Slashdot Mirror
Tubes vs Transistors: An Audible Difference?
cgenman writes "Are those vaccuum tubes worth the extra price? This paper, a transcript of a speech to the Audio Engineering Society of New York, indicates so, though the reason is surprising: Overloaded tubes behave better.
While the speech itself is from the early 70's, the paper takes on new importance with the recent trend in louder is better music."
Bob Carver did this already. If I remember correctly, it's the increaed resistance from a tube amp using output transformers that creates the 'soft' sound that characterizes glass audio.
Carver created a solid state amp which pretty much mimicked a $10K tube amp and no one could tell the differencec in blind tests.
I can't speak for the HiFi crowd but when it comes to Ham Radio tubes still have a job to do.
The front ends of receivers ALWAYS behave better when a tube is used because of the gradual distortion that has already been mentioned. On some of the bands that hams use receivers overload easily and the tube characteristics coupled with a high voltage power supply (80 volts or so compared with 12 volts for a transistor rig) can save the day.
Power amps for transmitters are always best when a valve or two is used. There are amps out there that use FETS and exotic technology but if you want to shove 2Kw up an antenna the only way to do it is with some heavy duty tubes.
Ed Almos
Budapest, Hungary
The more corrupt the state, the more numerous the laws. - Tacitus, 56-120 A.D.
I am shocked that this old crap has no annotation from the 1990s when phychology tests proved tubes sound more appealing than solid state op-amps.
The reason ?
Odd harmonics vs EVEN Harmonics !!!!
Odd harmonic overtones sound HARSH to human brains and are an unwelcome side effect of all solid state electronic amplification.
That was new data in the 90's that this ancient speech being discussed had no idea about.
Valve amps (the original name for tube amplifiers) are basically voltage driven, so when they distort, even-order harmonics are produced (2nd, 4th, 6th, etc...) while transistor amps are current driven and produce odd-order harmonics (3rd, 5th, 7th, etc....)
I cannot believe at the time i posted this i am still the only one to point this out.
All those years of subscription to The Absolute Sound taght me at least why tubes were better and an oscilloscope visibly points out the harmonics.
This trend really only came to light in the 90s, particularly the mid- to late-90s. Compression is used to squeeze all the dynamics out of the music in order to make it sound "louder" than the other songs on the radio. It's different from just loud rock instruments. This has to do with the wretched trend of signal compression.
Recording too loud is bad, but labels feel it gives them a comparitive advantage because it's the only way they can effect the final listening volume, and subjectively louder music sounds better.
As a Bass Player who has been in on more than a few sessions, I can tell you that my ears tell me that there is a difference between a nice Mesa Boogie or classic tube amp, and a straight transister amp.
I own both types. Both have pluses and minuses. But for bass, you can not beat the tube sound, even sythetic tube is just not the same, the ear knows.
"Who are in control, they are not in control of anything - they don't even control themselves!" - Glen Beck
Please stop spreading this nonsense. The vast majority of tubes have MUCH lower distortion than any solid state device. In fact the big DHT's (300B, 845) are probably the lowest distortion amplification devices ever made. Look at the curves if you don't believe me. Tubes at least look VAGUELY linear, transistors most certainly do not.
Solid state competes only by having very high gain and using feedback. There is absolutely no way for solid state devices to compete with tubes in terms of distortion in the forward path.
And feedback has a whole bunch of fun problems. It's great when you're driving resistors or simplified R/L/C 'dummy speakers'... but it has real problems when you drive REAL speakers. Real speakers have dozens of resonances all over the frequency range that throw all kinds of garbage back at the amplifier. Feedback has to take this trash and RE-AMPLIFY THE GARBAGE in order to cancel it out and present a lower output impedance.
With tubes (especially push-pull transformer-coupled tube amplifiers running heavy Class A) you can achieve VERY low distortion numbers with no feedback whatsoever. You do require speakers of higher-efficiency of course, but this is not hard to do. There are very good-sounding speakers in the 95db/watt range and up that can run great on tube amps in the 16w range. Horns up around 100db/watt are happy with much less.
Yes, SOME tube amps sound very 'warm' and distorted, but quite frankly, that was 5 years ago. Things have come a long way. Class A push-pull is really taking off and people are achieving EXTREMELY fast, detailed, low distortion tube amps that have all kinds of advantages over solid state.
Digging up an ancient speech which probably SPARKED the religious war in the first place is idiotic, in my opinion.
Maybe its a war for audiophiles, but for musicians, there is no dispute. The vast majority of professional musicians use tubes for the reasons stated in the article and others. Transisitors are used for different things, such as when size and heat are a consideration, like in a practice amp.
I never understood why there was a debate anyway. Tubes sound better, transistors are much easier to work with. They each have their place. You can make each sound good or bad by design, but when all is equal, tubes sound more pleasant to the ear, while transisitors look better on paper. I tend to believe my ear rather than a piece of paper. My home stereo is transistor, my guitar amps are tube. This is because I want good sound at moderate levels and excellent reliability from my home stereo. For my guitar amp, I am willing to put up with lower reliability and higher maintenance to get the dynamic range, uncolored sound, natural compression and punch that only tubes can bring.
The flame wars are pretty silly, its like arguing "horse vs. car". Obviously the car is better in most cirumstances but the horse is handier if you are where there are no roads.
Oh, the relevence today is that the quality of transistors today are not as good as they were years ago in some respects (as you state, they have changed) yet my guitar amps still use the same tubes other amps used 50 years ago: 6V6, 6L6, 5880, EL34, EL84, 12AX7, etc.
Tequila: It's not just for breakfast anymore!
The problem with this is you end up with horrible range that you can't do much with. Loud sounds end up clipped so that the softer sounds can sound 'louder'. Here's why it sucks: You lose a lot of the music's quality. When I turn up this song, my stereo dac becomes the limiting factor. When you turn up crap like this, the sound waves are already clipped. The jokes on them.
People like tube amps because they add a little bit of harmonics that sounds nicer to our ears. Tubes sound 'warm' and they fail gracefully when overdriven. It's an old battle that no one will win, but most muscians go with tube amps so they can't all be wrong
[Fuck Beta]
o0t!
h**p://club.aopen.com.tw/News/News_showAnswer_Old
and.. site with some comments.
h**p://techreport.com/news_reply.x/3670
As I recall, Don Lancaster had a treatment of how to make a transistor-based amplifier sound like a tube amp...
1) adding hum (60 and 120 Hz)
2) adding harmonics
and so on.
Since the first real Rock and Roll music appeared approaching, 60 years ago now, louder has been better. That's a "recent" trend?
That's not the way the poster meant it. You should read the first article he links to. It's about how recently CD's are made to sound louder and how this causes the music to sound bad. There are some examples or rock CD's from not many years ago that did not exhibit this awful practice.
MPEG can keep higher frequencies, but it is difficult to do so without throwing away all compression gains because there is only a single control for the 16kHz to 20kHz region. This is usually not a problem even for people who hear those frequencies because of a thing called masking. When we hear one frequency, others are essentially blocked. In normal music (though some music is different, and of course not counting white noise in silent parts of the mucis) there is no way to hear those frequencies anyway.
This whole phenomenon is well understood today. You can buy a little "tube amp emulator", with emulations for famous tube amps. Choose your own harmonic distortion. There are product lines of amp modellers.
Most of the trouble in audio today is not tube vs. transistor vs. digital. It's from artifacts introduced during compression of the dynamic range. The real problem is the car audio listening environment, which is noisy. Radio stations need to sound good in cars. This led radio stations to compress their audio into a narrow dynamic range. People got used to this. Then, when cars got CD players, CD mixes began to be compressed like car audio. ("You don't want your record to be the softest one in the changer"). Now, most popular music is so compressed that musicians have totally lost the musical use of volume. You can't have a soft passage; it will be pumped up. Sharp attacks are clipped, so that tool has been taken away. The end result is popular music that has no texture. Background music.
Izotope ozone is a (non free) winamp/directx plugin that emulates some of the distortion effects that sixties amplifiers produce using tubes. I've been using it for quite some time and it really enhances the listening experience. I can recommend it and it sure is worth the small license fee (which is peanuts compared to what you would need to invest in hardware otherwise). I haven't found any other plugins that produce a similar improvement in sound. There are many plugins that just beef up the bass a bit or add cheap 3d effects. Izotope Ozone is in a different league.
The plugin clearly demonstrates that the distortions (when used with care) can really enhance music. It also demonstrates that you can get the same effect by processing the sound digitally instead of with tubes. Izotope ozone actually goes way beyond what traditional tubes can do because it doesn't have the physical limitations.
Of course most commercial rock and pop music is processed and filtered in the studio before it is put on cd whereas older music (or indie records) tend to sound better when played back on equipment that adds the distortion effects. Of course the amount of distortion is a matter of personal taste and I find that I enjoy my music more with a little bass compression and a bit of sparkle in the higher ranges. Studios tend to optimize for cheap equipment (i.e. it has to sound nice on cheap radios) so you can gain a lot by adding some distortions.
You can also use sound distortion to compensate for lossy compression or lousy speakers. Just boost the bass digitally for the frequency range that your subwoofer can actually handle; add a little sparkle to compensate for loss of higher frequencies during the mp3 compression; add some overdrive on a guitar track. Distortion is not necessarily about reproducing sound as it was when it was recorded but about making it sound as nice/pleasing as possible. Much of the distortion effects in sixties equipment is deliberate and not accidental. Electrical guitars are a good example of how distortion can be used to produce a wide range of sounds.
Jilles
It's not that the loud parts are clipped, rather they are compressed. Yes, audio hardware/software compressors while they can be a godsend at times, their overuse has makes things sound flat, loud, and boring. (Wouldn't it be great if everytime someone whispered to you, your brain instead cranked that whispering up to the equivalent shouting db level? That's what compressors do - so that the music is always shouting).
A compressor is a device that says when the music reaches a certain decibel level, reduce the volume by X (X=compression ratio). So with a compressor you can take a song and crank it up super loud, without fear of ever actually clipping the signal or the system (it hovers right below 0db).
The result of this is that if you looked at a compressed waveform, they are no dynamics in it at all. The peaks and values of the entire wav are all maxed out. While this is louder, you have almost no dynamic range. Compression comes at a cost - most engineers these days don't seen to realize this.
CDs aren't actually recorded like this. The recordings are fine - it's when they go in to get the whole song (and CD) mastered that this happens. Audio Engineers are under increasing pressure to make the CD "sound louder" by the PHBs.
Processing signals at a resolution of 24 bits (maybe even floating point calculations, 32 bit, 64 bit???) combined with billions of operations per second can probably simulate whatever positive effects that vacuum tubes provided. The trick would be trying to figure out what those qualities are and developing efficient algorithms to implement them. The main benefit? It would be low power, easy to reprogram, and it won't change with time, temperature, or RF interferrence.
Of course, this doesn't make the analog components any less imporant. Once you get out of the realm of digital signals, there is still a great need for efficient analog circuitry. It's just how much of this circuitry you need to focus on can be reduced by using DSP.
Consider though, how much easier it is to repair tube gear than transistor gear. If a tube goes bad, you pull it out, and stick a new one in. And its usually fairly obvious which tube it is. With transistors, its entirely different. Bust out the 'scope, get a schematic, and start tracing. Ok, no signal here.. lets unsolder that part, test it... shit, its ok... hmm... maybe its the summing amp... unsolder that... nope, hmm... Overall, tube gear is really easy to fix, compared with solid state. I guess thats why so much broken solid state equipment gets trashed or replaced today, rather than fixed. Also, for an audiophile, spending $100 per year on completely new kick-ass tubes is no big deal, its totally worth it. Consider that the life of a tube can extend to over 30 years too. If the sound changes, re-bias. And if its really bad, spend the $10 on a new tube.
The other thing is that it's extremely unlikely that your speakers are able to output 20kHz at the same loudness as 10k and 16k. They will have a non-flat frequency response too, which generally goes from somewhere around 30-40Hz up to 18kHz. Reliably, anyways, which is to say without a lot of distortion. So that will affect what you hear. If you could hear the 20kHz sound at all, and the 16k, too, your hearing is probably fine. If there's a medical school around you somewhere, you could probably get it checked for free, which is interesting, and a good idea if you're going to be involved in music professionally.
Finally, yes, your hearing does get damaged only in the range of the things you hear. The little hairs (cilia) in your inner ear that respond to a particular frequency will wither if they get over-stimulated. They won't grow back on their own, but there are people working on it!
Perhaps you have ears capable of discerning a 0.00066 dB amplitude fluctuation at 40 kHz - I know I don't. Perhaps you can hear time smear of 70 ns - I know I can't. Perhaps you also have 25 meter - that's 80 some-odd feet for those of us in the US - interconnects; my whole apartment isn't 80 feet long. However, I think you have none of those things.
/. server blocks my IP until tomorrow which may (ahem) impede my discourse a bit. I do not have a /. account. Thanks for the interaction thus far.
Including the wire in any audio output transformers, coils internal to audio amplifiers, AND in the voice coils of the speakers, surely I do have quite a distance of wire involved in these low-impedance loops.
Again, I'm not selling interconnects or speaker wire OR stating that these alone would make a difference. Perhaps they would, and there are many who adamantly say that. But I'm not suggesting such a remedy. What I'm saying is the differing use of impedance between the typical vacuum tubes versus solid state audio circuits makes a difference. I'm also saying this difference is due to the skin effect, which also happens to be a function of the circuit impedance. I'm also speculating that differences in the the mechanism of switching an electron beam versus a doped crystal junction might play a role.
If my "voice through a pipe" analogy muddled things, I apologize. Certainly that was only indirectly related to impedance and skin effect. However that is the mental image I use, since there can be many shapes and internal textures to a pipe and it is relatively easy to visualize the distortion of audio waves.
PS: I do likely have an anonymous post or two left before the
What is interesting to me is the cult status that tube amplifiers have achieved. Some forty-odd years ago vacuum tube engineers (my father was one) jumped on the transistor bandwagon because of the numerous advantages it conferred over tubes. Now, for some unaccountable reason we look back at the heyday of the pentode in some twisted nostalgic fashion making unprovable claims about the wonders of the good old days. Fact is, all they were is old.
Some people like the sound of the tube amp better, others don't see any significant difference, and there are those that don't like it at all. Put it like this: what is an amplifier supposed to do? Why, it is supposed to amplify, of course, and the more precisely, predictably and accurately it does that is a good measure of the quality of the amplifier. The closer you come to achieving a one-to-one correspondence between the input waveform and the signal presented to your speakers the better your amplifier. Conversely, an amplifier that modifies, distorts or otherwise results in significant variation between the input and output waveforms is a worst a lousy amplifier and at best functioning as a signal processor in its own right.
What it comes down to is that the extremely-hard-to-model non-linear responses of a bewildering variety of kinds that you describe indicate that the tube amplifier is not faithfully reproducing the original recording and is distorting it in complex and unpredictable ways. Yes, it may do so in a pleasing manner and one may very well prefer the modified sound, I have no problem accepting that. But that is not intrinsically different from saying that I like what my 20-band equalizer or my Alesis effects processor does to the sound. And given the decades-long controversy on the subject, the presumption by tube amp afficionados that their sound is inherently "superior" is a bit hard to swallow, particularly as we are talking about one of the most subjective experiences that human beings can share. Personally, I like the sound of some of the tube systems I've heard, but for my part I wouldn't say that they are, under all circumstances, simply "better."
The higher the technology, the sharper that two-edged sword.
The 'lower' frequency you are hearing is probably caused by using a 44.1kHz DAC, which isn't really up to the job of reproducing these very high frequencies accurately. To do this particular hearing test properly, you really need to keep everything in the analogue domain.
The explanation for this is quite messy, but google the terms "Nyquist limit", "Shannons Sampling Theorem" and "Aliaising Noise" if you want to know more.
Try generating a slow sweep from 10k to 22.1k and you'll probably hear a multitiude of sounds from the DAC rather than a nice smooth rise, especially if the fundamental moves outside your hearing range along the way.
A pizza of radius z and thickness a has a volume of pi z z a
>I don't get the 20A thing, dude...
:)
V = I * R
P = I * V
therefore, substituting I*R for V, we get
P = I^2 * R
If P = 3000 watts, and it's hooked up to a standard home speaker (8 ohms), rearranging, we get:
I^2 = P / R
I = (P/R)^1/2
I = (3000/8)^1/2
I = (375)^1/2
I = 19.364916731037084425896326998912 amps
If you managed to squeeze 500 volts into your standard speaker, you'd have:
P = V^2 / R
P = 500^2 / 8
P = 31,250 watts
At a current of:
I = V / R
I = 500 / 8
I = 62.5 Amps
(I'd suggest using 00 AWG cable for that setup, LOL!) [ok, ok, you'd probably get away with 4 AWG]
Of course, input current would be higher. Also, speakers are a reactive device, so therefore these numbers are rough guesstimates. I don't feel like busting out my Algebra book to do the complex math required to give you an exact number.
If you could be told what you can see or read, then it follows that you could be told what to say or think - BoC
One reason for this, IMO, is that people are listening to more and more music in cars. Cars suck for listening to music in: they're loud. Even a quite car has got a fairly high level of background noise. This means that you've only got a limited amount of range left to present the music in, which means that listening to high-dynamic-range music just doesn't work.
(Ever tried to listen to classical music in the car? Ever found yourself adjusting the volume to make the quiet bits louder and the loud bits quieter? You've just run out of range. Modern music is easier but it tends to have a much smaller range anyway, even without compression.)
Compressing the range makes the music much more accessible in cars (and other high-noise environments). Of course, this makes it suck when you're listening to it on real audio equipment. But since radio is a major market, and most radios these days are in cars, there's a major push towards compression.
(Incidentally, as anyone who actually knows anything about audio equipment will tell you --- unless you're in the habit of listening to music in your car with the engine off, spending serious money on a car audio system is just not worth it. That background noise will ruin everything, every time. Spend the money on a digital jukebox instead and leave the high-quality audio at home, where you can listen to it in the environment it deserves.)
It's funny to see the myth about the cost of tube amps and preamps on slashdot.
Check out bottlehead.com and see a very good tube preamp for cheap.
There are tons of schematics out there, were you can build a cheap amp. Cheaper than Best Buy junk.
You can emulate valve clipping with a couple of small FETs and a handful of passive components per channel. It's basically just soft clipping, although it's easy enough to add in some hum (high-value resistor and capacitor from the top of the power supply's main rectifier, assuming a series-regulated or similar PSU), and white or pink noise (capacitor from the top of an unfiltered zener diode).
Or you can get silly about it and emulate the valve clipping and noise in each stage of the amp instead.
Got time? Spend some of it coding or testing
I can easily see that there would be less inertia involved in switching a beam of electrons than there would be with changing the saturation levels across multiple junctions of doped solid silicon.
Sorry, but that whole rodomontade just got funnier and funnier as it went on, and that last sentence cracked me up.
Having studied microwave transistor structures in both Si and GaAs, I can tell you that at audio frequencies, "intertia" of electrons, perhaps you want to mean dispersion or diffusion current velocities, is quite irrelevant until you start going into the 100s of MHz.
Yes, a legitamte concern with BJTs is time taken to discharge the newly formed "capacitor" at the reverse-biased P-N junction. No, this does not make BJTs useless, it just means you have to be smart about your circuit design - make sure there's enough current to drive the base as fast as you want it.
And, I'm sorry to sound snide, but what exactly about a high impedance circuit "favours" voltage over current? I'm no valve expert at all, but I was under the impression that valves were voltage devices! An ideal thevenin equivilent voltage source should have a low impedance!
Honestly, I can't believe so many people think audio is some kind of black voodoo magic. Try designing the frontend/filtering/amp stage for a GPS reciever, or carefully calculating intricate patterns on a PCB to create matching transformers for GHz signals using nothing but the shape of the copper!
In the beginning it was LP versus CD. (Nobody mentioned cassette, except to ask how come a bootleg recorded from an LP on a 99p ferric cassette using a 49 quid midi system sounded better than a store-bought original.) Now that the recording companies have all but killed off LP, hi-fi bores (if I called them "audiophiles" there would most probably be a mob of News of the World readers standing outside their homes, waving placards and pouring petrol through their letter boxes) need something else over which to disagree.
So we're back to silicon vs. vacuum. Now, in the 1960s and 1970s, transistors were still just expensive enough that they were still competing with valves, and a tranny amp from that vintage -- if it's been fitted with new capacitors, which degrade over time -- will sound as good as a cheap valve amp from the same vintage. It had to, because the competition was there. Today, valves are strictly in the realm of esoterica, and modern IC / transistor kit doesn't have to try to compete with them.
But it's a highly subjective area, and "scientifically perfect" reproduction (identical waveshapes, just different amplitudes) is not necessarily right for the ear. There is little doubt that the distortion characteristic of transistors is harsher than that of valves. This is because, by trying to be "scientifically perfect", they hit the supply rails easily. (Recall that valves use supply rails between 100-500V and require transformers to match to low-impedance loudspeakers; transistors are driving the speaker directly, 20W RMS at 8 ohms is 36Vp-p or +-18V). So with valves, there is more headroom. Deliberate slew rate limitation also helps, by giving a different type of distortion (never quite making it, which gives even harmonics, rather than trying to overshoot and maxing out, which gives odd harmonics). Odd harmonics are reckoned to have a harsher sound than even ones. In fact, modern op-amps, with almost DC-RF bandwidth and consequently slew rates in volts/nanosecond, are as harsh as you'll get.
Bottom line, if somebody spent a fortune on an amplifier -- beyond the point where the Law of Diminishing Returns sets in -- they must think it's good, otherwise they wouldn't have bought it. And there's unlikely to be any way of convincing them any different.
BTW, the first commercial use of transistor power amps was in juke boxes. My dad has a 1962 Seeburg with a 25+25 watt power amp (transformer coupled, has 100V line outputs, C/T to chassis so you can easily arrange mono speakers, taking 1/2 of LH signal plus 1/2 of RH signal in series) and also a power oscillator to run the motor at 45RPM (it does 33rpm on 50Hz so it needs 68Hz for 45RPM; it actually cheats by starting at 33RPM then switching to 45RPM, so it doesn't need to cope with the starting surge. A stationary motor looks like a short circuit). I don't think this was the first juke box to have a transistor amplifier, though, because I've seen one in a 1957 Wurlitzer (but this may have been a retrofit).
Je fume. Tu fumes. Nous fûmes!
Well, a long & drawn out old debate. First off; As that article is from the very early 70's, it doesn't really apply. Transistors from those days *did* suck & *did* sound bad. In one post, someone mentioned F.E.T's as a viable alternate to tubes. True, but expensive & very vulnerable to both static & the nast cascade effect should the input of the amp go pop. Second, tubes don't have crossover noise at all. Much more like a dimmer switch. Like FET's as well mind you. Sure they colourize the sound when distorted. Any amp would do this regardless of it being either tubes, transistors or FETS. All really irrelivant. Personal preference as the specs can be great depending on the way its tested. Who cares what type it is as long as the listener likes it. Thats all that matters.