are you after definitive studies on audibility or on the benefits of training. Actually daft question as there is no 'definitive' study. I should note I broadly agree with the work of Earl Geddes in coming up with better metrics. His free papers are well worth a read.
GedLee LLC
Of course because it upsets the status quo of THD and IMD measurements it's not gained popularity.
GedLee LLC
Of course because it upsets the status quo of THD and IMD measurements it's not gained popularity.
Here are some scholarly articles on the perceptibility of distortions (linear and nonlinear) in audio. It has been studied. I tried to avoid names we already know in high end audio. That would be easy and not really helpful. It is a real subject and in some cases has critical life or death importance (mostly at levels far higher than we are concerned with) so real research has been done. There is lots more but this is just to illustrate the point.
AES E-Library Speaker Auralization ' Subjective Evaluation of Nonlinear Distortion
AES E-Library Loudspeaker Measurements and Their Relationship to Listener Preferences: Part 1
AES E-Library The Audibility of Frequency Response Irregularities
http://asa.scitation.org/doi/abs/10.1121/1.1354986
AES E-Library The Effect of Nonlinear Distortion on the Perceived Quality of Music and Speech Signals
AES E-Library Speaker Auralization ' Subjective Evaluation of Nonlinear Distortion
AES E-Library Loudspeaker Measurements and Their Relationship to Listener Preferences: Part 1
AES E-Library The Audibility of Frequency Response Irregularities
http://asa.scitation.org/doi/abs/10.1121/1.1354986
AES E-Library The Effect of Nonlinear Distortion on the Perceived Quality of Music and Speech Signals
You seem to be lumping everything into distortion, it would take a serious study to eliminate all other possibilities. When we tackle that then maybe we can discuss the signal chains with double digit distortion that are highly regarded.
EDIT missed that you already know about this, so just what is your point?
Last edited:
I tried 3 different op-amps in a cmoy circuit recently, i listened over a few nights to each one, i noticed there was small differences in presentation. The one i settled on was the 4556, it best suited my Grado cans. All 3 had different spec so i was not that surprised, maybe the 4556 suited the feedback values or layout better. I don't really know for sure but maybe there is more than just cost involved when he choose that critter for the Grado headphone amp. One thing i'd add is that i would be hard pressed to hear much if any difference through loudspeakers. If you want to evaluate op-amps i would use a decent set of headphones if possible as well as speakers and plenty of time to form any opinions. Any differences are likely to be very small (unless something is badly wrong)
I very much like some signal chains with double digit distortion, including my modified Fender Blues Jr. guitar amp, and sometimes with a DS-1 or other pedal. Yet, I don't like the edgy harmonic the Lynx adds to acoustic guitar or vocals, although it isn't terrible and makes a good backup for when I want to record more tracks at once. However, the HEDD doesn't do that, it sounds cleaner and smoother, more musical, don't know how it measures in comparison to the Lynx. I found out after I got mine that Bob Katz rated HEDD as an A+ quailty A/D and something like a B+ quality D/A. I would say he got that right. I would use the Lynx D/A before the HEDD. Part of the perceptual experience of what sounds bad about some of the D/A's is I guess what people call "smeared," whatever that means, and whatever causes that perceptual experience. But if heard, "smear" seems to fit pretty well. The aural experience bears some similarity to the the visual experience of moving picture motion blur, although there is a little more to it than just that.
Moving along, looks like some people refer to linear and nonlinear distortion, with frequency distortion presumably falling into the linear category. I assume things like THD, IMD, TID, etc., are classified under nonlinear. So, regarding my "lumping," I think the things that I don't like mostly would fall under nonlinear. But, as I was trying to say, some nonlinear can be quite aesthetically pleasing.
I wonder what other "possibilities," though, you have in mind. I'm all ears...
Last edited:
For the transformers, I was just responding to Bill's aside on Neve stuff. The aside and my response were a bit off topic from the rest.
I was at a CES, back when it was at the, ummmm, Sahara??
Keith Johnson was there with some modified Ampex (iirc) with 1/2 tape...
...of course it is possible that my hindsight is "flawed".
Many of the CDs appear to be LF limited, not just shelved or rolled, but brickwalled. Some years back I did some FFTs of various CDs because there was a debate, put the graphs up, probably on this forum. Dafos does have LF information. I got into it because when I got certain reissues I "was like where's the bottom end?"
The infamous Telarc 1812 Overture was the one that jumped the stylus out of the groove, I think on WQXR, but maybe another classical station and I think dumped the transmitter too...
When the big thump comes on Dafos here, the back wall moves out... just a pressure wave is what the sensation is... was it the big hanging kit being dropped or something else??
No opamps were used or overdriven in this post <--- On Topic portion of post. 😛
Thanks Scott!!
Many other CDs I did the same thing with look like a sharp drop below ~35Hz at best.
"We" have that already.
Ed is making a "box".
Apparently he will let me listen to it and report what I find, good bad or indifferent.
Perhaps others will have a go at it.
Ed will report to us (hopefully in a separate thread) the results.
"We" will scream, shout, run in circles, throw mud, ponder, elucidate, and do all manner of things relating to what happened.
This will be "phase 1".
The next step will be subsequently determined.
Perhaps it may garner active participation by a group, wherein a more ambitious "box" or "system for testing" can be assembled, perhaps embodying a DBT/ABX sort of method. Or perhaps another series of preliminary tests needs to be done first.
I can think of a number of aspects that ought to be changed, those being let's say "variables" WRT the circuit design and implementation, that would be prudent in order to rule out or rule in said elements.
(some of these were mentioned earlier in the thread...)
So, consider thinking of this as an iterative process, if continued in its logical direction and vector, that may lead to something real and of use to all - regardless of their current views or position. Myself included.
_-_-
BTW, and fwiw, I hope to pull in some unwitting dupes to have a listen as well - potentially I can report what they find, even though you may think that I have "influenced them" - probably by the sheer weight of my imposing personality and the emission of actinic radiation. No doubt the results will be that they hear essentially what I hear, they hear some of what I hear, or they hear nothing like what I hear. 😀
What may we find ultimately?
Perhaps, a conjecture:
It is possible to make a 4558 sound as good as a 5534, and as good as a 797.
BUT, it requires $50 worth of precision shunt regulation on each rail.
I don't know, but perhaps it is this sort of thing that may emerge ultimately.
One can't find this in a single test. IF it can be found at all.
_-_-
Perhaps, a conjecture:
It is possible to make a 4558 sound as good as a 5534, and as good as a 797.
BUT, it requires $50 worth of precision shunt regulation on each rail.
I don't know, but perhaps it is this sort of thing that may emerge ultimately.
One can't find this in a single test. IF it can be found at all.
_-_-
Beatles? Seriously that is the best you can come up with Robert?
Mark: sorry missed your edit whilst writing my response. As Scott says, would be interesting to find out what your Lynx is doing that you don't like. I doubt its as simple as THD. Did Bob do any detailed measurements?
Mark: sorry missed your edit whilst writing my response. As Scott says, would be interesting to find out what your Lynx is doing that you don't like. I doubt its as simple as THD. Did Bob do any detailed measurements?
But then I'd have to listen to the Beatles, twice. I sat down and tried to watch "Hard Day's Night" with my older daughter and she asked if I actually watched this when I was a teenager. Even though I love Richard Lester I have to admit the music no longer does anything for me. And I realized that Lennon had a personality that grated me even back then.
EDIT Channeling Bill again. 😀
Last edited:
I may be confused here, but is the HEDD the Crane song unit you were talking about before, the one that has a USP of adding different distortion profiles? If so would be interesting to get some proper measurements of that compared to the Lynx as a cynic would say that the HEDD will be less 'clean' but may well produce a nice sound, as its clearly a musical instrument type device rather than a 'designed to be accurate' device.
Also note if you are recording music rather than measuring you are in 'nice' territory and accuracy be damned, especially if you want repeat business 🙂
Also note if you are recording music rather than measuring you are in 'nice' territory and accuracy be damned, especially if you want repeat business 🙂
As I read through some of the work of Geddes and Howard, I see what they have been doing, which is all good, but I still haven't seen any convincing evidence of ear training equivalent in effectiveness to cymbal based training, or any evidence of prolonged and repetitive distortion recognition training sessions.
How come it takes lots of practice to play an instrument, or to transcribe recorded music, but anybody can learn to expertly hear distortion in an hour listening to some symphony or something with a little added distortion of limited variety? I hope nobody thinks it's possible to become a professional mastering engineer by reading some books on what the equipment does, various format standards documents, and then practicing listening for a couple hours.
Again, I would like to know if anybody searched through more than a few tens or a hundred people, and rather went out a found a few one-in-a-thousand natural talents, then trained them and gave them ample practice, then tested them on equipment known to accurately reproduce low level distortion signals?
If you take the best listeners and play them music through a HEDD D/A converter, the D/A might well mask much of the low level distortion the listeners could otherwise potentially detect.
For the most sensitive measurements of the most sensitive listeners, I would probably want to use Benchmark converters, or the equivalent if there is any, for the testing.
And people keep insisting all this has all been done flawlessly 40 years ago?
I think some of you guys are skeptical when people claim to hear stuff you don't, but not skeptical enough that previous testing was done years ago as well as it could be done today. And I don't think you would say the same about other testing.
Probably everybody would like to have the latest Audio Precision test gear, presumably because it can show stuff not so easily seen before. And then hopefully you can make better amplifiers.
But, no matter the test equipment, you can't make your hearing better, so who even wants to go there? And in the worst case, it could even turn out that some of the people looked down upon and derided so much weren't actually all that crazy. Boy, wouldn't that be bad news?
Not that I know of. I imagine he may be been only judging by ear.
Regarding the HEDD itself, it has an internally or externally clocked A/D, a digital distortion function that can be applied to the A/D, D/A, or SPDIF/AES-EBU stream, or turned off entirely, and an externally clocked D/A presumably with cascaded PLLs that can track any incoming frequency up to 192kHz. Because the D/A is clocked separately from the A/D, the box could be used for sample rate conversion by going into the D/A at one clock rate, then analog back into the A/D at some other clock rate. In the old days before good software SRC people sometimes used it for that. It's distortion algorithm has largely been supplanted by the Crane Song Phoenix digital plugin, although that only runs on Macs and with Pro Tools. The primary use for the HEDD box at this time is the A/D converter which still sounds exceptionally good. The other stuff is there if one wants to use it, and occasionally it does get used, but not too often.
Last edited:
tests on thresholds of hearing are continually being done. The Geddes work is this century. Oddly few flaws have appeared in the old work. The end result is the same. Amplifiers can be made transparent if you want them to be. Or they can be made very audible.
As a point of order I think this whole discussion is going off the rails with recording engineer's effects boxes thrown right in with the op-amp discussion. I made my position clear in my LA article, I deeply respect the art of a recording engineer/producer, but my approach in a design article is instrumentation to preserve the signal without amplitude or temporal changes. Echoed mainly in the B&K technical notes, a microphone's job is to translate sound pressure vs time to a voltage vs time with as much analytically measurable performance as possible. A general purpose amplifier's job is to predictably preserve signal fidelity, if the IC designer does his/her job the customer needs to do less work/ needs less esoteric knowledge to achieve that.
Recording mics or amplifier products for that matter with known built in sonic effects, presence peaks, etc. are fine but then it is all a subjective issue. Why bother blind testing things that are all audible to anyone but strike different subjective cords with each individual.
Recording mics or amplifier products for that matter with known built in sonic effects, presence peaks, etc. are fine but then it is all a subjective issue. Why bother blind testing things that are all audible to anyone but strike different subjective cords with each individual.
- Status
- Not open for further replies.