interesting article about "transistors vs tubes" in audio

[Sceadwian]

People seem to forget something about audio with these increased sampling rates, the sampling rate doesn't only effect the frequency reproduction ability of the audio but the phase reproduction. What relies on a high definition of phase reproduction? Multi-channel (positional) audio. The BASEMENT spec for Blueray audio sampling is 48khz, and it goes up to 192khz.

Professional audio studio's record at this rate from multiple microphone's or digital spatial effects can be added with incredible precision at those sampling rates by very carefully processing an manipulating the delay between channels.

I'm a hardliner though, although I can appreciate the effects tube amplifiers have on some audio this can be designed into the audio itself! You can actually post process 'crap' audio signals through something that mimics the effects of a tube amp and play them back on modern solid state equipment with the same effect! In my opinion the amplifier design should never flavor the audio, there is one and only one purpose for and amplifier and speaker system, and that is to as accurately as is possible given current technology reproduce the signal.

The rest should be left up to audio signal engineers.

 

[unclejed613]

this can be designed into the audio itself!

and indeed they do.... only one problem, if you DSP model an amp to sound like a Marshall stack (or VOX 60 or Fender Bassman), every amp you add this DSP magic to sounds like the SAME Marshall stack., or VOX 60, or Fender Bassman. all of these classic amps each had their own unique sound, and you could play two Marshall stacks side-by side, and no two sounded identical. they might sound real close, but not identical. this was due to the use of components, many with tolerances of 10% (or more, because 20% tolerance resistors weren't uncommon in the 50's and 60's), and carbon composition resistors also contributed unpredictable nonlinearities and noise to the signal. i've worked in music stores, and guitarists can be very picky about the sound of an amp.

on the other hand a PA amp, or an amp used for reproduction, should be as close to "a piece of wire with gain" as possible.

 

[Sceadwian]

and you could play two Marshall stacks side-by side, and no two sounded identical. they might sound real close, but not identical. this was due to the use of components, many with tolerances of 10% (or more, because 20% tolerance resistors weren't uncommon in the 50's and 60's), and carbon composition resistors also contributed unpredictable nonlinearities and noise to the signal. i've worked in music stores, and guitarists can be very picky about the sound of an amp.

Yep, you can even post process THAT in.
As I said, it's all up to the audio signal engineers that study this! All you have to do is process each channel separately with different randomization on the relevant components.

 

[Mr RB]

Except that CD Players have a steep roll-off filter for that exact reason, as is standard in D2A systems.

Yep that's my point (and Crutschow's). The amp has no need to be able to reproduce frequencies above 20kHz.

People seem to forget something about audio with these increased sampling rates, the sampling rate doesn't only effect the frequency reproduction ability of the audio but the phase reproduction. What relies on a high definition of phase reproduction? Multi-channel (positional) audio. The BASEMENT spec for Blueray audio sampling is 48khz, and it goes up to 192khz.
...

Sure, but that's not the frequency of the sound it's the frequency of the digital sampling. I have a digital multitrack in my home studio that samples at 48kHz, with 96kHz internal processing freq if I remember right. But all that does is reduce the digital aliasing errors and improve sound quality, it doesn't mean the music will have any 48kHz sound content!

In fact it's the opposite, the sampling freq is raised high so the aliasing content is moved far above the roll off frequency applied to the music itself in analogue, so again that doesn't mean the amp needs to be able to reproduce frequencies above 20kHz.

Anyway I'll stand by my point that the detrimental effect of the speakers, and room characteristics, and people's ears have MUCH more effect on the sound than the modern solid state amps which are incredibly good at reproducing the sound. Has anybody else here spent time in a recording studio with live rooms or anechoic rooms? Or heard good monitor speakers with flat frequency response (not nasty coloured "audiophile" speakers)?

I get annoyed with people (not so much here, but like many people on the DIYaudio forum) that think they are expert at "listening" to music but have no clue to the technical details of the equipment and processes used to make, record, mixdown and playback the music.

 

[Sceadwian]

Mr RB, you must have missed the entire phase reproduction section...

 

[Nigel Goodwin]

Mr RB, you must have missed the entire phase reproduction section...

He probably ignored it as irrelevent, as I did

How can you talk about posible phase errors in the same thread as valve amps, who's high distortion, poor frequency response and colouration of the sound (plus variable phase shifts through the amp) alter the sound drastically from the original.

 

[tcmtech]

NOTE: This was drawn as the theory before I actually built it, the final circuit was close but with some extra components.


I will find the the Video of it in operation

The Distortion was crazy the fidelity was low, but it one of those things I built because I wondered if it was possible.

The power came from a 2HP AC motor that would belt drive both alternators at the same RPM.

The faster it was spun the more the 'amplification' haha.

Will also dig up some of my old Mag Amp projects.

I will start a new thread for these.

Well you certainly have my curiosity now!

 

[Sceadwian]

Nigel, I didn't say anything about phase error.... I said PHASE REPRODUCTION. The simply delay of a 22khz signal with sufficient resolution to provide fine aural positional details needs to run at WELL more than 22khz for a 22khz signal. If the phase reproduction is kept at the same rate as the sampling audio the positional quality of the audio at it's absolute frequency limit is limited to left or right only with no leaning! As the frequency goes down from the sampling rate the left/right positioning becomes more highly detailed.

A 22khz signal in a 192khz sampling range for delayed channels has at least 8 possible physically perceived positional ranges from left to right. At 22khz it has only left or right.

This is basic math, work it out.

 

[Mr RB]

Yeah sorry Sceadwian I did dismiss it as irrelevant because I assumed we were talking about slightly different things. An audio channel (say the right channel on a CD) is typically a 16bit 44.1kHz digital waveform, a mono waveform. That is filtered using an analogue (or digital+analogue) filter into a very close approximation of the digital waveform, but with less HF content done deliberately to reduce the aliasing errors. Then that mono channel is played back through a channel of an amp (basically a mono amp).

I'm not sure I understand your point? In that digital mono channel that is then amplified by a mono amp, where is the phase issue? Surely the amp just needs to have enough bandwidth to accurately reproduce the mono waveform it receives from the CD player, and that will be shown in the freq response and THD figures for the amp?

 

[Nigel Goodwin]

Nigel, I didn't say anything about phase error.... I said PHASE REPRODUCTION. The simply delay of a 22khz signal with sufficient resolution to provide fine aural positional details needs to run at WELL more than 22khz for a 22khz signal. If the phase reproduction is kept at the same rate as the sampling audio the positional quality of the audio at it's absolute frequency limit is limited to left or right only with no leaning! As the frequency goes down from the sampling rate the left/right positioning becomes more highly detailed.

A 22khz signal in a 192khz sampling range for delayed channels has at least 8 possible physically perceived positional ranges from left to right. At 22khz it has only left or right.

This is basic math, work it out.

Nothing worth working out, the various distortions of a valve amp make anything so utterly minor completely pointless.

I would also suggest you're wasting your time bothering about phase positioning at 22KHz

 

[Blueteeth]

Oh wow, can of worms! Solid state amps, valve amps, to speakers, to digital sampling. I too get annoyed by the audiophile followers who buy into using opamps with 100Mhz of bandwidth for audio signals (conveniently ignoring the fact the opamps other specs are awful for audio). But to get back on track, I usually think solid state for accuaracy, valve for vintage/novelty. Trying to seperate the 'engineering' and 'how they sound' of each one is like analyzing the colour spectrum of paint, and then asking people to name their favourite colour

[rant]
I totally agree with Mr RB about 'other factors' being far more detrimental to sound reproduction than the amplifier itself. Speakers, room size/shape/material, and the most dynamic of all - the human perception of audio. Listen to bass-heavy music for 10 minutes, then listen to your favourite track - it'll sound different. Similarly, listen to something loud, then your favourite track at the original volume - it'll sound different. As incredibly detailed models of the physical human ear are, the psycho-acoustic model, whilst well documented (being the corner stone of almost all lossless compression techniques) still has wide variation among people - as well as variation within the individual based on time of day, what they've heard previously, and even what they see/smell/taste.

From an engineering point of view, I think there will always be a battle between those with scopes, trying to maintain 'consistancy' in reproduction, and those with $2400 systems trying to justify their purchase to their spouse. Then there's those who walk the line inbetween, designing 'audiophile' circuits/products and selling them at high prices - because poeple will pay it purely for the sake of owning a brand name. And of course, all manner of shades in between.

As a guitarist, I really do want a valve amp - or at least I did up until a few years ago when digital technology started to blur the line between 'true valve' and all its poor reliablity, low efficiency, weight, and digital processing (with many claiming it 'sounds' digital, even though they are obviously hearing in analogue). That said, again going back to Mr RB's statement - speakers can compeltely change a system, for better, or worse, thats up to the persons perception. I've used headphones for everything (even piano) for years, and I'm afraid its spoiled me to the point where I really shun the idea of speakers, unless I need things to be heard by more than one person, or it needs to be loud.

'Sound scape' is all well and good for films, but it'll never be accurate because the dynamic range is so compressed - and rightly so, otherwise explosions would deafen us, or rupture internal organs, and we wouldn't hear the cheesy lines whispered in romantic comedies. Also, films sets in rooms, are rarely in an actual 'room', but a two wall set - so do they post process the audio to add in the reflection fom the two missing walls? Just to please someone who watches films with their eyes closed? Unless a system can detect where exactly the listener is in a room, and adjust the pbhase characteristics of a 7.1 system, as well as analyze the room specs - it quickly gets to the point of diminishing returns as more of a way of making money rather than pracitcal and 'useful'. Of course I'm bound to be proven wrong in the next few years :/
[/rant]

BT

 

[DJDAudio]

I still feel the problem lies with the irons, I just can not see a single transformer capable of handling that broad a frequency range without some adverse effect on the audio stream.

 

[Nigel Goodwin]

I still feel the problem lies with the irons, I just can not see a single transformer capable of handling that broad a frequency range without some adverse effect on the audio stream.

Fairly obviously not.

 

[unclejed613]

output transformers introduce a lot of phase errors in signals over the audio spectrum, and a lot of amps in the 60's used output capacitors, which also introduce their own set of phase errors. it's best to leave reactive components out of the signal chain wherever possible. one place where it seems unavoidable, though is in the speaker system. but that's one reason for having a very low output impedance in the amplifier. it may not eliminate the reactive effects altogether, but it does help.

 

[bountyhunter]

[rant]
speakers can compeltely change a system, for better, or worse, thats up to the persons perception. [/rant]

BT

It's not just perception, speakers mangle the sound more than all the other factors combined.

 

[Mr RB]

Don't forget the room acoustics!

Fortunately there are some VERY nice goodies now like this for $350;

**broken link removed**
**broken link removed**

It's an intelligent DSP spectrum analyser/compensator etc. You connect it before your amp, to compensate all speaker and room errors.

It actually has a "learning mode" where you connect a microphone were you would sit, and it plays frequencies and monitors the result and then adjusts all its parameters to remove errors caused by the amp, speaker and room frequency response. So you can get close to a perfect frequency response at least in that one exact position where YOU sit!

The features are pretty amazing...

 

[bountyhunter]

Don't forget the room acoustics!

No question, that's the second largest offender for buggering up the sound. Speakers need to be about 8 feet or more from all walls, floors, and celings just to minimize the resonances. Compensating equalizers are cool, I have one (don't use it anymore). If I still had good hearing, I'd probably buy one of those digital ones.

 

[tvtech]

Speakers need to be about 8 feet or more from all walls, floors, and celings just to minimize the resonances.

Not quite correct for all speakers. I have a pair of Wharfedale speakers from the Eighties and they are back-ported and need to be on stands off the ground but however close to a wall behind them. Amazing bass from such small speakers.

Cheers
TV TECH

 

[Sceadwian]

I don't care what kind of equalizer you use, you can't fix bad acoustics in a room, you might be able to slightly compensate for the worst of it, but you can't fix it, it's just the natural resonances of the physical space. I'm not sure if 350 is worth the price, but to each their own! I'm sure not gonna tell anyone what to spend their toy money on =) Personally if I want really good quality audio I'd buy a pair of studio quality monitor headphones. Even in a bad room slight tweaking of the channel delays on a typical surround system will get it to the point where it's pleasing.

 

[Nigel Goodwin]

I don't care what kind of equalizer you use, you can't fix bad acoustics in a room, you might be able to slightly compensate for the worst of it, but you can't fix it, it's just the natural resonances of the physical space. I'm not sure if 350 is worth the price, but to each their own! I'm sure not gonna tell anyone what to spend their toy money on =) Personally if I want really good quality audio I'd buy a pair of studio quality monitor headphones. Even in a bad room slight tweaking of the channel delays on a typical surround system will get it to the point where it's pleasing.

I suspect you're rather missing the point - the Behringer is a PA device, not a home device.

It's normal to EQ the room in PA, analysers and equalisers have been common for decades.