determine how impressive bass will be.
It is not costly to make amps that can provide enough to power all but very rare and inefficient headphones.

There you have it. It's all science not magic and the laws of that science cannot be broken.
Audio may seem like magic and things may seem 'inexplicable' but they aren't.

Aside from all these technical things that need to be right, including distortion and frequency response and in lesser sense phase, there are other aspect that also have great impact.

Seal of the headphone
Bass extension and frequency response
recording itself
dynamic range (loudness compression) also a recording thing
Hearing sensitivity (changes over the years, you don't notice this till it is really severe)
Taste (think bass heads)
The brain itself is also a major part of perception.

Nope... it is not simple and science is a part of it. Only that part is measured.

 

[bustled_hedgerow]

Quite honestly, I think if I were to drive 1W of power in my Sundara's, the drivers would become damaged. This seems to be recurring however, the whole needing 1W to drive the Sundara's properly. It commonly pops up in forums and I'm always a bit taken aback. I don't listen to them much anymore but I can confidently recall listening to them for several months, always using an ifi zen dac/amp-all-in-one, using the 6.3mm SE connection on the high gain mode on that particular device, and never adjusting the potentiometer past the 2'o'clock position. The max power for SE mode for the ifi zen dac are as follows: 6.3mm (SE) 230mW@32 Ohm. I can't report what the exact volume level was that I was listening at but it was definitely loud, impactful, and at the time impressive to me compared to what I had been listening to. Track 9.) on the Interstellar soundtrack, Mountains, became one of my reference tracks based on the clear dynamic range on display, the build-up to the climax at the 2-minute mark, and the subsequent sub-bass rumble that presents itself, and all those elements were clearly evident and on display at the power level that I was at. I understand that not everyone experiences audio the same way however there comes a point when physics kinda determines things and they are no longer subjective.

 

[solderdude]

I think if I were to drive 1W of power in my Sundara's, the drivers would become damaged.

They won't. 1W short peaks in music is not going to destroy those drivers. average levels with peaks up to 1W are perhaps 30mW to 100mW.

The thing is that each time when you want to go a bit louder the required voltage needed rises a LOT.
To go twice as loud you need 10x the power (3x the voltage).

1W is needed when you want to play almost uncomfortable loud (for short moments) and EQ the bass up to Harman levels.
For comfortable loud levels (can listen to a whole song without the urge to turn down the volume) 100mW is enough.

 

[paolomo]

average levels with peaks up to 1W are perhaps 30mW to 100mW.

This is a very useful piece of information, thank you! How would this picture change with EQ? For instance, let's say I have a +6dB low shelf for my Sundara, how much would it affect the short peaks? Would they still be approximately 1W, or would they increase significantly?

 

[holicst]

I'd like to get some clarification here. I'm absolutely on the science/measurement side in general, but now I'm confused about power requirements. Reading all the above would suggest that once the amp has enough power make the given HP loud enough for personal preferences, the sound quality would be practically the same (only THD+N difference).

However, my experience is that this is not true. I've just made another test right now and I can immediately tell the difference between a plain laptop 3.5mm output and my E1DA 9038D using my Shure SRH940 and KB Ear Believe. I also tried to match the volume, but it was not needed at all because the difference was so obvious. Naturally a sighted test done by myself...

So now comes the question: as the THD(+N) of the source is lower with order of magnitudes than the IEM/HP, how come that sound is vastly different between the two sources? Both are loud enough, but the laptop sounds muddy, veiled etc. The E1DA makes it tight, clean, cripsp etc.

If the power does not matter if it is loud enough, then the 0.0X THDN vs 0.000X THDN difference would not be as audible as the difference I have heard. It seems more logical to me that the difference in max power could be the reason instead. It is also logical that providing more and more power would mean less and less improvement in sound quality, possibly nothing above 1W for the gear I used.

Also, when I tried some THD listening tests linked on ASR, I was not able to tell the difference with much higher THD+N added to music samples (sometimes even 1%+). This makes me doubt that the difference I heard is related to THD+N.

 

[solderdude]

This is a very useful piece of information, thank you! How would this picture change with EQ? For instance, let's say I have a +6dB low shelf for my Sundara, how much would it affect the short peaks? Would they still be approximately 1W, or would they increase significantly?

Then bass level power is increased by a factor of 4.

So if you normally would listen at average levels of 20mW (pretty loud for Sundara) and thus have about 200mW of bass peaks in there then at the same perceived loudness you would be needing 800mW.
That's why one needs a lot more power when EQ'ing the low bass extension.
Of course this would only be the case when those frequencies are really there.

 

[solderdude]

So now comes the question: as the THD(+N) of the source is lower with order of magnitudes than the IEM/HP, how come that sound is vastly different between the two sources? Both are loud enough, but the laptop sounds muddy, veiled etc. The E1DA makes it tight, clean, cripsp etc.

Do you happen to know the output resistance of the laptop. The secret might lay there.
Could also be that the HP out of the laptop has output caps which are too low in value for the KB (17 ohm).
Could also be that you are already running into clipping already on the laptop.

The dongle, most likely, can go louder and the laptop output voltage may well not go nearly as high as the dongle.

ONLY measurements can tell. The ears are not suited as measurement device for this.

SINAD and THD would have to be very bad to become audible (or lots of higher harmonics).

You don't need 1W for SRH940 as it is 114dB/V so 2V is already 120dB SPL = 100mW.
The fact that it is rated at 1W does not mean it NEEDS 1W.

The KB is 115dB/V so also goes loud enough with 0.1W

 

[VintageFlanker]

Why not?

 

[jawbfl]

What is your basis for comparison and is this purely a sighted observation? The 2017 MacBook Pro's output is capable of delivering about 40mW of power into a 37 Ohm load, which is more than adequate for the Sundaras.

Even with a -6dB pad for low frequency EQ, my MacBook Pro is perfectly capable of driving the Sundaras to uncomfortable levels without clipping.

I'm running Mid 2012 not sure if the specs of the amp/dac are any different, I have no amp to compare against just going by what I heard online on the performance of the Sundara, on some tracks my M50x sound subjectively better.

 

[paolomo]

Then bass level power is increased by a factor of 4.

So if you normally would listen at average levels of 20mW (pretty loud for Sundara) and thus have about 200mW of bass peaks in there then at the same perceived loudness you would be needing 800mW.
That's why one needs a lot more power when EQ'ing the low bass extension.
Of course this would only be the case when those frequencies are really there.

Thank you so much for the detailed and quantitative answer! To stay on topic with power vs. Sundara, some people lament lacking soundstage with low power. Is there any way power shortage/availability affects phase? (Phase is the first I think about when I think of soundstage).

Also, I understand that's many questions, feel free to ignore them

 

[jawbfl]

So if supposedly drive Sundara from Macbook and a JDS Atom volume matched to a level that the Macbook isn't clipping, would they provide the same amount of power since the load(Sundara) is the same? and the sound difference will only be because of the noise performance difference?

 

[Weeb Labs]

So if supposedly drive Sundara from Macbook and a JDS Atom volume matched to a level that the Macbook isn't clipping, would they provide the same amount of power since the load(Sundara) is the same? and the sound difference will only be because of the noise performance difference?

Yes. If both amplifiers are capable of delivering the current drawn by the load at a given output voltage (volume dependent) and both exhibit low output impedance (less relevant to Sundara), noise and distortion will be the factors which determine whether the amplifiers can be distinguished from each other.

In practical terms, almost any modern source will exhibit noise and distortion below the threshold of audibility.

 

[solderdude]

some people lament lacking soundstage with low power.

Soundstage is something that is recreated in the brain. It depends on many factors. A channel separation of 30-40dB is enough.
Phase response between channels will be similar and with the vast majority of amps will never be a sharp change in phase.
So it is extremely unlikely this has anything to do with the used electronics (despite people reporting this) but everything to do with the recording itself, properties of the headphone and what the brain can (re)construct from the signals.
perhaps when amps start clipping this may confuse things but 10mW from a phone or from a 1W amp will not change anything but the brain.
Assuming, again, THD, output R, frequency response and phase response are up to snuff.

 

[Zensō]

Again... there is no increased demand for power. Amps do not 'push' power nor 'grab' drivers and certainly not demand more power than dictated by the voltage present.
As Otaku+ already mentioned the output of a music source provides a voltage. That is not a constant voltage but varies. That voltage has a limit. That limit depends on the design of the amp.
The maximum voltage also depends on the current capabilities of the same amplifier.

So when either the voltage reaches its limit OR the current reaches its limit (both have a limit) and it depends on what limit is reached first the output voltage clips. Yes, the output voltage can also clip because a current limit is reached and that can happen before the voltage limit is reached.

The lower the voltage limit and the lower the current limit and the lower the impedance of the headphone and the lower the voltage efficiency of the headphone the earlier the device 'clips' meaning the output voltage clips meaning the SPL is limited.

When this happens and you turn up the volume beyond that point (depending on how 'hard' the clipping is) the loudness (volume) of the sound will still increase and bass levels will 'suffer' first as they have the largest voltage swing. The sound will turn 'nastier' but the volume of the sound keeps increasing but NOT the bass. That is clipped so stays 'behind'.

In these cases using an amplifier that has a higher voltage and or current limit moves the voltage clipping point upwards and so the clipping occurs at higher listening levels.

At some point, where peak SPL of 120dB is reached, the sound is so loud you won't go there simply because it is unpleasant for longer periods.
You can listen to 120dB SPL peaks for about a minute and this is impressively loud and because of equal loudness contours has tremendous bass power (makes Amirs earlobes shake). But you won't be doing that for long.

This means that when the amp chosen for driving your headphone can deliver the needed voltage (= drawn power, not delivered power) without clipping and the amplifier has enough gain and the music source has enough output voltage then having more available power is pointless.

Yes, having 3dB more headroom is desirable but not needed for normal listening levels.

When amplifiers (even a dongle has an amplifier inside) clips at SPL levels of say 90dB SPL in a specific headphone then it will sound as if it struggles with the bass. And it does. It clips the bass as that has the highest amplitude.
90dB SPL for bass is NOT loud at all. 90dB of bass is about 30dB above the hearing threshold.
60dB SPL of bass makes the fundamental of that bass inaudible already while the rest of the music is at a enjoyable soft background level.
90dBA is loud and for bass fundamentals to reach that level the actual SPL = 120dB !!

So in short... when you want to use your headphone at all kinds of levels it needs to reach voltage levels that can make it produce 120 to 123dB SPL. You don't need as much if you never play loud. Around 110dB is good for comfortable loud and 100dB ensures normal listening levels never clip.

That said.. efficiency levels are often given at 1kHz or 500Hz. Looking at the frequency response of headphones you will find most of them roll-off and some 'boost' in the lows. This too has an effect of how bass is perceived. When you have to boost it the voltage limit is reached much, much, much earlier as well.

It is a complicated 'play' between headphones, voltage efficiency and voltage levels that can be reached which determine how impressive bass will be.
It is not costly to make amps that can provide enough to power all but very rare and inefficient headphones.

There you have it. It's all science not magic and the laws of that science cannot be broken.
Audio may seem like magic and things may seem 'inexplicable' but they aren't.

Aside from all these technical things that need to be right, including distortion and frequency response and in lesser sense phase, there are other aspect that also have great impact.

Seal of the headphone
Bass extension and frequency response
recording itself
dynamic range (loudness compression) also a recording thing
Hearing sensitivity (changes over the years, you don't notice this till it is really severe)
Taste (think bass heads)
The brain itself is also a major part of perception.

Nope... it is not simple and science is a part of it. Only that part is measured.

@solderdude I just wanted to thank you for your never-ending willingness to educate those of us who are less technically fluent. Your level of patience is really quite remarkable.