I realized this is why the Altec- lansing amp sounded MUCH better with IT'S speakers than mine. They designed them as a complete package. Might also be why this system might be usable for 40 years or more.
OS
Hard to say with very limited information and 43 years later. Did you ask an experienced pro at the time to figure out what the differences were? Source impedance and resulting frequency response are possible. Poor level matching, also possible. Stability, ditto. Non-blind A-B tests are subject to conscious and unconscious bias. Many other possibilities, but no time machine here, so you can file that one under "cold case."
Guys I've just removed a couple of borderline posts, let's keep to attacking the topic "whatever that might be" and not each other..Here is a very good explanation and importantly some (effective) 4 bit
files undithered / dithered / dithered+noise shaped that you can listen to.
That Dither Thing - [English]
They say a picture is worth a thousand words in this case it should read
an mp3 is worth a thousand posts (on the BT thread)
cheers
T
I find this statement on that website to be totally inaccurate " Noise shaping cures the latter, as the purple plot testifies: in the critical mid-band the noise level is on the average 20dB below that of the dithered tone's, while it is much higher around 19kHz, where noise is much less audible, if at all."
I can hear the noise & a bad, unclear reproduction of the tone! So much for dither if this is meant as a good example of it - ruins the original tone & introduces noise which then has to be moved to an inaudible higher frequency - not very successful really, is it? I know it's only 4 bits but it degrades the original sound even at that bit level
Last edited:
Did you actually read my conditions, SY? Level, output impedance, stability, were all checked. Of course, I was a degreed working 'engineer' at the time, designing servos, low noise design, etc at the Ampex Audio Division. The test was set up by two experienced PRO's at the time. One a working engineer, and his associate who still works at Meyersound Labs to this day. Yes, MY conscious bias made the other guy's design sound better than mine.
Last edited:
Yes, I did. They are pretty vague, as one might expect after nearly half a century (e.g, how close were levels matched, what were the source impedances and how did they vary with frequency, what were the details of the test setup).
Do you still have your lab notebook from those experiments so that the details can be looked at?
Do you still have your lab notebook from those experiments so that the details can be looked at?
Bearing in mind it is a 4 bit signal, so for CD we are talking the equivalent
of encoding a musical signal around -72dB or so I think it is pretty good
(when noise shaped).
I can't say I really hear it the way you do. To me the noise shaped / dithered
signal is a lot better than undithered - within the constraints of the 4 bit
signal, which has quite obtrusive signal dependant distortion undithered.
To put it in perspective, if this were recorded to analog tape at same level
the noise would be at least 10 x higher and broadband. You probably
would be struggling to hear the signal at all.
T
For my '68 test the INPUTS were PARALLELED, so they shared the same input impedance. This would normally not be so good, because it does not separate the two components as completely as possible, but it did NOT matter.
To reduce any further confusion, the SLEW RATE of my solid state power amp was very low, due to the fact that I OVERCOMPENSATED it, in order to virtually eliminate any serious ringing when driven with a square wave and any capacitance.
My FINAL take was that: To get the same essential results, ie distortion, damping factor, etc, the tube unit used 20 dB of negative feedback. The solid state unit used 40dB. That is when I started to read Matti Otala's work more seriously.
To reduce any further confusion, the SLEW RATE of my solid state power amp was very low, due to the fact that I OVERCOMPENSATED it, in order to virtually eliminate any serious ringing when driven with a square wave and any capacitance.
My FINAL take was that: To get the same essential results, ie distortion, damping factor, etc, the tube unit used 20 dB of negative feedback. The solid state unit used 40dB. That is when I started to read Matti Otala's work more seriously.
Simple answer John. It was 1968 and SS amps were crap. Period. Things have moved on since then. Interestingly, Bob Cordell ran an A-B test at RMAF a while back between a tube amp and a SS and contrary to what we are all told, most of the audience (all audio nuts, otherwise they would not have been there) preferred the solid state sound. Big, deep, wide sound stage, extended top and bottom end, smooth and silky mid range, crisp, tight bass. All the things you can expect from a well designed SS amp circa 2011. There is no black magic in this.
Actually, my little solid state power amp was pretty good, measured well, was Class A for the first 4 Watts, (1/2A Iq on the output stage), the circuit was all complementary balanced, push pull throughout, and the final iteration concluded with my first complementary differential input stage, as a final effort to make it 'perfect'. I used it on MY K horn for years. Mark Levinson heard it in action in 1968, and 5 years later, based on HIS listening experience with it, went out of his way to hire me to do audio design in 1973. Yes, that is how good that little amp was, BUT it did NOT sound as good as the Radio Craftsman TRIODE power amp, designed around 1950, by the same guy who later designed the Marantz 9 power amp, a few years later.
However, this example should show anyone who cares to note, that we were careful with OUR listening tests, fully 10 years before Dr. Lipshitz started recommending close adjustment between components. We met every criterion THEN, yet while I should have been prejudiced to prefer MY design, I heard the triode amp sounding better.
Of course, my solid state design beat my DYNA MK3 power amp with matched KT88 output tubes, but that was ULTRA LINEAR, not TRIODE. The Dyna measured worse, too.
However, this example should show anyone who cares to note, that we were careful with OUR listening tests, fully 10 years before Dr. Lipshitz started recommending close adjustment between components. We met every criterion THEN, yet while I should have been prejudiced to prefer MY design, I heard the triode amp sounding better.
Of course, my solid state design beat my DYNA MK3 power amp with matched KT88 output tubes, but that was ULTRA LINEAR, not TRIODE. The Dyna measured worse, too.
There have been a few very special amps in my life experience.
The first was the Radio Craftsman Triode amp, the second was the Marantz 9 in triode mode, the third was the Mac 3500 power amp, the fourth was the original Otala solid state power amp, the fifth was the Audio Research D-150 power amp, especially with on headphones. Any one of these amps, in my opinion, show exceptional sonic quality, above any others that I ever heard, when these were still available, let's say 30 years ago. I call these amps, KEEPERS, because they will always be a reference to what an audio amp should sound like.
The first was the Radio Craftsman Triode amp, the second was the Marantz 9 in triode mode, the third was the Mac 3500 power amp, the fourth was the original Otala solid state power amp, the fifth was the Audio Research D-150 power amp, especially with on headphones. Any one of these amps, in my opinion, show exceptional sonic quality, above any others that I ever heard, when these were still available, let's say 30 years ago. I call these amps, KEEPERS, because they will always be a reference to what an audio amp should sound like.
Here is a very good explanation and importantly some (effective) 4 bit
files undithered / dithered / dithered+noise shaped that you can listen to.
That Dither Thing - [English]
They say a picture is worth a thousand words in this case it should read
an mp3 is worth a thousand posts (on the BT thread)
cheers
T
A picture is still worth a few posts. The problem has been dealt with in printing images for years. The difference with and without something like error diffusion dithering in a quantized image is pretty obvious.
http://www.cs.princeton.edu/courses/archive/fall00/cs426/lectures/dither/dither.pdf
Ah but there is the rub, John asked "what did we miss" (I assume as in objective measurements). It's all very well to say SS was crap back then, but I think the point was that the measurements done didn't show this
Tony.
mostly what people want to use to "discredit" audio measurements is the ad copy used at the time - not exactly the measurements engineers would use - 1 KHz THD at near full power was probably chosen deliberatley to "hide" early SS amp crossover errors to look "better" than tube/output xfmr with % 2nd, 3rd harmonic distoriton at the same levels - they were trying to move product
actually even THD sweep plots with both frequency and amplitude do show nasty cross over distortion, possible dyanamic bias error like shoot through or deadband
the "measurements didn't show differences" is largely a strawman - BBC noise fill test is a very complex "multitone" IMD test, slew rate, power bandwidth were standard amplifier tests - even if JC wants to claim they were only brought to "audio" engineers atttention in the 70s - the tests, limitations they measure in electronic cricuit were known to engineering in the large sense - Volterra series, complex IMD were used early in RF engineering, slew rate was a known Analog Computing tube "op amp" limit explained in 50's texbooks...
Last edited:
It came from slewing gun barrels in WWII with primitive electronics. http://www.antiaircraft.org/directors.htm#M9 AND M10 DIRECTORS
Last edited:
Slew rate was a mostly 'forgotten' term in the making of audio amps. This can be shown by attempting to find it in the reference books of the time. It re-emerged when IC OP AMPS became common, about 1965. IC's like the uA709, were so slow that they got in their own way in many applications. The uA741, was also slow, BUT convenient for servos, because transport motors could not slew very fast, anyway. I used them by the dozen in my servo designs.
It is amazing how 'history' gets rewritten by people who were not even there, in order to remove any hint of discovery on anyone else's part.
It is amazing how 'history' gets rewritten by people who were not even there, in order to remove any hint of discovery on anyone else's part.
Last edited:
You may be right, John, but absence of evidence is not the same thing as evidence of absence. People may not mention something because they don't think about it, or because it is so blindingly obvious that they assume everyone already knows. A later generation comes along and 'rediscovers' something which is only new to them.
- Status
- Not open for further replies.