Thoughts on a bunch of DACs (and why delta-sigma kinda sucks, just to get you to think about stuff)

[KeithEmo]

You're asking a bit much, not a pro-studio engineer. Saw some of those though (nothing of say abbey road size & prestige) and my general impression was: very detailed and clear sound but not a relaxing listening space. Some studio speaker samples: k+h o300-400, focal (biggest 3way ones), barefoot, pmc (dont remember models, pretty big & shinny new ones). I did not measure anything but the sound was quite linear, no obviously accentuated freqs and all had some sort of room treats.
Their components did sound quite linear/neutral but also kinda shrill for my ears. Same for many ultra highend (and suposedly very neutral) hifi setups in various shops.

As about the slewrate, I think most modern amps (at least the ss kind) are linear to 50khz or more .. even "lousy" receivers are. Doubt that is a big issue nowadays. A poor slew rate would prolly also defeat the linear to 20khz rule .. same about the peaks you mentioned.

So, I think I still have a Q. Other than odd order harmonics why would a linear-to-20k component sound shrill?

 
Slew rate is really "how fast the signal passes through the amplifier". In an amplifier with absolutely no feedback, slew rate is unimportant.... because you certainly can't hear whether what you're listening to takes an extra few millionths of a second to play. (This is ignoring the fact that a really low slew rate probably implies a poor frequency response as well.) However, negative feedback relies on the output signal, which is being fed back to the input, being EXACTLY 180 degrees out of phase with the input signal - and anything less than an infinite slew rate "throws this off" to a degree. Therefore, as you use higher levels of feedback, it becomes more critical that you have a sufficiently high slew rate. 
 
To take this to two extremes:
1) If you used zero feedback, and assuming everything else was "theoretically perfect", slew rate wouldn't matter at all.
2) If your slew rate was theoretically perfect (infinite), and everything else was also perfect, there is no reason for ANY amount of feedback to cause audible distortion.
 
(Of course, since nothing is perfect, neither of these two situations ever happens 100%.)
 
There are two reasons why most modern audio equipment doesn't quote slew rate any more. First, this relationship is pretty well known - and it's not that difficult to design an amplifier with a slew rate that is more than sufficient for audio. Therefore it's not really a problem with modern designs. Second, while the particular slew rate that you need to avoid causing problems depends on the design itself, an insufficient slew rate will show up in the normal specifications. (An amplifier with an insufficient slew rate will exhibit THD that rises rapidly at high frequencies. So, if you see an amplifier that has 0.05% THD from 20 Hz to 2 kHz, but has 10% THD at 20 kHz, then there's a very good chance it has insufficient slew rate for the amount of feedback and gain it uses. SInce slew rate varies with both frequency and level, this rise will occur sooner and more prominently as the output level is raised. Conversely, if a design maintains a reasonably low THD over the entire audio band, and at reasonable power levels, then it has sufficient slew rate to go with its other design parameters.)
 
"Back in the old days" the overall distortion specs where high enough that this anomaly would often be "buried" under the normal noise and distortion floor. However, with modern designs, having the THD vs Frequency curve go from "very low" to "almost straight up" at some high but audible frequency would be pretty obvious in the standard graphs... so no separate test is needed.

 

[prot]

KeithEmo
"It's mostly in the source" is of course another 100% valid possibility. I'm not in audio (IT guy) but I hope noone ever records drums & cymbals from only cm/inches away ... there's a reason why drum players wear earplugs.

And yep, live acoustic music and recordings don't seem to sound the same, even in my very limited experience. But lately I checked some pretty good binaural recordings and I am quite impressed ... at least they seem to capture the room and that elusive feeling of realism much better. E.g. binaural recordings of rain sound quite wet already . Don't know if it's the methodology or just the newer/improved recording equipment, but I surely like what I'm hearing.

As for the initial "why would neutral sound harsh" question, that is surely loaded and "harsh" is of course very personal. But it's not just my observation, many people seem to agree that the very neutral components that the pros usually use are kinda harsh .. the pros included. And even the big OP review here puts studio DACs like Benchmark/Lavry/Mytek in the (somewhat) shrill category ... same for other components which are known to measure very neutral like Odac/O2

 

[KeithEmo]

@KeithEmo
"It's mostly in the source" is of course another 100% valid possibility. I'm not in audio (IT guy) but I hope noone ever records drums & cymbals from only cm/inches away ... there's a reason why drum players wear earplugs.

And yep, live acoustic music and recordings don't seem to sound the same, even in my very limited experience. But lately I checked some pretty good binaural recordings and I am quite impressed ... at least they seem to capture the room and that elusive feeling of realism much better. E.g. binaural recordings of rain sound quite wet already

. Don't know if it's the methodology or just the newer/improved recording equipment, but I surely like what I'm hearing.

As for the initial "why would neutral sound harsh" question, that is surely loaded and "harsh" is of course very personal. But it's not just my observation, many people seem to agree that the very neutral components that the pros usually use are kinda harsh .. the pros included. And even the big OP review here puts studio DACs like Benchmark/Lavry/Mytek in the (somewhat) shrill category ... same for other components which are known to measure very neutral like Odac/O2

 
I think the way that final sentence is phrased sort of tells the story.
 
If someone thinks that some components that measure neutral still sound shrill, while others do not, then there could be something else going on.
 
However, if someone thinks that "all components that measure neutral sound shrill" and concedes that "the components that don't sound shrill don't measure neutral", then they simply don't like components that are actually neutral - or, to say that differently, they find that components that are actually neutral sound shrill to them.
 
Of course, since this is a headphone forum, and we all know that different headphones sound very different, you really need to qualify that further as to whether they only find that those components sound shrill with certain headphones. Personally, I know that a lot of headphones aren't especially flat, so I don't find it at all surprising that certain headphones may sound shrill when presented with a perfectly neutral signal, and may sound less so when presented with a signal which is NOT neutral, but has some roll off that de-emphasizes the frequencies that they over-emphasize.
 
I know that I personally find many headphones to sound very harsh when presented with a neutral input signal, and so to sound smoother when presented with a signal that is rolled off, and so simply avoids the frequencies that sound harsh on those headphones. (This is why, even though I much prefer solid-state headphone amps in general, I find that certain headphones sound better to me when run with tubes.)
 
(However, to me, the flaw is with the headphones, and I find having to select a flawed DAC or amp to make my flawed headphones sound good to me to be an unacceptable compromise. My personal goal is to have each piece of equipment I own be as good as it can be INDIVIDUALLY, so I don't have to worry about which ones I use together, and so I tend to avoid "picky" headphones - or amps, but some people seem to not mind having to figure out specific combinations that "go well together", or even to "enjoy the chase to perfection".)
 
I would also note that, at least to me, I find that all DACs that use the Sabre DAC chips have a somewhat distinctive sound. To me they sound as if the upper frequencies are slightly emphasized - even though they have a very flat frequency response. I use the analogy of looking at your carpet with a very bright LED flashlight; you don't see anything that wasn't there under incandescent light, but the dust and dirt seems more obvious, as if the LED light picks them out more carefully. You tend to notice more of the details that were there all along. To me, this goes beyond neutral, and is an actual emphasis - and I'm pretty sure it's related to the digital oversampling filter they use. If you read reviews, you will find that many people who like Sabre DACs describe them as "detailed" or "revealing", while many who don't describe them as "etched". 
 
In the context of this post, I can definitely see how this would quickly become very annoying with headphones that already emphasize the upper treble - and could well tip the balance between "clean" and "harsh". (And I could see how a headphone or amp that rolled off the high end would make it much less noticeable. I know that I barely noticed the difference with a pair of HiFiMan planars I used to own.)

 

[jcx]

slew rate is a little different than described above - it is the rate of change of the signal at a given instant in time
 
we are usually interested in the maximum slew rate of a signal and the limits of the amplifying electronics
 
for those with basic Calculus its easy to see that a 20 kHz, 2 V sine wave has ~ 250000 V/s maximum slew rate at its zero crossings - usually we use Volts per micro-seconds in audio electronics so that would be 0.25 V/us
 
the max slew rate possible is a function of frequency and the max V so it really needs to normalized to the max V at a particular point in your system - the 1 Vrms of most DAP, 2 Vrms of  consumer DAC or the max level after gain of a headphone amplifier
 
since Vpeak is sqrt(2) higher than Vrms - consumer desktop DAC output a "conventional audio" 20-20kHz signal is ~0.36 V/us and with special digital signal tricks you could ask for ~2x that to be "safe"
 
with higher sample rates than CD 44.1k you certainly could record proportionately higher frequency "ultrasonic" signals with higher slew rate - but many surveys of music show generally that music has a "power bandwidth" of <3-5 kHz == amplitude falls above that frequency fast enough that musical signal slew rate is seldom larger than needed for the 3-5 kHz at max amplitude - that's ~ 4x less than the 20 kHz max amplitude calculated number
 
 
the link between feedback errors and slewrate was the source for some controversy in JAES in the '70s when Matti Otala started publishing his theories on Transient Intermodulation Distortion - there are links depending on the amp electronics internal design, Otala incorrectly assumed a particular linkage was unavoidable and dogfights ensued
his fundamental assumption has been shown wrong in theory, by hardware realizations, custom built measurement hardware using his TIM definitions - but it is a powerful meme that many wanting to criticize high feedback amplifiers have refused to give up
 
 
another thing that has happened without getting credit from Otala/TIM adherents is that over time op amp manufacturers have improved the semiconductor processes and internal circuit topologies - modern chips intelligently selected for the application have vanishingly small slew rate caused errors for audio
and that's off a "baseline" of NE5534, TL072 that managed ~ 10 V/us in the '80s
 
just take a "prosumer" $150 sound card like the ESI Juli@ and try to show TIM in analog loopback - despite the only slightly better than generic op amps - certainly not using anything a "op amp roller"would put in, "audiophile" manufacturer would brag on (if they admitted to using op amps at all) 

 

[evillamer]

Can anyone explain why hi hats sound different on sigma delta vs r2r (ch ch ch vs tsh tsh tsh)

Is it because of accumulator overload or non-filterable aliasing or lack bits in modulators(quantization errors) or feedback issues?

 

[prot]

I would also note that, at least to me, I find that all DACs that use the Sabre DAC chips have a somewhat distinctive sound. To me they sound as if the upper frequencies are slightly emphasized - even though they have a very flat frequency response.
... and I'm pretty sure it's related to the digital oversampling filter they use.

that ... why would that happen?
You say Sabre & oversampling filters ... but I do not think it's only coming from the DACs, my very linear speakers also had their harsh moments ... wish I could remember all the DACs I tried in that system but I'm pretty sure there was some non-DS stuff (like old CD players with philips chips) ... but the speakers were always kinda sparkly

Oh well, may be better to put this whole "linear is harsh" thing to rest ... according to all else I read there are prolly no definitive and/or generic answers ... just had a hope that maybe I missed something and some of the experts could bring the light in a few clear words.

 

[jcx]

show the signals, and the listening test conditions, statistics and more people who know a little about DAC architectures, specs, errors would have a better chance of usefully speculating
 
Audio DiffMaker files of the DAC's outputs with the signals that you can show that you differentiate in controlled listening would be a place to look with various signal analysis tools

 

[prot]

show the signals, and the listening test conditions, statistics and more people who know a little about DAC architectures, specs, errors would have a better chance of usefully speculating

Audio DiffMaker files of the DAC's outputs with the signals that you can show that you differentiate in controlled listening would be a place to look with various signal analysis tools

Life would be so easy if I could just do that .

Like said, it's all based on many, many auditions/tests plus similar impressions I did read in various places. And may very well be that it's all 100% subjective.

 

[KeithEmo]

  slew rate is a little different than described above - it is the rate of change of the signal at a given instant in time
 
we are usually interested in the maximum slew rate of a signal and the limits of the amplifying electronics
 
for those with basic Calculus its easy to see that a 20 kHz, 2 V sine wave has ~ 250000 V/s maximum slew rate at its zero crossings - usually we use Volts per micro-seconds in audio electronics so that would be 0.25 V/us
 
the max slew rate possible is a function of frequency and the max V so it really needs to normalized to the max V at a particular point in your system - the 1 Vrms of most DAP, 2 Vrms of  consumer DAC or the max level after gain of a headphone amplifier
 
since Vpeak is sqrt(2) higher than Vrms - consumer desktop DAC output a "conventional audio" 20-20kHz signal is ~0.36 V/us and with special digital signal tricks you could ask for ~2x that to be "safe"
 
with higher sample rates than CD 44.1k you certainly could record proportionately higher frequency "ultrasonic" signals with higher slew rate - but many surveys of music show generally that music has a "power bandwidth" of <3-5 kHz == amplitude falls above that frequency fast enough that musical signal slew rate is seldom larger than needed for the 3-5 kHz at max amplitude - that's ~ 4x less than the 20 kHz max amplitude calculated number
 
 
the link between feedback errors and slewrate was the source for some controversy in JAES in the '70s when Matti Otala started publishing his theories on Transient Intermodulation Distortion - there are links depending on the amp electronics internal design, Otala incorrectly assumed a particular linkage was unavoidable and dogfights ensued
his fundamental assumption has been shown wrong in theory, by hardware realizations, custom built measurement hardware using his TIM definitions - but it is a powerful meme that many wanting to criticize high feedback amplifiers have refused to give up
 
 
another thing that has happened without getting credit from Otala/TIM adherents is that over time op amp manufacturers have improved the semiconductor processes and internal circuit topologies - modern chips intelligently selected for the application have vanishingly small slew rate caused errors for audio
and that's off a "baseline" of NE5534, TL072 that managed ~ 10 V/us in the '80s
 
just take a "prosumer" $150 sound card like the ESI Juli@ and try to show TIM in analog loopback - despite the only slightly better than generic op amps - certainly not using anything a "op amp roller"would put in, "audiophile" manufacturer would brag on (if they admitted to using op amps at all) 

 
You're quite right - slew rate itself is "how fast the voltage can change" - which is a very specific measurement, and I conflated that with various other causes of phase shift and delay.

 

[KeithEmo]

that ... why would that happen?
You say Sabre & oversampling filters ... but I do not think it's only coming from the DACs, my very linear speakers also had their harsh moments ... wish I could remember all the DACs I tried in that system but I'm pretty sure there was some non-DS stuff (like old CD players with philips chips) ... but the speakers were always kinda sparkly

Oh well, may be better to put this whole "linear is harsh" thing to rest ... according to all else I read there are prolly no definitive and/or generic answers ... just had a hope that maybe I missed something and some of the experts could bring the light in a few clear words.

 
I suspect there are many answers.....
 
To me, the biggest question there is whether a given component is adding harshness, or simply showing you exactly what's there in the source material, and so allowing you to hear harshness that's already there. If the component is accurate, and your source material really sounds harsh, then it is the job of a high fidelity component to let you hear exactly how bad your source material is - and, if that bother you, then it is your job to pick other source material, or to find better recordings. Likewise, if your headphones really sound harsh, then it isn't the job of your amp or DAC to alter the sound so you can't hear the problem. Much as I prefer a smooth sound, I simply can't "abide by" any component that makes things sound good by hiding how they really sound. (Even though it may make certain poor quality sources sound better, or avoid frequencies that sound harsh on a particular pair of headphones, I'm always inclined to believe that it is also preventing me from hearing all of how good really high quality sources and other components can sound. Obviously not everyone agrees with my philosophy there.)
 
I was not at all suggesting that harshness is always caused by the DAC, or even that Sabre DACs tend to sound harsh (I would not actually describe the way they sound that way). What I was suggesting was that most headphones don't have a very flat frequency response, and that combining a Sabre DAC, which tends to emphasize certain frequency ranges and certain aspects of the sound presentation, with a set of headphones that also either emphasizes those same frequencies, or simply doesn't handle them very well, could lead to an overall result that would be characterized as harsh. (To phrase that differently.... I would personally specifically avoid pairing a Sabre DAC with headphones that I already considered to be at all harsh or bright. I would much prefer to pair a Sabre DAC with somewhat laid back headphones whose sound character complements it - like planars.)
 
 
 
 
 

 

[prot]

KeithEmo
that's why I wated to put the Q to rest .. too many answers.
And your're right .. even if the DS DACs were indeed too bright/harsh/shrill/etc, there are still thousands of dark amps/HPs/speakers that would be a perfect match. IMO, there is quite a lot of ado about nothing in this thread.

 

[evillamer]

On the topic of Audio Diff Maker, Furman/Panamax did a test with it that shows that their equipment does provide a difference to the sound.
 

 

[evillamer]

@KeithEmo
that's why I wated to put the Q to rest .. too many answers.
And your're right .. even if the DS DACs were indeed too bright/harsh/shrill/etc, there are still thousands of dark amps/HPs/speakers that would be a perfect match.

Well I feel it's like wearing dark sunglasses to cover up the tiny dust and dirt of a room.

 

[prot]

Well I feel it's like wearing dark sunglasses to cover up the tiny dust and dirt of a room.

Could be. But same as true you could say it's like combining sweet & sour for an even better taste.
IMHO, audio is subjective enough, no need to muddy the waters further with (dubious) analogies

 

[evillamer]

 

Well I feel it's like wearing dark sunglasses to cover up the tiny dust and dirt of a room.

Could be. But same as true you could say it's like combining sweet & sour for an even better taste.
IMHO, audio is subjective enough, no need to muddy the waters further with (dubious) analogies

 
This thread is full of analogies, Purrin uses cars, Keith uses LED lights and Carpet. What's dubious?