Not anytime soon, we are still in lock-down and vaccination is very slow. I'm hearing he will use a Kawai, though, you may be disappointed. OTOH, I've spent enough time in Italy, at the University of Trento and Rome, during my post doc studies, so I have little interest in visiting again, even less for listening to clocks.
Last edited:
Not even a little scope? Basically, use a square wave on the input, and check the output and the supplies for ringing.
These ultra-everything opamps are often quite fast (like 50MHz GBW) so they require attention to decoupling and layout. Fast opamps are quite sensitive to inductance in their power supplies. So, for example, it is quite common to see people here replace an opamp like a NE5532, which is slower, with a fancy LME49xxx. But... if the board was designed for a NE5532 or a 4580 or something like that, there will be some electrolytic decoupling caps somewhere on the board, with quite long power traces to the opamps. This is fine with slower opamps, but not for 50MHz opamps. Especially at the output of a DAC, where the signal contains high frequency, the fast opamp will process it, output it, and to do so it will draw some high frequency current from its power supplies. If there are no decoupling MLCCs and ground plane to get a low inductance power supply, this will result in excess HF noise on the rails and in the output, ringing, and worse settling time. Also the internal compensation of these amps is referred to the power supply, so if it has extra inductance, stability can be compromised.
there's a nice series of articles about this by Kendall Castor Perry:
https://www.edn.com/yet-more-on-dec...ing-the-complete-op-amp-power-supply-circuit/
Basically, if the opamp is relatively fast (like 20MHz or more), then I'm not interested in opinions about how it sounds in a circuit if I can't see the layout, because if the layout and decoupling are not proper, then the result is not the "sound" of the opamp (if it has one...) but rather the result of the interaction of a PCB and its parasitics with the opamp. And that provides no useful data, since every bad PCB is different.
There's one thing I've noticed though: when the layout is good and the power supplies are clean, then it works. Not really surprising
Last edited:
Kawai or Steinway don't matter, they measure the same. Evgeny Kissin makes an audible difference.
Anyway you can still found good wine to get the clock sounds better.
Great news, first vaccine tomorrow!
I'm almost ready for Canada!
P.S.
I would suggest Scuola Superiore Sant'Anna in Pisa for post-doc.
Last edited:
Correct. There is some amount of 3D information and some amount of spatial information even in a mono signal.
Here is an easy example, I place a Sony in-ear-monitor in my ears playing a 100% mono song. Ok, the sound sounds around 2 centimetres away from my ears, very nearby.
Now, I play the exact same mono song on a sub-woofer 10 metres away, with a tweeter 1.8 metres away. Now the mono song sounds spatially divided with the bass 10 metres away & the treble 1.8 metres away.
Now if someone is going to argue it's about reflection on the walls, ok, we put the sub-woofer & tweeter in an anechoic room, they still sound 10 metres & 1.8 metres away, since we have two ears & if the sub-woofer signal arrives 900 nanoseconds later to the left ear, than the right ear, then the formula goes into a cluster of neurons which are programmed to know distance.
Plus, a Titanium tweeter sounds different than a plastic tweeter, it could be a little in FR, a little in time domain, it's not important, it's just the way it is.
----
It's not interesting for me to discuss with audio people that there is spatial information even in mono signals, there should be some acousticians somewhere which have the measurements on what you are referring to, in mono signals. Like, when Sony or Canalworks are developing a new product, actually I think they don't care about measuring the spatial effects in a single channel in a mono signal, they just create the product & the spatial sector is rolling a dice. For example, here is the product sheet of a Sony IEM, it basically covers some of their 'mission agenda' so to speak ---> Sony MDR-EX1000 In-Ear Monitor Headphones MDREX1000 B&H Photo
So, 1 - "liquid crystal polymer film diaphragm for rigidity & responsiveness", I saw the pictures & details of their measurements & they experimented with different plastic diaphragm materials until they found a material with smooth FR lines in the USF, plus other companies similar to Sony measured diaphragms for the IR to zero time (kind of like settling time), so for example Titanium is much faster than various types of plastic, however Sony may think the liquid crystal polymer film they used is more transparent, even if it's less fast in the IR to zero. In the 1990's, Sony engineers were looking for a diaphragm material which could have the speed of Titanium & yet still have a natural sound at the same time, so they spent a lot of time with different plastics & different materials, eventually in their headphone drivers they decided on nanocomposite diaphragms made of Silica, so the labs had these jars of Silica which were transferred in nano dust / micro dust onto the drivers. Then in earphones in the 1990's / early 2000's they eventually decided on different types of plastics, such as Polyimide (in 200 or so micro layers), so they kept the plastic culture until now however changed to liquid crystal polymer in the EX1000.
The above paragraph isn't fancy talk since the THD measurements of the EX1000 are much lower than everything in any K-Mart & any smartphone earphones & any of the popular brands such as Apple & Skullcandy. EX1000 is much lower in THD than pretty much every single headphone in existence, as well, so it's not important if it's a $300 Apple or a $2000 Ultrasone or a $7000 iBasso or a $10,000 American electrostatic nonsense, I have the lab data & lab pictures to back up my hearing.
2 - "Powerful neodymium magnets".
3 - "Magnesium housings", just like the brochure says they used Magnesium housings to greatly reduce unwanted vibration, dissonance & echo. It's just like when you are in a small karaoke room, the echo is constant & overwhelming. Another example you can press buttons on the karaoke panel which introduce artificial time domain changes.
Everything in this paragraph above has nothing to do with FR!
-------
I don't need to spend time reading audio blogs where people say " FR is everything " &/or " THD under 0.5% is inaudible ", I think they are not audio enthusiasts & they are schooled in a way where they need evidence about everything or they don't care, so they're in this kind of myopia place & then trying to pull everyone else down into the myopia place as well; it's kind of like they want to 'teach people' that blu-ray is a scammer snake-oil thing & 144 Hz monitors are a scammer snake-oil thing & all this beautiful sounding amplifer & DAC equipment are scammer snake-oil things. If someone asked me for evidence that blu-ray is visible compared to DVD then I can't even find that evidence.
Audio enthusiasm is a low-income area where most of the product designers are not making much money, even medium income in audio enthusiasm is rare, the high income area in high quality audio is only rare exceptions such as Sony & Audio Technica & so on. Then low quality fashion/brand-consensus audio such as Skullcandy created a net worth of at least $177 million Usd. So, if people want to look for scammers they should go to the real estate sector or smartphone sector or financial sector where lazy people make $100,000 a week doing nothing, for their overpriced smartphones (like the $2,200 Usd+ Oppo, Huawei, Apple, Sony, Samsung), or their inflated overpriced real estate portfolio with 100 tons of cement & wood & stalactite, valued at $60 million dollars, or the expensive financial sector where consulting firms charge $50,000 Usd just to explain to someone how to run their company; I am pretty sure the Shanghai saying, " when someone is does nothing & counts the huge money, you have found the scammer. " is the accurate saying to locate the scammer [i.e. not in audio/visual].
Thank you so much for your refreshing comment.
Correct, the time-domain sector in millisecond patterns provides information on the family of instrument.
Then underneath the family of instrument, you get type of instrument. Then underneath the type of instrument, you get coloration / tone / timbre.
So! Here is the part where I start to care.. for instance, if someone changes a violin string from a Pirastro Chorda (Pirastro Chorda Violin String Set 4/4 STARK | eBay), to a Pirastro Eudoxa (Eudoxa Violin String Set 4/4 Wound E Loop (RIGID D,G) | eBay), then the tone / colour difference in the acoustic audio signal is just some nanosecond constant fluctuations......
Yes.
Last edited:
It looks completely believable, he says he applied distortion to vocal recordings until there was a subtle change, so he's looking for something which is easy to pick up in distortive characteristic.
He didn't say he's looking for a noisy sound like pink noise at -130 SPL.
For example if someone plays a chainsaw sound from a speaker at 70 dB then I need to hear a distortion of the chainsaw sound, I'm guessing most people will not hear a minor distortion, it's already very noisy & distorted sound. A distortion of a chainsaw, I assume there will not be any interesting result.
Then if there is pink noise played in a soundproofed very very super quiet room, ok, they performed those tests already. They already know the approximate level of pink noise someone can hear inside a very very quiet soundproofed room.
The amount of distortion applied to a 60 Hertz sine wave is easy to differentiate 0.001 THD versus 0.005 THD since I just tried it yesterday & not in a quiet environment, I am not listening for anything super quiet I am listening for a change. It's like on a pc monitor, a quiet change would be a change in brightness, amplitude, I am listening for a change in colour, it has nothing to do with quiet.
He says in his environment the " SPL is low ", so I assume he's saying the music / sound is played at normal volume. Not played at let's say 125 dB volume (the sound of a jet engine) while listening for 5 dB (the sound of a pencil scratching on paper), which 'sounds impossible'. He's just looking for a distortive change in texture or coloration & not a change in amplitude.
Again I assume the texture change of a chainsaw is difficult, while the distortive color / texture change of two humans having a conversation into a condenser microphone is much easier.
His description sounds like the difference between say LT1115 & LME49720 which is a change of color / texture / tone / timbre at normal volumes. There is also some phase characteristic for example the AD797 seems to have more phase accuracy or inaccuracy like there is something happening there.
It's difficult to use exact analogies or descriptors, although it's a change in color / tone / texture / timbre, it's definitely never trying to hear a noise floor or buzzing sound.
Plus if I could hear the noise floor of the LT1115 at the sound of an ant scratching it's antenna against another ant at 0.0000000009 whatever decibel, then I would need to be a cyborg-alien on a different planet, so it's not a serious conversation.
~-*-~
On this topic of THD, I am not sure if THD numbers are responsible for DAC sound since all the different DAC chips have quite similar THD numbers, I don't think people should obsess about which DAC achieved the lowest THD on the planet ever, for example I will still enjoy listening to the sharper & edgier & phasier AD797 sometimes, compared to the ultra-transparent fresh sounding LME49990. I like the sound of ES9018 however in my LG smartphone it's only ok, it's not stellar like when you hear ES9018 from a desktop DAC with AD828 in the line-out then it's like... what is this!!
-
The OPA1611 doesn't sound fresh & natural like the LME49990.
The OPA1611 only sounds transparent & detailed like it's really a step above all the standard chips in transparency although it sounds kind of empty like there is something missing. (plus I placed LT1010 in the buffers of one of my iBasso, then changed to zero buffers, so I already know increasing the mA output does not change the OPA1611 signature / lack of signature).
Last edited:
Well thank you but... i don't find my comments refreshing: with the definition of a timbre it was the first 30mn of the introduction to the 1 year and half course i used to teach to future audio engineer (students) 15 years ago in an audio engineering school ( whose mother ship/headquarter was located in Byron Bay )...
This is the basis.
Ok let me comment your first example: by locating your sub away from your tweet you did not really introduced some 3d information in mono but you created an aberation in group delay. It may give a feeling of space but it is not really how that works when you are tracking ( recording) something.
If you want more info on how we perceive acoustic clues i repeat take a look at David Griesinger work. He is the guy who created Lexicon which is the benchmark when it comes to professional reverberation units ( he now work on different things (related) which are worth a read believe me).
)...This is the basis.
Ok let me comment your first example: by locating your sub away from your tweet you did not really introduced some 3d information in mono but you created an aberation in group delay. It may give a feeling of space but it is not really how that works when you are tracking ( recording) something.
If you want more info on how we perceive acoustic clues i repeat take a look at David Griesinger work. He is the guy who created Lexicon which is the benchmark when it comes to professional reverberation units ( he now work on different things (related) which are worth a read believe me).
Last edited:
Hi indra1,
I'm sorry, I don't understand your question. All DACs require a filter, its just a question as to what order roll-off they are. High order filters have terrible phase response, some may ring badly. Higher sampling frequencies allow the use of low order filters (a very good thing).
-Chris
I'm sorry, I don't understand your question. All DACs require a filter, its just a question as to what order roll-off they are. High order filters have terrible phase response, some may ring badly. Higher sampling frequencies allow the use of low order filters (a very good thing).
-Chris
You mentioned :
You are saying >88k would still require the use of a filter but can but less nasty lower order type?
Hi indra1,
Yes, exactly. Double the frequency again and you end up with a much more benign filter. That's why there is a push on to higher sampling rates. The sound quality increases as the sampling rate and depth go up.
From my own direct experience, this is true. I have worked on straight DACs since the sampling rate was 44.1 KHz. There are other tricks too. One of the nicest are the PCM1702 like chips that are collinear. Look up that data sheet and have a read.
All these new chips (multi-bit) can run at low speed, but why would anyone do that? You throw away a massive advantage. You will notice that almost no people pushing "NOS" DACs use a reconstruction filter (that's what it is called). That shows you how little they understand about what they are doing. They get something that works (= functions), but it doesn't work properly.
Ever see the signal before it hits the filter? Do that if you have a chance. It will explain everything.
-Chris
Yes, exactly. Double the frequency again and you end up with a much more benign filter. That's why there is a push on to higher sampling rates. The sound quality increases as the sampling rate and depth go up.
From my own direct experience, this is true. I have worked on straight DACs since the sampling rate was 44.1 KHz. There are other tricks too. One of the nicest are the PCM1702 like chips that are collinear. Look up that data sheet and have a read.
All these new chips (multi-bit) can run at low speed, but why would anyone do that? You throw away a massive advantage. You will notice that almost no people pushing "NOS" DACs use a reconstruction filter (that's what it is called). That shows you how little they understand about what they are doing. They get something that works (= functions), but it doesn't work properly.
Ever see the signal before it hits the filter? Do that if you have a chance. It will explain everything.
-Chris
It's worth noting here that virtually everything we "know" from science is based on statistical measurements.
I'm not absolutely sure that no one can hear it, but I do know from experience as well as reading that even substantial distortion of only second harmonic is hard to hear.
Indeed that happens. If an amplifier has 1 percent second harmonic distortion (1/100th the fundamental) and is played through a speaker with 1 percent second harmonic distortion (also 1/100th the fundamental), the amplifier's second harmonic will be distorted by the speaker and come out as 0.01 percent (1/100 * 1/100 or 1/10,000th the fundamental) 4th harmonic distortion, as well as the two 1 percents possibly adding or even subtracting, depending on their phase relationship. To understand that, look up addition of complex numbers in polar coordinates.
Your (and so many others') constant use of THD (Total Harmonic Distortion) is like biting into a food and saying "I taste fruit!" It could be apples or oranges, and me trying to explain why is about to drive me bananas.
But yes, solid-state electronic devices usually generate high-order harmonics (many times the input frequency - long story short, it's all because of that evil feedback), and even at much lower levels they're a lot more audible than "low-order" (as in 2nd) harmonics that loudspeakers generate.
Read the links in one of my previous posts about the GedLee Metric. Or watch that video interviewing Mr. Geddes. I quite enjoyed it myself. The interviewer asked this very question about distortion, and Geddes mentioned MP3s and the masking effect, and uses this to explain how we can hear some types of distortion through equipment that creates other types of distortion.
So I'll try this. Suppose you're listening to a pipe organ, the organist plays a low A note (55Hz) really loud. Now suppose the organist plays the A an octave higher (110Hz, it would have to be on another keybooard playing another rank of pipes, so the rank can be in an enclosure (see here) and it doesn't sound as loud. You won't hear it very well, and you might not hear it at all, because 110Hz is only twlce 55Hz, and due to masking the 55Hz will dominate.
Now if the low A note is played along with an A in the midrage or higher (880Hz), you'll hear them both very well because they're so widely separated in frequency.
There's a Wikipedia article on this, but I don't like it much. The example of a cat scratching and a vacuum cleaner aren't bad, but I think it goes downhill from there, and is a bit too dry and technical to effectively get the idea across: Auditory masking - Wikipedia
Total Harmonic Distortion is like Total Fruit Content. You know it's got something in it, but you don't know whether to expect mango, strawberry or what.
This is not true. Solid state amplifier that use negative feedback, can have H2 dominant distortion. Even, one of my amplifier which measured by Thimios, it have H2 dominant distortion.
- Status
- Not open for further replies.