• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Stax Tube Amplifier Distortion vs Solid State

YSC

Major Contributor
Joined
Dec 31, 2019
Messages
1,252
Likes
849
#21
You don't listen to continuous test tones. Momentary peaks especially in low frequencies are not loud at all. Threshold of hearing at 20 Hz is 70 dBSPL. If you are listening at 65 dB, you are not hearing anything at that frequency let alone it being loud!

Really folks, this whole dBSPL thing is so misunderstood it seems. What did you measure with it anyway? Noise? We are not talking about noise.
Sorry I don't mean to say you would never listen to anything at 112dB SPL, just saying when peak level at any frequency is at 112dB one likely would listen to much lower SPL, my measurement was just using the phone with dB meter app while I play some music so no steady tones for that sake. just wanna say that at music generally listening level should be well below 112dB SPL so should not be the limiting factor for almost all usecase
 

phoenixdogfan

Major Contributor
Joined
Nov 6, 2018
Messages
1,373
Likes
1,272
Location
Nashville
#22
You lose me here. Normally I would seldom listen to steady state music at even 90db. How much louder would the peaks go for that music? Over 112 db? Over 114 db? It looks to me like the solid state provides around 2.6 db of additional headroom which is quite a bit, but in terms of driving a speaker with an amp, wouldn't this like going from 600wpc to 1000wpc? Just don't see why either headphone amp would not be adequate for most users' needs, and I am most definitely not a tube guy.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,396
Likes
11,853
Location
Riverview FL
#23
My rule of thumb is to add 20dB for peaks over average.

Example right now:

Gentle piano, bass, and saxophone

Carla Bley - Life Goes On

1609138318295.png



An RTA from the disc over the same period seems to indicate a broad range of frequencies with roughly equal peak power.

1609138462330.png


While the piano and bass do not hold peaks, the saxophone can from time to time, which could excite audible distortions at higher listening levels.
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
33,645
Likes
117,741
Location
Seattle Area
Thread Starter #24
Sorry I don't mean to say you would never listen to anything at 112dB SPL, just saying when peak level at any frequency is at 112dB one likely would listen to much lower SPL, my measurement was just using the phone with dB meter app while I play some music so no steady tones for that sake.
Your meter is not fast acting enough to give you proper peak rates. And it is not calibrated to boot. Really, there is no correlation between what you think the SPLs are and what I am reporting with calibrated instruments.

I remember being at an audio show where Andrew Jones was demoing his speakers. Someone asked him what level he was playing at. He asked the audience to provide guesses. People were throwing numbers like 85 or 90 dB. At the end he said that the peaks are likely over 105 dB or something like that.

As I said, there seems to be massive confusion regarding what levels you all listen to. Either that you are not representing the rest of us who using the power and dynamics we have to get realistic sound levels.
 

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
223
Location
Beaverton, OR
#25
You lose me here. Normally I would seldom listen to steady state music at even 90db. How much louder would the peaks go for that music? Over 112 db? Over 114 db? It looks to me like the solid state provides around 2.6 db of additional headroom which is quite a bit, but in terms of driving a speaker with an amp, wouldn't this like going from 600wpc to 1000wpc?
Transients can be 20db+ louder than the average listening volume and will get clipped off if there is not enough headroom for them. Bass also has much higher power requirements compared to mids for the same SPL.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
33,645
Likes
117,741
Location
Seattle Area
Thread Starter #26
You lose me here. Normally I would seldom listen to steady state music at even 90db. How much louder would the peaks go for that music? Over 112 db? Over 114 db? It looks to me like the solid state provides around 2.6 db of additional headroom which is quite a bit, but in terms of driving a speaker with an amp, wouldn't this like going from 600wpc to 1000wpc? Just don't see why either headphone amp would not be adequate for most users' needs, and I am most definitely not a tube guy.
The difference is not 112 vs 114. As I have explained now multiple times, distortion from the amp rises well before that. It is just hidden in that graph. I can tell you that the difference is plainly audible and neither amp is able to get ear bleeding levels driving these headphones.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,396
Likes
11,853
Location
Riverview FL
#27
in terms of driving a speaker with an amp, wouldn't this like going from 600wpc to 1000wpc?
Yes, the amplifier would need to output a few more volts.

Power is the square of voltage divided by impedance, so those few volts difference can make an impressive difference in watts, at higher voltage levels.

1609139914940.png
 
Last edited:

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,396
Likes
11,853
Location
Riverview FL
#28
Bass also has much higher power requirements compared to mids for the same SPL.
But the recording above does not show higher voltage for the bass (few do from what I've watched), so, if the impedance is the same, how does bass require more power?

I've never been able to figure this out.

I'm almost ready to tag this as "Urban Legend".

---

Here's another disc, just next in line here, no cherry picking, rock type band

1609139664009.png
 
Last edited:

the_brunx

Active Member
Joined
Apr 7, 2020
Messages
168
Likes
368
#29
Add eq to the mix and things will get from bad to unbearable real quick. I love electrostatic headphones but never found an amp that totally cuts it with this problem. It’s the reason to buy a SRD7 mk2 transformer and use with a very low distortion and powerful speaker amp.
 

pma

Major Contributor
Forum Donor
Joined
Feb 23, 2019
Messages
1,726
Likes
2,790
Location
Prague
#30
Your meter is not fast acting enough to give you proper peak rates. And it is not calibrated to boot. Really, there is no correlation between what you think the SPLs are and what I am reporting with calibrated instruments.

I remember being at an audio show where Andrew Jones was demoing his speakers. Someone asked him what level he was playing at. He asked the audience to provide guesses. People were throwing numbers like 85 or 90 dB. At the end he said that the peaks are likely over 105 dB or something like that.

As I said, there seems to be massive confusion regarding what levels you all listen to. Either that you are not representing the rest of us who using the power and dynamics we have to get realistic sound levels.
When I am in doubt what is the peak level I put the digital scope at the amp output terminals. This displays and stores wave shape and thus peaks accurately. Speaker sensitivity is known so it is possible to estimate peak SPL. The second possibility is calibrated microphone + PC.
 

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
223
Location
Beaverton, OR
#31
But the recording above does not show higher voltage for the bass (few do from what I've watched), so, if the impedance is the same, how does bass require more power?

I've never been able to figure this out.

I'm almost ready to tag this as "Urban Legend".

---

Here's another disc, just next in line here, no cherry picking, rock type band

View attachment 101998
Impedance of an estat is not linear, by far. It is essentially a capacitor, electrically. I've also personally experienced hard clipping in bass on energizers at listenable volumes, where the mids exhibited no such clipping, so it's definitely not something I'm just theorizing.

Plus, even on planar drivers with a flat impedance curve, there are still momentary differences in electrical characteristics as the driver excurses, which can make the difference between following through on a transient or falling flat. Ask anyone who has tried something like an HE6 from a speaker amp. Suddenly the bass starts to slam much harder even though FR is the same. This is not something you'll be able to measure with a sine wave, because a constant tone already establishes some of the required momentum and makes the amp's job considerably easier.
 

Maki

Active Member
Joined
Oct 24, 2018
Messages
223
Likes
380
#32
Impedance of an estat is not linear, by far. It is essentially a capacitor, electrically. I've also personally experienced hard clipping in bass on energizers at listenable volumes, where the mids exhibited no such clipping, so it's definitely not something I'm just theorizing.

Plus, even on planar drivers with a flat impedance curve, there are still momentary differences in electrical characteristics as the driver excurses, which can make the difference between following through on a transient or falling flat. Ask anyone who has tried something like an HE6 from a speaker amp. Suddenly the bass starts to slam much harder even though FR is the same. This is not something you'll be able to measure with a sine wave, because a constant tone already establishes some of the required momentum and makes the amp's job considerably easier.
Can we measure that? I've got an HE6 and a miniDSP EARS right here. I can make a recording of starting a 20hz tone from my A90 and a much less powerful amp, I guess the Lotoo PAW S1 would work. Then we can inspect the behavior at the start of the test tone. That should give insight into this alleged phenomenon, no? Hopefully the EARS isn't limited in this respect, even though obviously the FR measurements it produces are only slightly better than nothing.
 

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
223
Location
Beaverton, OR
#33
Can we measure that? I've got an HE6 and a miniDSP EARS right here. I can make a recording of starting a 20hz tone from my A90 and a much less powerful amp, I guess the Lotoo PAW S1 would work. Then we can inspect the behavior at the start of the test tone. That should give insight into this alleged phenomenon, no? Hopefully the EARS isn't limited in this respect, even though obviously the FR measurements it produces are only slightly better than nothing.
Yes, that could be quite interesting. Maybe generate a short 20hz tone (or even just a few cycles) in audacity and see what the result is across different amps. See how it correlates to audible differences in music as well. Accuracy *might* be a problem tho since this would require a very high resolution mic.
 
Last edited:

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,512
Likes
1,583
Location
Iasi, RO
#34
Yes, the amplifier would need to output a few more volts.

Power is the square of voltage divided by impedance, so those few volts difference can make an impressive difference in watts, at higher voltage levels.

View attachment 102000
The graphs is very representative and a good information to get an idea about Watts vs. SPL.

For 85 dB/W/m I know an easier to remember "formula" (listener in front of one single speaker, at 1 meter distance):
- 85 dB @ 1W
- 95 dB @ 10 W
- 105 dB @ 100 W
- 115 dB @ 1000 W
Adding a second speaker will add another 3 dB to the chart.
Doubling the distance will subtract 6 dB from the chart.
Based on the above, with two such speakers, at 2.5 meters in front of them and a power amp delivering 2 x 100 W/ch. that would be exactly 100 dB SPL for the listener. Below the THX standard, but definitely enough power to enjoy the music.

However, I see that you got 112 dB SPL at 1000 Watts and 115 dB at 2000 Watts, not sure why, maybe the dB SPL values are shifted with one row. Based on Sound pressure level and amplifier power I got 118 dB SPL for 2000 Watts and 121 dB SPL for 4000 Watts, applied to one single 4-Ohms speaker at 1 m in front of it. Wondering where's the catch.
 

Maki

Active Member
Joined
Oct 24, 2018
Messages
223
Likes
380
#35
Yes, that could be quite interesting. Maybe generate a short 20hz tone (or even just a few cycles) in audacity and see what the result is across different amps. See how it correlates to audible differences in music as well. Accuracy shouldn't be too problematic in bass regions.
I'll get to that hopefully tomorrow (it's getting late where I am). I'll create a new thread since it's not really relevant to this one, and tag you in it.

To get back on topic, I've read lower frequencies are the easiest part to drive for an electrostat given the headphone is a capacitive load.
 

RayDunzl

Major Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
11,396
Likes
11,853
Location
Riverview FL
#37
However, I see that you got 112 dB SPL at 1000 Watts and 115 dB at 2000 Watts, not sure why, maybe the dB SPL values are shifted with one row. Based on Sound pressure level and amplifier power I got 118 dB SPL for 2000 Watts and 121 dB SPL for 4000 Watts, applied to one single 4-Ohms speaker at 1 m in front of it. Wondering where's the catch.
My chart was 4 ohms. the (I think) "standard" 2.83V signal yields 2 watts at the speaker via increased current.

If my chart is 8 ohms, it shifts a row, with 1 watt as the power at the 2.83V sensitivity..

1609142494292.png


It's not meant to be anything more than a general example of "louder requires more power and much louder requires much more power".
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
9,113
Likes
19,440
Location
The Neverlands
#38
How loud is +112 dBSPL?
It is playing uncomfortably loud. It is the type of level where you turn up the volume because you either like the song very much or want to listen ‘deep’ in the recording to evaluate. Not a level you would gladly endure for the whole song.

113dB peaks (DR 16 recording) has an average level of 97dB SPL (not Phon that may be lower)

Note that if one EQ's the lowest frequencies up by a few dB (some prefer Harman lows and they will have to) the average level will drop as well. Add to that clipping of amps is more audible than the soft clipping of drivers is (except for some Focal HP's) this will make sounds distort sooner than one thinks.

I such EQ'ed state one may be looking at 90dB average with 104dB peaks where bass reaches 112dB peaks which is comfortably loud level which is easy to listen to for 1 or 2 songs before getting the urge to turn it down to more comfortable levels.
 
Last edited:

Degru

Active Member
Joined
Feb 19, 2019
Messages
230
Likes
223
Location
Beaverton, OR
#39
I'll get to that hopefully tomorrow (it's getting late where I am). I'll create a new thread since it's not really relevant to this one, and tag you in it.

To get back on topic, I've read lower frequencies are the easiest part to drive for an electrostat given the headphone is a capacitive load.
In practice, lower frequencies are the hardest part of driving stax and the biggest area of improvement after moving to a transformer setup. I don't know enough to confidently describe exactly why tho, aside from needing more current and energizers having trouble providing it.

As for the measurement test, I am actually having doubts about whether EARS would be sufficient since you would want to record at a high sample rate with a very high resolution mic to capture the behavior of the driver accurately. I was thinking accuracy in terms of FR versus a professional HATS, which is a different thing. Basic measurement mics like EARS are generally made to be "good enough", and EARS only supports 48khz afaik. But it's worth a shot I suppose if you hear a noticeable difference in bass impact switching between for example A90 and apple dongle.
 
Last edited:

Maki

Active Member
Joined
Oct 24, 2018
Messages
223
Likes
380
#40
In practice, lower frequencies are the hardest part of driving stax and the biggest area of improvement after moving to a transformer setup. I don't know enough to confidently describe exactly why tho.
As for the measurement test, I am actually having doubts about whether EARS would be sufficient since you would want to record at a high sample rate with a very high resolution mic to capture the behavior of the driver accurately. Basic measurement mics like EARS are generally made to be "good enough".
I've never had any trouble from my STAX with the low frequencies. You may have heard clipping in the lows because lower frequencies have much higher representation in a lot of music than mids/highs. Plus, test tones at lower frequencies are much more tolerable. I cringe at the thought of a 90dB 1KHz sine wave. A 30hz sine wave at the same SPL? Not that bad.

If you're saturating the transformer core you may be getting additional harmonics in the bass frequencies leading to more bass. I've never seen any measurements of what it takes to saturate a SRD-7 or similar converter box, nor measurements of the THD the transformer itself produces so the jury is still out on that one.

The EARS does 24bit/48KHz. That should be enough for bass frequencies at least. We'd get just over 2,000 samples per cycle at 20hz, 200 samples/cycle at 200hz if I did my math correct. That should be more than enough precision on the sampling side of things, though if there are any other issues please do bring them up.
 

Similar threads

Top Bottom