Skip to main content
Topic: Sound Cards (Read 9561 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Sound Cards

Hey,

Don't know if you guys saw the TekSyndicate video on audio.
https://www.youtube.com/watch?v=bjTxEwlypA0
https://www.youtube.com/watch?v=d1rXcJuEsy0

There it is claimed that a sound card is basically useless. I mean, for starters sound cards have features onboard might not have. Onboard also varies from motherboard to motherboard. I think it's a bit ridiculous to say that all onboard will not see an upgrade in sound quality once one gets a sound card. I have an old 2008 computer which had pretty bad onboard. I think a $30 sound card can be justified, because it skips the onboard if it's too crappy. In a Youtube conversation Logan from the video mentions that he feels sound cards have high output impedance and that it colors his sound. To me, it's odd that Logan claims to be able to hear and get bothered by high output impedance while at the same time say he cannot ever hear an improvement in sound quality going from onboard to a sound card. I mean, wouldn't a sound card have a lower output impedance than an onboard solution (if the sound card is good)? Wouldn't that alone be enough to hear a difference from onboard vs sound card? A sound card could have a lower noise floor... Although, that's one of the "either it's a problem or it's not" things. Like having cracks in your audio due to crappy onboard, either it's there and it bothers the heck out of you, or it's not audible. And SNR... SNR is another way of displaying total distortion... I am under the impression that a regular <$10 onboard dac chip is audibly transparent in the distortion category. And then shielding... Does that even matter?

And then we have the Tom's article where the people there failed to tell Onboard from Objective dac/amp with HD800s!
There is so much contradiction in the audio world it's insane.

And on impedance... An HD800 is rated to have 300omh impedance. Measurement charts show again and again that this isn't the case, that the impedance varies up to like 530ohms depending on the frequency. I mean, that's getting close to double the impedance listed... That's kind of misleading. And so, I hear people saying things like 'This amp is good for up to 300ohms'... Which, first of all, I'm not sure if it even works that way. But ok, when I state that the HD800s vary in impedance all over the place, it is pointed out that a higher impedance requires less current is therefore easier to drive. I still don't get it. To my knowledge, you have current and voltage. In general, excluding the oddball headphones, the ones that require high current are the low impedance headphones and the ones that require high voltage are the higher impedance headphones. Clearly in either end of the extreme, it is problematic for the amp. So a headphone that is struggling to "drive" a 300ohm headphone will have issues when the headphone wants 500+ ohms at 200hz, no? So there will be problems.

Another question: Yea, I know that some people can hear the difference between lossless and a lossy file. But some people are telling me their non-audiophile girlfriend can easily tell lossy from non-lossy apart on some budget headphones. I'm skeptical. In your experience, how often do you run into a person that actually can hear a difference between lossy and lossless?

And while I'm piling on the questions here, my opinion is that balanced audio doesn't produce better audio, especially not in a way that is audible. Do you agree?

There has been a significant backlash against Logan's two videos on TekSyndicate. Lots of facepalms were had. And I personally can spot more than one error in their video. But a lot of posts were simple facepalm, rage posts or a jab here or there, criticism here and there. No exhaustive list detailing every single error in the video. Which annoys me. Now depending on which forum you go to, people have different philosophies... Some are at the 'onboard all the way' crowd, some are anti-sound cards, some are tube lovers... I mean, with tubes you can blind test them easily, right? That's the entire freakin' point of tubes, to distort the sound in a way the person finds pleasurable. Lots of differences in opinion. I'm always curious about what Hydrogenaudio thinks about this though.

Sound Cards

Reply #1
There it is claimed that a sound card is basically useless. I mean, for starters sound cards have features onboard might not have. Onboard also varies from motherboard to motherboard.


Yeah it depends on your motherboard and what you are doing.  But for most people, yes its fairly useless on modern hardware, particularly if you aren't recording. 

In a Youtube conversation Logan from the video mentions that he feels sound cards have high output impedance and that it colors his sound. To me, it's odd that Logan claims to be able to hear and get bothered by high output impedance while at the same time say he cannot ever hear an improvement in sound quality going from onboard to a sound card.


He has a good point.  Surprisingly many sound cards have terrible headphone out. 

I mean, wouldn't a sound card have a lower output impedance than an onboard solution (if the sound card is good)?


You would think, but a lot of sound cards aren't made with headphones in mind, so its not unheard of for them to perform worse than good onboard. 


And on impedance... An HD800 is rated to have 300omh impedance. Measurement charts show again and again that this isn't the case, that the impedance varies up to like 530ohms depending on the frequency.


The rated impedance is usually for 1 KHz IIRC.  Since impedance is by definition a function of frequency, it will not be constant.

I mean, that's getting close to double the impedance listed... That's kind of misleading.


No, thats how headphones work...

And so, I hear people saying things like 'This amp is good for up to 300ohms'... Which, first of all, I'm not sure if it even works that way.


No it doesn't work that way.  Higher impedance means less load => easier to drive but lower volume.  An amp won't have problems with higher impedance per se, but rather higher impedance headphones will tend to need higher voltages (but not always!).  Any amp will drive high impedance headphones, jsut maybe not with as much volume as you want.  Its more useful to look at the sensitivity (in dB/volt) because that tells you what you really want to know (how much voltage for a given loudness).

Sound Cards

Reply #2
To me, it's odd that Logan claims to be able to hear and get bothered by high output impedance while at the same time say he cannot ever hear an improvement in sound quality going from onboard to a sound card.


High output impedance can cause frequency response variations as high as 6 dB or more; it can roll-off bass or treble so badly as to make them disappear altogether. It can also reduce the output voltage (i.e. volume into the headphones) by similarly large amounts. And it will attenuate or eliminate headphone damping, resulting in less punchy, more rumbly / muddy bass.

Conversely, the difference in frequency response linearity, SNR and distortion between a discrete sound card and onboard audio, can be the difference between "total overkill" and "well beyond audibility" (i.e. irrelevant, as far as playback is concerned).

Here are my RMAA measurements for my EMU 0204 USB ("total overkill") and my Realtek ALC663 onboard audio ("well beyond audibility"). Both suffer from high output impedance, unfortunately (22 and 75 ohms, respectively, reducing volume into my 25? Denon headphones by 5.5 dB and 12 dB respectively).

Sound Cards

Reply #3
The ALC663 has more periodic passband ripple than I'd like to see though - spec is +/-0.02 dB, which means a pre-echo less than about 60 dB down. This, however, is an older chip; modern Realteks tend to be at +/-0.0005 dB or so on the playback side. So as stated, these ought to be just fine, with headphone out characteristics being the dominant factor - and you can upgrade those with an external amp if needed.

Things are looking less good on the recording side, where you average standard quality chip will have +/-0.03 dB of passband ripple (since low ripple + low latency = high computing effort), distortion tends to be OK but unexciting, and even a "low noise bias supply" for the mic tends not to be very low noise at all (though admittedly better when compared to chips with a standard-quality bias supply, which qualifies as terribly noisy).

Sound Cards

Reply #4
PS: Note that there are some scenarios in which significant amounts of dynamic range - or rather a low noise level - are required. One would be direct connection to a pure power amplifier, possibly with speakers of the more sensitive kind.

For example, let the speakers have a nominal sensitivity of 90 dB SPL / 2.83 V / m (which is somewhat higher than average but by no means outlandish), and the power amplifier have a gain of 26 dB. Target noise level be 0 dB SPL @ 1 m anechoic (which is 3 dB from two speakers!).
Then amplifier noise output should be no bigger than 89.5 µV, or noise input no bigger than 4.5 µV.
That's -108 dB re: 1.1 Vrms or -114 dB re: 2.2 Vrms, requiring a corresponding amount of DAC dynamic range if there is no analog (e.g. PGA) volume control. And you definitely don't want any ground loop to be running across your interconnect.
Of course you could reduce requirements by including a line-level attenuator and sacrificing some maximum level - or upgrade to the fancy version involving a preamp with a volume control of its own for better dynamic range management.

Sound Cards

Reply #5
PS: Note that there are some scenarios in which significant amounts of dynamic range - or rather a low noise level - are required. One would be direct connection to a pure power amplifier, possibly with speakers of the more sensitive kind.

For example, let the speakers have a nominal sensitivity of 90 dB SPL / 2.83 V / m (which is somewhat higher than average but by no means outlandish), and the power amplifier have a gain of 26 dB. Target noise level be 0 dB SPL @ 1 m anechoic (which is 3 dB from two speakers!).


Back in the day JJ told us about the 0 dB SPL targeted listening lab that they built at AT&T labs. Their budget was huge.  One major stumbling block was a freeway maybe a half mile away. I think they actually were able to actually set up a room that was like NC 10 which is still way above 0 dB SPL.



Sound Cards

Reply #6
Hey,


And while I'm piling on the questions here, my opinion is that balanced audio doesn't produce better audio, especially not in a way that is audible. Do you agree?


As to the sound card, my inboard was quite noisy (I believe it was from the outlandishly large graphics card), an outboard solved it immediately while offering me a very nice feature set (more outputs!)  I can't tell the difference from the headphone out on the card to any of my DACs.

I am interested in the quoted question above though...what do you imply by saying "balanced audio"?  Do you mean balanced cabling?  Balanced FR?

Sound Cards

Reply #7
Hey,


And while I'm piling on the questions here, my opinion is that balanced audio doesn't produce better audio, especially not in a way that is audible. Do you agree?


As to the sound card, my inboard was quite noisy (I believe it was from the outlandishly large graphics card), an outboard solved it immediately while offering me a very nice feature set (more outputs!)  I can't tell the difference from the headphone out on the card to any of my DACs.

I am interested in the quoted question above though...what do you imply by saying "balanced audio"?  Do you mean balanced cabling?  Balanced FR?

I think a few headphones have output that is typically XLR.
I think for headphones that don't fit in that category people will just get custom cables that is balanced to plug into a dac/amp that has balanced plugs/

Sound Cards

Reply #8
I think a few headphones have output that is typically XLR.
I think for headphones that don't fit in that category people will just get custom cables that is balanced to plug into a dac/amp that has balanced plugs/


I thought as much, thanks for clarifying.

I've heard balanced a couple times, but I've never heard the same gear balanced/unbalanced

My basic thoughts are balanced is only terribly useful for long runs, but it have no backup.  It would be interesting to hear what any others have to say on balanced headphones.  My understanding is it's used primarily for noise rejection...

Sound Cards

Reply #9
Balanced headphone driving is not generally necessary, assuming total shared ground return impedance does not exceed about 1% of headphone driver nominal impedance as a rule of thumb. That's usually the case if the headphone manufacturer bothered to include separate ground returns for both drivers in the cable, otherwise this probably is a dominant contributor (I have some 16 ohm in-ears where shared resistance computes to over 1 ohm, which would have to be counteracted by reducing (L-R) amplitude, i.e. stereo width, to about 85-90% - thankfully it's a dynamic driver model with almost constant impedance). I also imagine a HDA front panel connector cable may contribute several tenths of an ohm. Finally the headphone jack and internal wiring and PCB tracks also tend to contribute a few dozen milliohms.

The value of balanced connections is in analog interconnects. As anyone who's ever had to fight a ground loop will be able to attest, you can get in trouble with unbalanced connections very easily. It's not like this can't happen in balanced (see "Pin 1 Problem"), not to mention some other drawbacks such as potentially higher noise - but when devices are built correctly, balanced can be a lot more robust.
Back in the day JJ told us about the 0 dB SPL targeted listening lab that they built at AT&T labs. Their budget was huge.  One major stumbling block was a freeway maybe a half mile away. I think they actually were able to actually set up a room that was like NC 10 which is still way above 0 dB SPL.

Actually you don't need a super quiet room to shoot for a 0 dB noise target.
For one thing, 0 dB at 1 m anechoic (which tends to be about the same at 2-3 m in typical rooms) is the level at listening position - stand close to the speaker, and it'll be a fair bit higher. Some people do walk around in their rooms at times and do not appreciate being greeted by hiss. Ideally I wouldn't want to hear much of anything at 20 cm or so. My nearfields (granted, they're less sensitive) only are about 50 cm away here.
And then we are basically talking about white noise - which has a power spectrum quite unlike the remaining noises in a reasonably quiet room, where high-frequency components tend to be pretty much absent. So the high-frequency hiss is not being masked well at all.

I have an amp + speaker combo that ought to produce about a 7 dB noise level @ 1 m anechoic (4558 opamp with sensible resistor values + 45.5 dB total gain = ~250 µV @20k, + 88 dB / 2.83 V / m), and the hiss is by no means terribly intrusive but still audible up to at least 80 cm from the speaker if you don't stray too far from on-axis listening at about tweeter level. (And that's with the computer running, which is a fairly quiet office model but by no means ultra-silent.) Hence a 0 dB target is not something I would consider particularly outlandish.

Yes, asking the DAC to handle the entire output-side dynamic range may be considered "brute force", but things like that do occur and have their uses in real life.

Incidentally, onboard solutions in fancy gaming boards (e.g. Asus) come with SNR claims as high as 120 dB these days, and HDA codecs claim as much as 115 dB when using differential output (see ALC1150). How much of this is usable in real life remains to be seen. (The ALC1150 datasheet hints at a few gotchas - only the front outputs benefit from the good DACs, while others including the headphone out have to make do with 96 dB - I hope they include a PGA -, and THD+N specs have more to do with their standard quality parts as well, in spite of significant analog current consumption.)

Sound Cards

Reply #10
Balanced headphone driving is not generally necessary, assuming total shared ground return impedance does not exceed about 1% of headphone driver nominal impedance as a rule of thumb. That's usually the case if the headphone manufacturer bothered to include separate ground returns for both drivers in the cable, otherwise this probably is a dominant contributor (I have some 16 ohm in-ears where shared resistance computes to over 1 ohm, which would have to be counteracted by reducing (L-R) amplitude, i.e. stereo width, to about 85-90% - thankfully it's a dynamic driver model with almost constant impedance). I also imagine a HDA front panel connector cable may contribute several tenths of an ohm. Finally the headphone jack and internal wiring and PCB tracks also tend to contribute a few dozen milliohms.

The value of balanced connections is in analog interconnects. As anyone who's ever had to fight a ground loop will be able to attest, you can get in trouble with unbalanced connections very easily. It's not like this can't happen in balanced (see "Pin 1 Problem"), not to mention some other drawbacks such as potentially higher noise - but when devices are built correctly, balanced can be a lot more robust.
Back in the day JJ told us about the 0 dB SPL targeted listening lab that they built at AT&T labs. Their budget was huge.  One major stumbling block was a freeway maybe a half mile away. I think they actually were able to actually set up a room that was like NC 10 which is still way above 0 dB SPL.

Actually you don't need a super quiet room to shoot for a 0 dB noise target.
For one thing, 0 dB at 1 m anechoic (which tends to be about the same at 2-3 m in typical rooms) is the level at listening position - stand close to the speaker, and it'll be a fair bit higher. Some people do walk around in their rooms at times and do not appreciate being greeted by hiss. Ideally I wouldn't want to hear much of anything at 20 cm or so. My nearfields (granted, they're less sensitive) only are about 50 cm away here.
And then we are basically talking about white noise - which has a power spectrum quite unlike the remaining noises in a reasonably quiet room, where high-frequency components tend to be pretty much absent. So the high-frequency hiss is not being masked well at all.

I have an amp + speaker combo that ought to produce about a 7 dB noise level @ 1 m anechoic (4558 opamp with sensible resistor values + 45.5 dB total gain = ~250 µV @20k, + 88 dB / 2.83 V / m), and the hiss is by no means terribly intrusive but still audible up to at least 80 cm from the speaker if you don't stray too far from on-axis listening at about tweeter level. (And that's with the computer running, which is a fairly quiet office model but by no means ultra-silent.) Hence a 0 dB target is not something I would consider particularly outlandish.

Yes, asking the DAC to handle the entire output-side dynamic range may be considered "brute force", but things like that do occur and have their uses in real life.

Incidentally, onboard solutions in fancy gaming boards (e.g. Asus) come with SNR claims as high as 120 dB these days, and HDA codecs claim as much as 115 dB when using differential output (see ALC1150). How much of this is usable in real life remains to be seen. (The ALC1150 datasheet hints at a few gotchas - only the front outputs benefit from the good DACs, while others including the headphone out have to make do with 96 dB - I hope they include a PGA -, and THD+N specs have more to do with their standard quality parts as well, in spite of significant analog current consumption.)

So in general, which probably benefits more from balanced connections, in ears or headphones?
Although the ASUS boards probably have decent audio, it is obvious their measurements are overblown. Based on the way they phrase SNR values and other measurements in their ASUS STX sound card manual, it looks like it's surpassing the Objective dac/amp. Nwavguy noted this in his blog. I am actually using the Objective right now.

So:
Balanced can mean higher noise but when implemented correctly tends to fix any ground loop problems and have lower noise? How about in a typical audiophile setup... Say, Audeze LCD 2, Sennheiser HD800s, with... Objective amp/dac some more expensive Schiit product... I mean, I don't think it will sound different going with balanced vs unbalanced. I don't think there is enough noise to be audible in these setups.

A balanced connection might help my Krk Rokits though. They have a small hissing problem. They had a ground loop problem, but then I stuck it through a ground loop isolator.

Sound Cards

Reply #11
So in general, which probably benefits more from balanced connections, in ears or headphones?
Although the ASUS boards probably have decent audio, it is obvious their measurements are overblown. Based on the way they phrase SNR values and other measurements in their ASUS STX sound card manual, it looks like it's surpassing the Objective dac/amp. Nwavguy noted this in his blog. I am actually using the Objective right now.

I wouldn't consider it impossible that the ST(X) actually does come close to its specs on a good day and with a bit of a following wind, but it would seem that the figures are easily degraded in practice. At this performance level, running loopback tests with unbalanced connections tends to be pretty hopeless, too - usually too much ground loop garbage. (Once two different slots are involved, it tends to become entirely hopeless, as you can tell by the results for a 2-slot Creative ZxR in a thread here.)

Comparing the STX to an O2/ODAC combo is a bit apples to oranges. The Asus card's headphone out relies on "brute force" in the form of high DAC SNR to keep noise down, much like a setup that connects a line-level output directly to a speaker power amplifier. Now if SNR drops to a "measly" 108 dB in 44.1 kHz (which uses a noisier PLL-generated clock), people with sufficiently sensitive cans start complaining about hiss. 108 dB isn't that great if your full-scale amplitude is 7 Vrms - it gives 28 µV worth of noise (-91 dBV), which is run of the mill headphone amp level. Granted, nothing that your average HiFi AKBeyerheiser (let alone isodynamics) will care much about, but most anything at 120 dB SPL/V up should make it plainly audible. You may be surprised to find what kind of total dynamic range may be required at a headphone or speaker power amplifier output with a range of different loads!
The O2 with its integrated volume control is a lot less critical when it comes to DAC SNR. The volume control allows placing the 70 dB or so "window" of instantaneous dynamic range that human listeners typically need somewhere within DAC dynamic range. So even with relatively sensitive cans, it does not need to be stuck at the very bottom, and you may find that even a modest 90 dB of DAC SNR gets the job done fine. Likewise, the sensitivity to ground loops running over the interconnect potentially is much reduced as well when you don't have to be running tiny levels.
(Obviously, getting a real-life 111 dB of DAC SNR from a relatively modest ES9023 on USB in the ODAC is more impressive than 114 dB with the fancier parts of the STX.)

So:
Balanced can mean higher noise but when implemented correctly tends to fix any ground loop problems and have lower noise? How about in a typical audiophile setup... Say, Audeze LCD 2, Sennheiser HD800s, with... Objective amp/dac some more expensive Schiit product... I mean, I don't think it will sound different going with balanced vs unbalanced. I don't think there is enough noise to be audible in these setups.

Headphone amps usually are built as IEC Class II devices (double insulated, with floating audio ground) and as such tend to be fairly immune to ground loop issues involving their source connection. Some hollow state gear is Class I though and in addition does connect its audio ground to safety earth (as does the PC) - uncomfortably high levels of ground loop noises are possible there. If you have to turn down volume so much that maximum levels are severely limited, it's definitely time to do something about them.

The wanted signal obviously should not sound any different, assuming the balanced input stage isn't a distortion generator in disguise (which normally it in fact is not).

A balanced connection might help my Krk Rokits though. They have a small hissing problem. They had a ground loop problem, but then I stuck it through a ground loop isolator.

If the hiss remains with input gain turned all the way down and no source connected, it's an inherent problem related to internal gain staging. The designers may have preferred to go easy on crossover levels to keep distortion down, or used ye olde TL07x series opamps for cost reasons... or maybe it's just a stupid design. It is unfortunately quite common for inexpensive active speakers to be hissy. It really shouldn't have to be like that.

My Tascams here have very well-controlled hiss levels (inaudible at my level setting), but unfortunately build quality wasn't up to par - one has had woofer issues at times (apparently there was a run with faulty relays), and the other has a bit of a buzz indicating a badly-soldered / duff filter capacitor or somesuch. ;-/ Too bad opening them can only be described as a royal pain in the rear end (the back panel is held in place by a gazillion screws and glue, grrr).

While I had been able to turn down input gain far enough to make ground loop noises disappear even with an unbalanced connection, I eventually splurged for a line-level transformer (for galvanic isolation and unbalanced/balanced conversion) when upgrading the none-too-reliable cabling. Since then, even cranking up input gain all the way reveals nothing but the expected hiss.
If I felt like soldering my own cables, I'd probably go with the cabling suggested by Bill Whitlock (of Jensen Transformers) - he uses a balanced cable and connects it to an unbalanced output as follows: "hot" to unbal signal, "cold" to unbal ground, shield to unbal ground as well. This way you get a halfway decently balanced connection, with impedance imbalance amounting to unbalanced output impedance - which still tends to be good for a decent amount of CMRR (at least 30-ish dB, and >=50 dB with low-impedance outputs), vs. none at all on the purely unbalanced connection. I have to agree that this is how "unbalanced to balanced" aftermarket cables should be built.
(The advanced DIYer may still choose to integrate a series resistor in the "cold" connection for optimum balance if source output impedance is known.)

Active monitor speakers often give trouble with unbalanced connections - like most pro audio gear, they are usually built as Class I devices with audio ground connected to safety earth. Connect a 1/8" stereo to 2x TS unbalanced cable to two of them, and you have created a ground loop without even having the cable plugged into the PC! Once you do that, there's two more...  Keeping out of trouble with unbalanced connections requires a fair bit of attention from equipment makers and users alike.

(It's not like balanced interconnection would be entirely free from trouble, but that generally has more to do with implementation issues like "pin 1 problem".)

Sound Cards

Reply #12
I plugged my 300ohm HD800s to my crappy 2008 netbook. It was actually plenty loud to my surprise. A friend of mine that I think is a bit loony is telling me that he prefers his $30 Razer Morays over every single headphone we tried out on a Head-fi meet, saying that his earphones are less accurate but more like a real live performance.  Anyways, he tried my HD800s with my O2/dac, and he says he can hear hissing when the audio knob is turned up. I think that's kindda insane, I'm not sure if I can hear it as long as I don't hit the gain switch, and the sound of the fans running on my PC is way louder than whatever hiss he is hearing...

So what exactly is different, sound wise, from a ground loop issue versus a bad dac that is affected by say, electromagnetic interference? The ground loop I got before I used the ground loop isolator on my monitors were pretty hardcore, and they were not just a hissing sound. I think before I used the isolator, the ground loop sound occured on top of a hissing sound. And now the louder, more obnoxious ground loop is out of the way, the hissing on the monitors is more obvious (but still not bad).

If I stick my HD800 into O2/dac and I hear a hiss when the volume is cranked up, that's not a ground loop related issue, right? That's just the dac's noise floor?

Sound Cards

Reply #13
If I stick my HD800 into O2/dac and I hear a hiss when the volume is cranked up, that's not a ground loop related issue, right?


True. Ground loop issues typically manifest themselves as hum.

Quote
That's just the dac's noise floor?


True, and the if your normal listening situation cranks the volume back from full to some degree, then it is an irrelevant test.

Sound Cards

Reply #14
Thanks for weighing in, Arnold.

So if the entire point of balanced audio is to prevent ground loop issues, then it seems pretty pointless for me right now. And also, we're all agreeing that although in terms of which gear has the larger number next to the SNR rating, the STX wins, the O2 actually wins in terms of noise floor?

So if ground loop doesn't manifest as a hiss, then what is actually causing the hiss on my Rokits? It's not the noise floor. Even unplugging the connection from monitor to amp/dac, there is no change in hissing. Changing amp/dacs did not affect the hiss. I'm assuming it's a fault of the speakers themselves. The hiss doesn't really bother me at this point. The ground loop issues drove me insane, but that's out of the way. But I'm curious as to why the hiss is audible when it's not the dac's noise floor.

AFAIK, the O2/dac is the best, measurement wise, of any amp or dac in their price range, or really, even double that, depends. So I was telling my friend, if he's not happy with the noise floor of my Odac, he is in for disappointment when he gets his new onboard audio. Of course we have to consider a few things with his disappointment: 1) Max volume on O2 is not equally as loud as max volume on his soon to be new motherboard. The O2 on max is WAY LOUDER. It's not fair to compare noise at max vs max volume, we need to compare at about equal volume settings and try to see which has more noise. 2) It's harder to fault the O2 because by the time you raise the volume to where hiss is an issue (at least to his oddball ears... I almost want to subject him to a blind test for the hiss), your ears will be damaged long term by the volume. 3) Volumes that are excessively high makes it harder to hear details in the music. 4) It's a little odd to be complaining about noise floor when the fridge next to me and the fans in my computer case are many times louder. Of course, say he gets an O2 and then slaps on a headphone that requires him to gun up the volume... or he uses sensitive IEMs... maybe the hissing would become a real problem... I'm not sure at this point. I have a little bit more faith in NWavguy than that though.

Ok, so some more questions:

1) I believe that, as long as solid state amps are performing correctly, they should sound identical. True/false?
2) I also think that, the goal of a tube amp is to color the sound in a way that seems pleasing. EQ can alter the sound, but that has some pretty harsh limitations IMO. NWavguy made the argument that because this means buying an amp to match a headphone, every time you use a different headphone, you've gotta switch the amp too. But if you're just using one headphone forever and ever and you want to color the sound without EQ, would tubes make sense? If tubes exist solely for being a novelty and coloring the sound, why are there tubes that try to be neutral?
3) I never got a chance to test cables, let alone do any sort of abx testing. I believe most of the people on this forum take the position that cables don't affect the sound, correct?
4) Honestly, I just never got this one: So as far as amps go, we have two things: Current and voltage. An insensitive headphone, like the HE-6, requires very high amounts of current, making it very hard to drive. But apart from that oddball headphone, often times people talk about "driving headphones" and amps as if, the higher the impedance of a headphone, the harder it is to "drive". Yet, when I was pointing out the impedance graph of HD800s, about how it peaks way above 300ohms in the bass region, a guy mentioned that higher impedance means lower current, hence they are EASIER to drive. I don't understand this at all. First we're acting like higher impedance is hard to drive, now all of a sudden we almost want higher impedance because it's lower current, and therefore easier to drive. As I understand it, needing too much of either the current or voltage causes problems driving a headphone. So wouldn't a impedance peak push a headphone towards an extreme, making it harder to drive?

Sound Cards

Reply #15
1) I believe that, as long as solid state amps are performing correctly, they should sound identical. True/false?

I'd doubt that. Different designers/manufacturers have different priorities and approaches to what they consider 'realistic' sound. And in my experience, they do sound different. Although I have only owned 5 different amplifiers in my life, unlike some people, going from a NAD to a Pioneer was a change from a softer to a more detailed sound. I liked both of them, but for different reasons.

Sound Cards

Reply #16
2) I also think that, the goal of a tube amp is to color the sound in a way that seems pleasing. EQ can alter the sound, but that has some pretty harsh limitations IMO. NWavguy made the argument that because this means buying an amp to match a headphone, every time you use a different headphone, you've gotta switch the amp too. But if you're just using one headphone forever and ever and you want to color the sound without EQ, would tubes make sense? If tubes exist solely for being a novelty and coloring the sound, why are there tubes that try to be neutral?
3) I never got a chance to test cables, let alone do any sort of abx testing. I believe most of the people on this forum take the position that cables don't affect the sound, correct?

As for questions 2 and 3, I did once hear a magnificent tube amp setup, which didn't seem to have the 'romantic haze' of colouration about it. I forget the source, but I think the amps may have been Art Audio, playing through Apogee Centaur (or Centaur Minor) speakers. Unfortunately, I didn't get to listen to that system for very long, but I could have stayed a lot longer. Ruinously expensive, though, and that was at least 20 years ago.

Cables and interconnects: for me, people should definitely step up from the 'giveaway' stuff, like ditching the cruddy earbuds supplied with most portable player. But the law of diminishing returns kicks in quite early at least in my opinion. I don't know what's on the market in the USA, but in the UK, Richer Sounds have some keenly-priced cables and interconnects that are worth a modest investment.

Sound Cards

Reply #17
Cables and interconnects: for me, people should definitely step up from the 'giveaway' stuff, like ditching the cruddy earbuds supplied with most portable player. But the law of diminishing returns kicks in quite early at least in my opinion. I don't know what's on the market in the USA, but in the UK, Richer Sounds have some keenly-priced cables and interconnects that are worth a modest investment.


What do they offer over something like the standard cables included with their CD player or DAC or whatnot?

Earbuds (and headphones and speakers in general) are mechanical devices with various physical tradeoffs, so it makes sense that they sound different.

But what exactly does "upgrading" your cables and interconnects offer on the sound quality front?

Sound Cards

Reply #18
In general, nothing. Unless you have a faulty cable or one which is mismatched for the application, even the cheapest cabling should be audibly indistinguishable from any other proper cable.

Sound Cards

Reply #19
@JabbaThePrawn:

Sighted tests are not objective. Your conclusions are not permissible here, according to our Terms of Service.

Sound Cards

Reply #20
@JabbaThePrawn:

Sighted tests are not objective. Your conclusions are not permissible here, according to our Terms of Service.

Did he mention that his tests are sighted? I don't recall him saying that.


My stance on cables and solid state amps are very similar: Ideally they are audibly transparent. If not, it is due to an imperfect cable and/or amp/dac. (Assuming accuracy is the goal - If you want to color the sound on purpose then this doesn't apply and obviously if you purposefully alter the sound something makes, it will sound different.)

Which still doesn't answer the question of why people want neutral sounding tube amps though.
Personally I have not done any test on cables, sighted or blind or otherwise. There's no chance for me to, and I didn't have time at the last Headfi meet to even think about cables. BTW, anybody know the answer to my 4th question?

Thanks

Sound Cards

Reply #21
BTW, anybody know the answer to my 4th question?

You pretty much have it right. As I described earlier, an amplifier runs optimally at a particular load impedance, i.e. it can supply the most watts into this impedance. If the load impedance is either higher or lower than this then the maximum watts are somewhat less. As a result, if the load impedance varies with frequency then it is possible to have too few amps at one frequency and too few volts at another frequency.

What this does not take into account is that the maximum amps or volts out of an amplifier can also vary with frequency, and in particular it is often the case that the maximum current is less at low frequencies (due to use of a capacitor to block DC). This results in limited bass into low impedance headphones.

Sound Cards

Reply #22
1) I believe that, as long as solid state amps are performing correctly, they should sound identical. True/false?


That is true of any good amp, tube or SS.


Quote
2) I also think that, the goal of a tube amp is to color the sound in a way that seems pleasing. EQ can alter the sound, but that has some pretty harsh limitations IMO. NWavguy made the argument that because this means buying an amp to match a headphone, every time you use a different headphone, you've gotta switch the amp too. But if you're just using one headphone forever and ever and you want to color the sound without EQ, would tubes make sense? If tubes exist solely for being a novelty and coloring the sound, why are there tubes that try to be neutral?


False.

All good amps try to be neutral and succeed. Active device type matters not.

It is true that we have a certain number of boutique products that are engineered to sound bad. It is harder and more expensive to make a good amp with tubes than SS.

Quote
3) I never got a chance to test cables, let alone do any sort of abx testing. I believe most of the people on this forum take the position that cables don't affect the sound, correct?


There is a generality that seems to need to be stated:

All good audio should be audibly transparent, no matter what it is.  Even if it is a device that is designed to make audible changes such as an equalizer, crossover or dynamics processor, it should have some benign mode where it does not make audible changes to the signals that pass through it.

Quote
4) Honestly, I just never got this one: So as far as amps go, we have two things: Current and voltage. An insensitive headphone, like the HE-6, requires very high amounts of current, making it very hard to drive.


Fractionally true.

The combination of low impedance and/or low efficiency and/or listener who prefers loud sound and/or ear damaged listener puts requirements for above average current on the amp that drives it.

The combination of high impedance and/or low efficiency and/'or listener who prefers loud sound and/or ear damaged listener puts requirements for above average voltage on the amp that drives it.

However, I am unaware of any combination of the above related to headphones that should tax the brain or budget of a good equipment designer.

Quote
But apart from that oddball headphone, often times people talk about "driving headphones" and amps as if, the higher the impedance of a headphone, the harder it is to "drive". Yet, when I was pointing out the impedance graph of HD800s, about how it peaks way above 300ohms in the bass region, a guy mentioned that higher impedance means lower current, hence they are EASIER to drive. I don't understand this at all. First we're acting like higher impedance is hard to drive, now all of a sudden we almost want higher impedance because it's lower current, and therefore easier to drive. As I understand it, needing too much of either the current or voltage causes problems driving a headphone. So wouldn't a impedance peak push a headphone towards an extreme, making it harder to drive?


There seems to be this recent obsession with headphone amps.  They only deal with microscopic amounts of power compared to power amps, and even power amps up to the tens of thousands of watts are solved problems, and quite economically thank you. I don't get it on any rational technical level, but I do get it on a human behavioral level.

Sound Cards

Reply #23
@JabbaThePrawn:

Sighted tests are not objective. Your conclusions are not permissible here, according to our Terms of Service.

Did he mention that his tests are sighted? I don't recall him saying that.



Doing a listening test with enough experimental controls to make it reliable is enough work that any money bet on the fact that a test was done that way would be stated as such up front, is pretty safe.

It would be like a BMW driver talking about his car and never mentioning that it was a BWW! ;-)


There is even a phrase: "Blind test" which sounds impressive but usually ends up being just a Single Blind test, which is generally a defective double blind test. IOW, its not reliable enough to be given any more credibility than a sighted evaluation.

Sound Cards

Reply #24
1) I believe that, as long as solid state amps are performing correctly, they should sound identical. True/false?


That is true of any good amp, tube or SS.


Quote
2) I also think that, the goal of a tube amp is to color the sound in a way that seems pleasing. EQ can alter the sound, but that has some pretty harsh limitations IMO. NWavguy made the argument that because this means buying an amp to match a headphone, every time you use a different headphone, you've gotta switch the amp too. But if you're just using one headphone forever and ever and you want to color the sound without EQ, would tubes make sense? If tubes exist solely for being a novelty and coloring the sound, why are there tubes that try to be neutral?


False.

All good amps try to be neutral and succeed. Active device type matters not.

It is true that we have a certain number of boutique products that are engineered to sound bad. It is harder and more expensive to make a good amp with tubes than SS.

Quote
3) I never got a chance to test cables, let alone do any sort of abx testing. I believe most of the people on this forum take the position that cables don't affect the sound, correct?


There is a generality that seems to need to be stated:

All good audio should be audibly transparent, no matter what it is.  Even if it is a device that is designed to make audible changes such as an equalizer, crossover or dynamics processor, it should have some benign mode where it does not make audible changes to the signals that pass through it.

Quote
4) Honestly, I just never got this one: So as far as amps go, we have two things: Current and voltage. An insensitive headphone, like the HE-6, requires very high amounts of current, making it very hard to drive.


Fractionally true.

The combination of low impedance and/or low efficiency and/or listener who prefers loud sound and/or ear damaged listener puts requirements for above average current on the amp that drives it.

The combination of high impedance and/or low efficiency and/'or listener who prefers loud sound and/or ear damaged listener puts requirements for above average voltage on the amp that drives it.

However, I am unaware of any combination of the above related to headphones that should tax the brain or budget of a good equipment designer.

Quote
But apart from that oddball headphone, often times people talk about "driving headphones" and amps as if, the higher the impedance of a headphone, the harder it is to "drive". Yet, when I was pointing out the impedance graph of HD800s, about how it peaks way above 300ohms in the bass region, a guy mentioned that higher impedance means lower current, hence they are EASIER to drive. I don't understand this at all. First we're acting like higher impedance is hard to drive, now all of a sudden we almost want higher impedance because it's lower current, and therefore easier to drive. As I understand it, needing too much of either the current or voltage causes problems driving a headphone. So wouldn't a impedance peak push a headphone towards an extreme, making it harder to drive?


There seems to be this recent obsession with headphone amps.  They only deal with microscopic amounts of power compared to power amps, and even power amps up to the tens of thousands of watts are solved problems, and quite economically thank you. I don't get it on any rational technical level, but I do get it on a human behavioral level.

About the tube amps and coloration:
Why do neutral tubes exist though? Forgetting the people who want coloration, or who get tubes for non-audio reasons. Why make tubes when you can get solid state? I mean, one could argue that, what sounds better or worse is up to personal opinion. The only thing you can really state is whether something is more or less accurate, not whether a person has to find it better or worse. Then that leaves room for a person justifying a tube amp that colors the sound. But that still doesn't explain the point of a neutral tube amp.

The Objective is the go-to performance DAC/amp... yet it is said by multiple people that it is not capable of powering the HE-6. Granted, HE-6 is a fringe example, but it just so happens that I tried an HE-6 the other day and quite enjoyed it. The modified ODac/amp at DIYaudio is said to be capable of driving HE-6 among other things, but there is nobody assembling them and selling them. Obtaining one is strictly DIY, something I am not capable of.

Regarding the cables what I was asking was, whether a normal cable is audibly transparent. I'm assuming the answer that would be given here would be yes.

 
SimplePortal 1.0.0 RC1 © 2008-2020