EEVblog + The Signal Path Discussion

[Hugoneus]

Do not bother arguing with user yym. If you go and check this individual's forum profile, the user technical contribution to this forum is zero. All the user does is bashing the forum owner. Interestingly enough, that started from Batteriser. Move on and do not feed the troll.

That is why I stopped.

[josecamoessilva]

Debating with religious people is like debating with audiophools as Dave so neatly explained in one of the videos at the top of this thread.  It will go nowhere.

Yes, I saw that part too... not a high point in the video, IMNSHO.

I don't want to wade into another swamp like the debunkmania, still it's important to separate between reasonable audiophiles (vast majority), unreasonable audiophools (much fewer), and annoying audiosnobs (quite a few).

And in the interest of fairness, I should point out that a lot of people mocking audiophiles do so from a position of ignorance (not here, but at places like Head-Fi and other audiophile forums); including some who say überstupid things like "with digital, all that matters are the headphones."

[Fungus]

some who say überstupid things like "with digital, all that matters are the headphones."

It's not all that matters, but I can buy a very good headphone DAC for about $10, so... 

Datasheet: http://www.ti.com/lit/ds/symlink/pcm2704c.pdf


(there's also a version with audiophile capacitors in it for $14, but where's the fun in that?)

Edit: There's also one with audiophile capacitors AND an external audiophile op-amp for $16. Really go to town!

[josecamoessilva]

There's also one with audiophile capacitors AND an external audiophile op-amp for $16. Really go to town!

The car equivalent is: "I can buy a Fiat 500 for $15,000 so don't understand how anyone -- call them carphools -- would buy a Bugatti Veyron for $2,500,000. Basically they're the same."*

This was written for the Head-Fi forum, but it illuminates some of the traps between a digital file and power delivered to headphones:

http://sitacuisses.blogspot.com/2016/09/the-problem-with-wireless-earbuds-for.html

Most relevant for that DAC/Amp IC would be problem classes 3 and 4.

I'm a big fan of OpAmps, having used 1000s of 741s to make assorted sound machines in my teen years (analog synthesizer modules, think homemade-Moog Modular --- and if you get that reference without Google, well, hello fellow traveller), but as sound quality goes, I find them lacking in comparison to big power transistors fed by over-dimensioned PSUs, as in PSUs with minimal internal resistance. Yes, even for headphones it makes a difference.

Also, while I understand that this is no longer the 80s, so no dance parties at home, in my respectable years I do occasionally throw dinner parties, so the amplifiers have to drive speakers, it's kind of a social thing.

- - -
* Ironically, not being a car guy, I would get the same use from the Fiat 500 and the Veyron, but not everyone is like me; understanding that last clause is a major part of growing up.

[Brumby]

some who say überstupid things like "with digital, all that matters are the headphones."

I don't believe it is excessively stupid at all.

The weakest link in any sound system is the electro-acoustic transducer - a.k.a. the speakers (whatever the size).  Back in the day, it was a given.  The priority was you spend as much as you can on getting quality speakers, then work the rest of your system.

With digital, fidelity is pretty much ensured until you get into the linear signal processing - the amplifiers - and you can get good fidelity these days for quite reasonable cost.  This still leaves the speakers (or headphones) that need to be selected.  While this technology has advanced over the years, it is still the weakest link, especially when price is factored in.

So I see this statement - as literally stated - being founded on reasonable grounds, but it would be unwise for me to suggest that it's usage in the indicated circles actually follows this logic.

[Brumby]

Ironically, not being a car guy, I would get the same use from the Fiat 500 and the Veyron, but not everyone is like me; understanding that last clause is a major part of growing up.

For me, the fuel bill alone would dictate a very different usage pattern.

[nctnico]

some who say überstupid things like "with digital, all that matters are the headphones."

I don't believe it is excessively stupid at all.

The weakest link in any sound system is the electro-acoustic transducer - a.k.a. the speakers (whatever the size).  Back in the day, it was a given.  The priority was you spend as much as you can on getting quality speakers, then work the rest of your system.

True. At one point in my carreer I have designed pro-audio and PA equipment. For many of these installations getting the speakers right was the most complicated part. For the amplifying & switching part of the electronics what matters most is keeping the noise, clicks & plops and distortion low.

[josecamoessilva]

Well, since you insist here's why you need to care about more than the headphones (note that I didn't say headphones are unimportant, I said that people who think only headphones are important "because digital" are überidiots).

Even if you ignored the phase distortion that some DAC timing circuits introduce (noticeable if you listen to solo instruments like harpsichord and cello both live and in recording), the power needed for the amplification stage is crucial to sound quality. Computers, iPods, phones don't have enough power to deliver accurate dynamic range, even the 96db of a CD track. What happens is this:



You can actually see this on an oscilloscope, with a simple experimental rig; I did it in class circa the late 70s, but the image above is a simulation (from the post I linked before). My own amplifiers are 150W/ch, with a 1000W power supply in a separate enclosure. But then again, I can tell the difference between the house red and Screaming Eagle 1997; musically speaking, that is.

If you don't understand why the power supply matters, consider the Rint in the picture below. That's 1/3 of what matters, but it's the most obvious problem. Louder sound raises i, which leads to a drop in voltage Vin which leads to amplitude compression (not clipping like output stage saturation). Very noticeable in iPods if you know how to tell the difference, but you can always find a test pattern MP3 and an oscilloscope.



(Yes, it's just one side of the balanced power supply; because simple examples make the point clearer.)

Most audiophiles are reasonable and know that any real difference will be visible in an oscilloscope or a spectrum analyzer. Most people criticizing them just can't bear that others know what they like and are willing to pay for it.

As I said too, there are some audiosnobs (B&O, Beats, anyone?) and some audiophools (audiophile-grade cat-5 cable for $1000), but the vast majority of audiophiles are neither.

[Brumby]

Well, since you insist here's why you need to care about more than the headphones (note that I didn't say headphones are unimportant, I said that people who think only headphones are important "because digital" are überidiots).

It depends on how literally you are going to take the statement.

[GeorgeOfTheJungle]

With digital, fidelity is pretty much ensured until you get into the linear signal processing

Well not really, here you can see what compression can do to the poor audio signal:



I remember an electronics magazine of the 70s: ETI (Electronics Today International) IIRC those guys used to publish hifi amp designs almost monthly, improving here and there this and that month after month. Back then the power mosfets were something new. I'd bet they would have had a heart attack if they had known that in the future we were going to be doing that to our audios.

[DJohn]

I've only watched the first episode but I have to comment on the engineers and deadlines reference. That is so me. I can procrastinate until the cows come home, Bob's my uncle or being trapped as a young player, and then finish a sprint to the deadline and sort of get the product out. Have been doing that since attending university in 1980 really.

Anecdote time, with a dash of Myers-Briggs, so take it with whatever quantity of salt you like.

Once, when the management at a previous job made the apparently delusional decision that I might make a good team lead, I attended some 'leadership training'.  I ended up escaping the lead position, but the training was very interesting and I'm glad I had the chance to do it.

At the start, they gave us all the Myers-Briggs test, and split us into the appropriate groups for a series of exercises intended to show the differences in each axis.  One of them in particular stuck in my mind.  We were given a collection of kindergarten-level construction materials (paper, foil, glitter, pipe-cleaners, and so on) and told that the instructor wanted us to make a star.  We had five minutes.  Group A set to work making a plan, assigning each person a role, and made steady progress towards their star.  Group B (mine) started playing with the materials.

After a while the instructor announced that we had one minute left, and she had changed her mind.  Now she wanted a boat.  Group A nearly threw things at her.  They were NOT happy.  But group B... the sudden challenge felt good.  Someone's playing had resulted in something that looked a bit like a sail, and within seconds we had a boat.

When you know where you are, where you want to go, and how to get there, group A are brilliant.  They'll get the job done quickly and efficiently.  When things are uncertain and changing rapidly, group B will pull a miracle seemingly out of nowhere.

After reflecting on the way that I work, I am led to believe group B's 'playing' is an exploration of the problem domain.  Instead of steady progress to a known goal, we get familiar with the area around it, until we reach the point where we can quickly put together the things we've learned in a solution to any related problem.

That's useful in an emergency, but we can be a little sloppy with the details at times.  In the real world, you need both skills.

(Group A's boat was judged to be the best)

And anyone who doesn't like cats is wrong.  That's scientific fact.

[Fungus]

With digital, fidelity is pretty much ensured until you get into the linear signal processing

Well not really, here you can see what compression can do to the poor audio signal:

a) 128Kbits mp3 - 'nuff said.
b) How much of that is actually audible?

[Brumby]

With digital, fidelity is pretty much ensured until you get into the linear signal processing

Well not really, here you can see what compression can do to the poor audio signal:

Can't argue that ... but what I meant was: Once the signal is encoded (ie before it leaves the studio) fidelity is pretty much ensured until you get into the linear signal processing.

b) How much of that is actually audible?

That's a key part of the digitising process.

[Cliff Matthews]

@DJohn - Postmodern levels of clarity and the inability to distinguish "doesn't like" and "hate" is scientifically disturbing as well (a hallmark of the lost art of reading and video enchantment?).

[Fungus]

http://sitacuisses.blogspot.com/2016/09/the-problem-with-wireless-earbuds-for.html

Most relevant for that DAC/Amp IC would be problem classes 3 and 4.

The page doesn't exactly start off well...



Somebody needs to watch this a few times: https://xiph.org/video/vid2.shtml

The bit at 6:00 will be especially interesting.

Most relevant for that DAC/Amp IC would be problem classes 3 and 4.

Problem 1: Irrelevant

Problem 2: The chip in that DAC is responsible for outputting the individual samples at a stable rate, it doesn't depend at all on the timing of how it arrived there. It has an internal PLL running at 12MHz and something called "SpAct" for timing.

Problem 3: Red herring. Dynamic range simply doesn't work that way. A quiet sound doesn't have the same dynamic range as a loud sound.

Problem 4: I don't know much about that but I assume the engineers who make the chips do know how much current is needed for a pair of headphones and they make their chips accordingly. The datasheet says the output is in the two-digit milliwatt range. That's not a lot of current, a decoupling capacitor might be needed because of inductance but I don't see it as much of a challenge for a chip designer to output that much power.

FWIW: I own two of those, one with the extra op-amp and one without. ON paper the extra op-amp can supply three times as much current as the bare chip but I can't hear any difference between them on my fancy Audio Technicas. Maybe it's only needed for more 'difficult' headphones.

Speakers? I'll simply note here that nobody has yet taken up James Randi on the challenge to tell the difference between monster cables and Walmart lamp wire. Surely one has more resistance than the other! You could win a million bucks!

(Has anybody here actually measured instantaneous current going into a pair of speakers? Do we really need huge amounts of amps like those big amplifiers claim they can supply?)

[GeorgeOfTheJungle]

b) How much of that is actually audible?

Every bit, more so when it's an mp3@128k hahaha.

[josecamoessilva]

Somebody needs to watch this a few times: https://xiph.org/video/vid2.shtml

Whoa! I'll make sure to watch that to explain how sampling, quantization, and signal compression work... No.



Here, read this intro textbook, do the exercises. That's how you learn engineering (doing the exercises), especially when it gets technical.

Edit: Actually, given the context of your comment, you might want to read this one first (and do the exercises, that's how one learns engineering; that and building machines after doing the exercises):

[coppice]

http://sitacuisses.blogspot.com/2016/09/the-problem-with-wireless-earbuds-for.html

Most relevant for that DAC/Amp IC would be problem classes 3 and 4.

The page doesn't exactly start off well...

At least a ZOH view of signals is a step up from join the dots.

[Fungus]

On a side note: That page does make an interesting point about wireless earbuds in that left+right won't be exactly in phase, there will always be a slight difference between them (unless Apple is very clever and can reset their clocks wirelessly).

It's probably no worse than sitting a few millimeters from the exact center of your speakers though.

[nctnico]

If you don't understand why the power supply matters, consider the Rint in the picture below. That's 1/3 of what matters, but it's the most obvious problem. Louder sound raises i, which leads to a drop in voltage Vin which leads to amplitude compression

Buzz.... wrong! You are forgetting that the amplifier's control loop takes care of this. Google for PSRR. Since most current is supplied at the really low frequencies the PSRR doesn't even have to be really good to deal with this. Ofcourse this assumes the amplifier has been designed with enough headroom in the power supply (think 100Hz or 120Hz ripple on the capacitors!) not to clip the signal but that goes without saying. This also debunks your whole theory about iPhones. Sure they can supply 96dB but don't ramp the volume up so far they clip. IMHO you should read more real theory!

[Hugoneus]

If you don't understand why the power supply matters, consider the Rint in the picture below. That's 1/3 of what matters, but it's the most obvious problem. Louder sound raises i, which leads to a drop in voltage Vin which leads to amplitude compression

Buzz.... wrong! You are forgetting that the amplifier's control loop takes care of this. Google for PSRR. Since most current is supplied at the really low frequencies the PSRR doesn't even have to be really good to deal with this. Ofcourse this assumes the amplifier has been designed with enough headroom in the power supply (think 100Hz or 120Hz ripple on the capacitors!) not to clip the signal but that goes without saying. This also debunks your whole theory about iPhones. Sure they can supply 96dB but don't ramp the volume up so far they clip. IMHO you should read more real theory!

What is the nature of disagreement here? Sorry, it looks like I am missing something.

[nctnico]

If you don't understand why the power supply matters, consider the Rint in the picture below. That's 1/3 of what matters, but it's the most obvious problem. Louder sound raises i, which leads to a drop in voltage Vin which leads to amplitude compression

Buzz.... wrong! You are forgetting that the amplifier's control loop takes care of this. Google for PSRR. Since most current is supplied at the really low frequencies the PSRR doesn't even have to be really good to deal with this. Ofcourse this assumes the amplifier has been designed with enough headroom in the power supply (think 100Hz or 120Hz ripple on the capacitors!) not to clip the signal but that goes without saying. This also debunks your whole theory about iPhones. Sure they can supply 96dB but don't ramp the volume up so far they clip. IMHO you should read more real theory!

What is the nature of disagreement here? Sorry, it looks like I am missing something.

josecamoessilva seems to think that a drop in the power supply leads to a drop in the output amplitude of an audio amplifier. That just doesn't happen unless the power supply voltage drops below the design limits of the amplifier.

[Hugoneus]

If you don't understand why the power supply matters, consider the Rint in the picture below. That's 1/3 of what matters, but it's the most obvious problem. Louder sound raises i, which leads to a drop in voltage Vin which leads to amplitude compression

Buzz.... wrong! You are forgetting that the amplifier's control loop takes care of this. Google for PSRR. Since most current is supplied at the really low frequencies the PSRR doesn't even have to be really good to deal with this. Ofcourse this assumes the amplifier has been designed with enough headroom in the power supply (think 100Hz or 120Hz ripple on the capacitors!) not to clip the signal but that goes without saying. This also debunks your whole theory about iPhones. Sure they can supply 96dB but don't ramp the volume up so far they clip. IMHO you should read more real theory!

What is the nature of disagreement here? Sorry, it looks like I am missing something.

josecamoessilva seems to think that a drop in the power supply leads to a drop in the output amplitude of an audio amplifier. That just doesn't happen unless the power supply voltage drops below the design limits of the amplifier.

I see. Well in some circumstances when you design the amplifier and have a limitation on how high your supply can be, this could be a big program. For a lot of applications you just make sure that the supply is well decoupled and everything has sufficient PSRR. But of course as long as you have sufficient headroom you should be OK.

In car audio those crazy people who put 100's of Watts into their cars have to use huge multi-Farad capacitors to quiet down the power supply and avoid drops/distortion. I detest loud car stereos, why some people think everyone in a half-mile radius wants to listen to their terrible music is beyond me.

[rx8pilot]

TOPIC CHANGE:

Shariar - In your day to day work, how much of your time is spent on practical experiments? One of my high school friends is also a PhD EE that works for a prominent computer manufacturer. From the way he has described his job - he has lab technicians that do most of the setup and execution of various experiments that he designed. He has to make a specific effort to stay on top of the tools and techniques to make sure he is taking advantage of all available resources. Is that true in your world as well? The way you setup, conduct, and interpret your video experiments shows that you have pushed the button yourself enough to be nearly second nature. My gut feeling is that you have the personality that would go nuts if you never pushed the buttons and did the experiments yourself.

The Bell Labs lore is that folks in your position kind of design their job to best fit they way they work best - curious if that is based in reality.

[xrunner]

Shariar,

Hello! Watched most of your videos. As a cat owner staff myself I know how they are. Mine currently likes to take over my computer chair and sit by while I do bench work. I have seen your gray cat in several videos. Does yours like to take naps on warm test equipment?

Thanks for reading my question and for your informative videos.