It is currently Mon Mar 16, 2020 7:14 pm

All times are UTC - 7 hours



Post new topic Reply to topic  [ 75 posts ]  Go to page Previous  1, 2, 3, 4, 5
Go to page Previous  1, 2, 3, 4, 5
Author Message
Post subject: Re: The future of analog amps
Posted: Thu Apr 30, 2015 8:50 pm
Offline
Amateur
Amateur

Joined: Sat Dec 13, 2014 6:18 pm
Posts: 141
philipk wrote:
Same technology now in digital only there's a lot more than 255 stair steps, so it approaches a smooth curve.


This demonstrates a complete lack of understanding of how a digital representation of an analog waveform gets converted back to an analog waveform. You can find the details many places online, but the summary is that if you sample any signal at a frequency at least twice as fast as the highest frequency in that signal, you can then turn that sampled representation back into an exact duplicate of the original signal. No "stairsteps". For the loudness component, use 24 bits if you are going to manipulate the digitized signal, otherwise 16 bits gives you much better resolution than the human ear, and way more than available from vinyl.

This has nothing to do with the difference between tube amps and digital amps. Certainly we can digitally model a lot of tube amp characteristics, but I suspect there are significant nonlinearities that are audible but much more difficult to model. Again, this has nothing to do with converting digital waveforms to analog.


Top
Profile
Fender Play Winter Sale 2020
Post subject: Re: The future of analog amps
Posted: Wed May 13, 2015 11:48 am
Offline
Rock Star
Rock Star
User avatar

Joined: Tue Aug 07, 2007 2:19 pm
Posts: 8827
Worst case we'll have to build them ourselves from DIY kits.[/quote]

I already do. It's fun, the amps sound great and I learn stuff.
It's not that difficult.

_________________
Life...... It's sexually transmitted and always fatal


Top
Profile
Post subject: Re: The future of analog amps
Posted: Wed May 13, 2015 9:26 pm
Offline
Professional Musician
Professional Musician

Joined: Thu Aug 07, 2014 3:25 pm
Posts: 1023
Quote:
summary is that if you sample any signal at a frequency at least twice as fast as the highest frequency in that signal, you can then turn that sampled representation back into an exact duplicate of the original signal.



Now this is where the digital Kool-Aid leaves a bad taste in my mouth.
Lets take a 44KHz sampling rate and a 20KHz signal. That gives you about two samples for each oscillation.
If the samples were taken from a sine wave and just happen to sample at the peak and valley of the wave, you would get an accurate output from the D/A converter.
The problem, as I see it, is that its unlikely that you would actually sample at those points. It's just as likely to sample at the crossover and output nothing. The D/A converters will only go from point A to point B to point C. If something changes in between, they won't see it.
Now that's a simple sine wave with errors. If you try to sample the incredibly complex waveform that is music, sampling errors must be abundant.
This will distort or erase fine detail in the music.
Fine detail is also lost in the accuracy of the sampled voltage level, the bit depth.
This was made clear to me when I first listened to the same song through different bit depth converters. First in 16 bit, then in 18 bit, then in 24 bit. We just looked at each other, a bit bewildered, and asked ourselves," Well if digital is perfect, how come it sounds so much better in 24 bit?"


Top
Profile
Post subject: Re: The future of analog amps
Posted: Fri May 15, 2015 7:35 pm
Offline
Amateur
Amateur

Joined: Sat Dec 13, 2014 6:18 pm
Posts: 141
TimsAudio wrote:
Quote:
summary is that if you sample any signal at a frequency at least twice as fast as the highest frequency in that signal, you can then turn that sampled representation back into an exact duplicate of the original signal.



Now this is where the digital Kool-Aid leaves a bad taste in my mouth.
Lets take a 44KHz sampling rate and a 20KHz signal. That gives you about two samples for each oscillation.
If the samples were taken from a sine wave and just happen to sample at the peak and valley of the wave, you would get an accurate output from the D/A converter.
The problem, as I see it, is that its unlikely that you would actually sample at those points. It's just as likely to sample at the crossover and output nothing. The D/A converters will only go from point A to point B to point C. If something changes in between, they won't see it.
Now that's a simple sine wave with errors. If you try to sample the incredibly complex waveform that is music, sampling errors must be abundant.
This will distort or erase fine detail in the music.
Fine detail is also lost in the accuracy of the sampled voltage level, the bit depth.
This was made clear to me when I first listened to the same song through different bit depth converters. First in 16 bit, then in 18 bit, then in 24 bit. We just looked at each other, a bit bewildered, and asked ourselves," Well if digital is perfect, how come it sounds so much better in 24 bit?"


Your comments about where in the waveform you sample are simply wrong. It doesn't matter where in the waveform you sample as long as you sample at least twice as fast as the highest frequency in the signal. Your analogies don't hold up. Not a slam at you, just facts. There are some things that just require doing the math to understand; this is one of them (the math in question is the Nyquist-Shannon sampling theorem).

As far as bit depth for the loudness dimension goes...16 bits covers the range of human hearing adequately, but 24 bits is needed for signal processing so that multiple signals can be combined without roundoff error. Conversion back to 16 bits at the end will preserve plenty of information to reproduce the 24 bit signal with no audible differences.

Digital signal processing algorithms must be carefully designed, and the conversions to and from the digital domain are particularly important; maybe some of your bad experiences were due to poor designs. But that doesn't invalidate good digital designs, nor the math on which they are based.


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sat May 16, 2015 8:25 am
Offline
Professional Musician
Professional Musician

Joined: Thu Aug 07, 2014 3:25 pm
Posts: 1023
This is what I mean by drinking the digital Kool-Aid, You either believe in the accuracy of digital theory or you don't. While the math may be accurate, the practical application falls short of predictions.
Wikipedia states,"Practical digital-to-analog converters produce neither scaled and delayed sinc functions, nor ideal Dirac pulses. Instead they produce a piecewise-constant sequence of scaled and delayed rectangular pulses (the zero-order hold), usually followed by an "anti-imaging filter" to clean up spurious high-frequency content.

In other words, A/D and D/A converters do not acheive the promise of the math.
The limitations of the electronics cannot take two samples and reproduce a complete sine wave accurately. There is not enough information to predict amplitude, phase, or waveshape.
The bit depth and the frequency are one and the same as far as circuitry is concerned.
Converters are not computer controlled oscillators, they are level shifters. They can only take information and go from point A to point B. What happens in the sample interval is lost information. And in music, alot can happen in between samples.
My hands on experience with digital circuits goes back to the mid 80s, including cd players, digital amps, and digital servos

Respectfully, I just can't swallow the Kool-aid


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sat May 16, 2015 9:00 am
Offline
Rock Icon
Rock Icon
User avatar

Joined: Wed Mar 31, 2010 12:48 am
Posts: 26417
Location: Tombstone Territory
TimsAudio wrote:
Respectfully, I just can't swallow the Kool-aid


+1!

I despise kool-aid.

:evil:

Arjay

_________________
"Here's why reliability is job one: A great sounding amp that breaks down goes from being a favorite piece of gear to a useless piece of crap in less time than it takes to read this sentence." -- BRUCE ZINKY


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sat May 16, 2015 10:17 am
Offline
Aspiring Musician
Aspiring Musician
User avatar

Joined: Fri Aug 21, 2009 1:45 pm
Posts: 852
Location: SoCal323
cthulhu wrote:
TimsAudio wrote:
Quote:
summary is that if you sample any signal at a frequency at least twice as fast as the highest frequency in that signal, you can then turn that sampled representation back into an exact duplicate of the original signal.



Now this is where the digital Kool-Aid leaves a bad taste in my mouth.
Lets take a 44KHz sampling rate and a 20KHz signal. That gives you about two samples for each oscillation.
If the samples were taken from a sine wave and just happen to sample at the peak and valley of the wave, you would get an accurate output from the D/A converter.
The problem, as I see it, is that its unlikely that you would actually sample at those points. It's just as likely to sample at the crossover and output nothing. The D/A converters will only go from point A to point B to point C. If something changes in between, they won't see it.
Now that's a simple sine wave with errors. If you try to sample the incredibly complex waveform that is music, sampling errors must be abundant.
This will distort or erase fine detail in the music.
Fine detail is also lost in the accuracy of the sampled voltage level, the bit depth.
This was made clear to me when I first listened to the same song through different bit depth converters. First in 16 bit, then in 18 bit, then in 24 bit. We just looked at each other, a bit bewildered, and asked ourselves," Well if digital is perfect, how come it sounds so much better in 24 bit?"


Your comments about where in the waveform you sample are simply wrong. It doesn't matter where in the waveform you sample as long as you sample at least twice as fast as the highest frequency in the signal. Your analogies don't hold up. Not a slam at you, just facts. There are some things that just require doing the math to understand; this is one of them (the math in question is the Nyquist-Shannon sampling theorem).

As far as bit depth for the loudness dimension goes...16 bits covers the range of human hearing adequately, but 24 bits is needed for signal processing so that multiple signals can be combined without roundoff error. Conversion back to 16 bits at the end will preserve plenty of information to reproduce the 24 bit signal with no audible differences.

Digital signal processing algorithms must be carefully designed, and the conversions to and from the digital domain are particularly important; maybe some of your bad experiences were due to poor designs. But that doesn't invalidate good digital designs, nor the math on which they are based.


This must be what happens when you put an engineer and a test pilot in the same room. :lol:

_________________
Its a lifestyle.....

Image


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sat May 16, 2015 2:49 pm
Offline
Amateur
Amateur

Joined: Sat Dec 13, 2014 6:18 pm
Posts: 141
socal323 wrote:
cthulhu wrote:
TimsAudio wrote:
Quote:
summary is that if you sample any signal at a frequency at least twice as fast as the highest frequency in that signal, you can then turn that sampled representation back into an exact duplicate of the original signal.



Now this is where the digital Kool-Aid leaves a bad taste in my mouth.
Lets take a 44KHz sampling rate and a 20KHz signal. That gives you about two samples for each oscillation.
If the samples were taken from a sine wave and just happen to sample at the peak and valley of the wave, you would get an accurate output from the D/A converter.
The problem, as I see it, is that its unlikely that you would actually sample at those points. It's just as likely to sample at the crossover and output nothing. The D/A converters will only go from point A to point B to point C. If something changes in between, they won't see it.
Now that's a simple sine wave with errors. If you try to sample the incredibly complex waveform that is music, sampling errors must be abundant.
This will distort or erase fine detail in the music.
Fine detail is also lost in the accuracy of the sampled voltage level, the bit depth.
This was made clear to me when I first listened to the same song through different bit depth converters. First in 16 bit, then in 18 bit, then in 24 bit. We just looked at each other, a bit bewildered, and asked ourselves," Well if digital is perfect, how come it sounds so much better in 24 bit?"


Your comments about where in the waveform you sample are simply wrong. It doesn't matter where in the waveform you sample as long as you sample at least twice as fast as the highest frequency in the signal. Your analogies don't hold up. Not a slam at you, just facts. There are some things that just require doing the math to understand; this is one of them (the math in question is the Nyquist-Shannon sampling theorem).

As far as bit depth for the loudness dimension goes...16 bits covers the range of human hearing adequately, but 24 bits is needed for signal processing so that multiple signals can be combined without roundoff error. Conversion back to 16 bits at the end will preserve plenty of information to reproduce the 24 bit signal with no audible differences.

Digital signal processing algorithms must be carefully designed, and the conversions to and from the digital domain are particularly important; maybe some of your bad experiences were due to poor designs. But that doesn't invalidate good digital designs, nor the math on which they are based.


This must be what happens when you put an engineer and a test pilot in the same room. :lol:


I've done that before in my day job; when the video shows the pilot pressing the button he swore he didn't touch, I win :lol:

Here's a link (that I trust more than Wikipedia) that may clear some stuff up:
http://xiph.org/~xiphmont/demo/neil-young.html


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sat May 16, 2015 7:19 pm
Offline
Professional Musician
Professional Musician

Joined: Thu Aug 07, 2014 3:25 pm
Posts: 1023
This is an interesting article that doesn’t address the subject at hand. That is, it is about recorded music. Any recorded music is inferior to live guitar playing through an amplifier. Comparing an inferior product to another less inferior product is a red herring. Live guitar music is the subject and the invisible hand of the marketplace has chosen tube amps as the highest quality product. All those ears can’t be wrong.
For feature driven customers, digital is what many prefer, due to their flexibility. The digital offerings on the marketplace are mid to low-end units with many compromises for affordability. While the processors are decent 32 bit computers, the converter is usually combined CODEC that is serial driven. Then low performance analog components are used to finish the job. It’s not that high quality digital components aren’t available, it’s that they cost too much for the market niche. Over time, many musicians abandon digital for the higher quality tube amps.

A few points about the article
So high sampling rates are used to increase high frequency response that no one can hear. The argument is extended to video and uses a comparison to ultrasonics and ultraviolet light. It knocks that strawman down easily.
However oversampling is not used for ultrasonics. It’s used to reduce the sampling interval. The digital blank spots in between the data.

The entire argument over ultrasonics causing intermodulation distortion can only be taken seriously by an engineer.
Musicians call those ultrasonics harmonics and rely on them for a bright, chimey tone. The average Vox AC30 amplifier has significant response up to 30 KHz. I’ve measured Ibanez acoustic amps up to 60KHz. Those ultrasonics do come down to the audible range and musicians have always depended on that for good tone.

16 bit systems suffer from accuracy of detail. 24 bit gives nuance to music.
The 16 bit digital format of 8 by 14 modulation was chosen for CD because that was the highest rate possible at the time and conveniently had already been written by the aerospace industry for satellite communications. It is hopelessly obsolete at this point.

The “War on loudness” bit ignores the goal of increasing dynamic range. It was to match the dynamic range of live music, which can exceed 125dB. The early formats could only reach 50-60 dB for home equipment.
The article attempts to justify the superiority of digital, but is misleading and self-serving. Further, it doesn’t deal at all with live music, only digital recordings. I could go on, but it's like refuting Fox news.....
Live music has 10 to 15 significant orders of harmonics impressed on the fundamental waveform, which usually isn’t a sine wave at levels to 110dB.
This requires high-speed sample rates and 24 bits to capture and reproduce music with all that detail intact and in proper harmonic phase.
So the goal of hi rate, hi bit digital, extrapolated out, is analog.


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sat May 16, 2015 7:35 pm
Offline
Rock Icon
Rock Icon
User avatar

Joined: Wed Mar 31, 2010 12:48 am
Posts: 26417
Location: Tombstone Territory
Image

:wink:

Arjay

_________________
"Here's why reliability is job one: A great sounding amp that breaks down goes from being a favorite piece of gear to a useless piece of crap in less time than it takes to read this sentence." -- BRUCE ZINKY


Top
Profile
Post subject: Re: The future of analog amps
Posted: Sun May 17, 2015 10:58 pm
Offline
Amateur
Amateur

Joined: Sat Dec 13, 2014 6:18 pm
Posts: 141
One last try:
Digital sampling and reproduction of audio signals, if the hardware and algorithms are properly designed and employed, can and do accurately reproduce audio signals. The link I posted earlier was to refute the FUD that Tim was posting. That the link was about Neil Young's Pono snake oil is a side issue. The info at the link does a good job debunking the digital audio myths that Tim posted. If you don't agree, then I see no point in trying confuse you with more facts :wink:

However, as I said in my first post in this thread, this is tangential to the main issue, which is whether digital amp emulation can be as "good" as the actual tube amp, for some reasonable definition of good. My suspicion is that the answer is yes, BUT not in a package that will be usable outside the lab, at least for several more years. Tube amps have lots of nonlinearities and interactions among the components; I doubt that anybody working in the field of digital amp design has tried to fully characterize all those nonlinearities and put them into algorithms that can run on hardware that can be produced and sold in any quantity. And I'm not seeing the market that would justify the investment. Amps like the Mustang III/IV are clearly "good enough" for a lot of folks including gigging musicians you can find on this board. Hell, my Yamaha THR-10C is a terrific practice amp and handles a lot of my amp needs.

But I also enjoy the hell out of firing up my tube amp and wailing away. Just like I prefer cars with manual transmissions. There's a joy that goes along with connecting to the older technology that amp modeling and paddle shifters just don't capture, at least yet, at least not for me. Sometimes convenience and practicality are more important, but when they aren't, it's great to have the choice to pull out those stone knives and bearskins, and get to places you can't go otherwise.


Top
Profile
Post subject: Re: The future of analog amps
Posted: Mon May 18, 2015 7:09 pm
Offline
Professional Musician
Professional Musician

Joined: Thu Aug 07, 2014 3:25 pm
Posts: 1023
Well, this has been fun, but I'll wind down too. As you can see, I'm free with my opinions, but I do respect yours even if I don't agree.


Top
Profile
Post subject: Re: The future of analog amps
Posted: Mon May 18, 2015 8:21 pm
Offline
Amateur
Amateur

Joined: Sun Jul 22, 2007 9:29 am
Posts: 139
I’m a fan of mathematics. It’s a beautiful pure science where the equal sign means something, where things can be perfect. Theories can be proved. And the proof’s itself make sense too. The term Nyquist may mean something on paper. But Isn't that all it is, a theory on paper, an ideal. Like a perfect sphere, only in Mathematics it’s perfect, imo.

I like solid state amps, like Fender’s Champion. It has some amp/effect modeling.
Isn't it Peavey's Transtube technology (solid-state circuit) that comes closest to a tube amp, right? What gets me is the Tube-Amp fans forget how good a Solid-State amp can sound. It’s got a punch to it.

And then truer Digital-
Fender has its Mustang, Peavey has its Vypyr. What’s lacking in both is the sound decay (the sustain), it drops off (on some of the modeling selections), a decaying fuzz-factor. If Digital-Amps are so good why are there so many Tube-amp fans?


Top
Profile
Post subject: Re: The future of analog amps
Posted: Tue May 19, 2015 2:56 am
Offline
Rock Icon
Rock Icon
User avatar

Joined: Tue Jun 29, 2010 1:20 pm
Posts: 9640
Location: Indiana
That's like asking; "If vanilla is so good, why do some people eat chocolate?"

There's no accounting for taste. There are lots of people, professional and amateur, that like/use SS amps and digital modeling amps, and there are lots of people, professional and amateur, that like/use tube amps.

I prefer, and own nothing but, tube amps, but I also get a thrill putting on headphones and using my POD XT. :o

_________________
---> "The amp should be SWITCHED OFF AND UNPLUGGED before you do this!" <---

Por favor, disculpe mi español, no se llega a la práctica con mucha frecuencia.


Top
Profile
Post subject: Re: The future of analog amps
Posted: Tue May 19, 2015 5:39 am
Offline
Rock Star
Rock Star
User avatar

Joined: Tue Aug 07, 2007 2:19 pm
Posts: 8827
The future will be determined by demand. The demise of tube amps has been around for thirty years and it never happened. As long as people want them, we'll have them.

Side by side comparisons are easy enough. Take a 100 watt or even an 85 watt Twin next to a Champion 100 or a 100 watt FM 212 solid state amp. Keep cranking them bit by bit to the point of distortion. This is where you will experience a feel and different dynamic response. For myself, there's just something missing. The punch and feel are missing.


I prefer tubes, but even though I have a bunch of pedals, I rarely use them unless the music I'm playing at the time requires them. The funny thing I find on some solid state amps is a "sag setting" trying to emulate the sag of a tube rectifier. On my 18 watt clone, I installed a switch and a couple of diodes to switch out the rectifier tube. It's funny I did it the reverse of what some of the solid state amps have, but I have and actual solid state rectifier not a modeled one.

As long as people want both, they'll be both. Old technology, especially 50 year old technology, is pretty much proven technology. The old amps weren't as fussy because of the way they were manufactured. New tube amps are built for profit. This is where the Boutique guys come in, but quality costs money, but when you factor in inflation, those amps are where well built, hand wired, high quality amp should be price wise. When my Father bought me a Vibrolux Reverb in 1966, it was about $285, today it would be somewhere around $2,076.97, but people have been conditioned to accept cheap, poorly made products, throw them away when they break and get another cheap, poorly made product. Solid state stuff fails too. Transistors and IC's can fail and short out too.

So really, it just boils down to preference. I'm happy lugging around my 50+ year old technology and enjoying what comes out of it just the the players who prefer solid state gear. I look at solid state amps, not as a comparison to tube amps, but a sound all their own, because frankly they really don't perform or really sound like a tube amp to me. Some sound good, some don't. I do like Quilter amps. Probably because they put their money into making a great sounding amp, not bells and whistles and complex menus. To each his or her own.

_________________
Life...... It's sexually transmitted and always fatal


Top
Profile
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 75 posts ]  Go to page Previous  1, 2, 3, 4, 5
Go to page Previous  1, 2, 3, 4, 5

All times are UTC - 7 hours

Fender Play Winter Sale 2020

Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
Jump to: