I seem to remember that you can make an inexpensive drivers sound like a million dollars by painting spots on them?
Now, using the same " technology", would it be possible, by re-arranging the spots, make an SS amp sound like a tube amp or vise versa?
I don't know about a million dollar, but they do provide serious value for the dollar. Frugal-phile™.
No-one knows why it works, but it does (that verified with a number of informal blind & double-blind tests).
And in some cases, we can deal with some driver issues that are not covered by the spots.
dave
I was just saying than they can be equaly good but not with the same attributes. But at the end one has to choose his subjective preference, this last being a tauthology of course !
That's exactly what I said to my first wife at 20 ! I'm going to run after my second wife just know with a pencil ! 😀
ABX testing does not relate to personal preferences, so I don't know what you did, but it wasn't ABX. ABX is a test for differences.
I don't know about a million dollar, but they do provide serious value for the dollar. Frugal-phile™.
No-one knows why it works, but it does (that verified with a number of informal blind & double-blind tests).
And in some cases, we can deal with some driver issues that are not covered by the spots.
If you read To the Simple, Everything Appears Simple | Stereophile.com
You will find this:
"In the words of famed mastering engineer Bob Katz: "There is no such thing as a 'casual' blind test. Blind tests are a serious business. Experimenters need training how to perform blind tests well. Blind tests can fail (produce statistically invalid results) if the experimenter neglected one critical detail. Weeks of intensive study are required to learn how to perform blind tests. Then weeks of preparation to create the test. Then weeks of testing to follow."
Mr. Katz is correct.
Reminds me blind tests with wine where the testers even don't take 3 minutes between two glasses ! Of course at TV you only see perfs which are good and not all the faked blind tests or the ones with funny results (like a white which can have red attributes when you hide the eyes of the testers )😉
I admire people able to make devices good and not too much expensive ! I find they add an item which not help them and even the listeners (as it has been said cheap devices can sound worst just at the look for too much people... maybe not the majority of the people here).
I admire people able to make devices good and not too much expensive ! I find they add an item which not help them and even the listeners (as it has been said cheap devices can sound worst just at the look for too much people... maybe not the majority of the people here).
It seems to me that sighted evaluations are far more prone to have false outcomes than that.
I'm not sure in which context you are saying that but in hi-fi context, sounding good refers to the accuracy of output compared to the input.
Well worth repeating. I have been beating on this for a long time. I do know how much work it takes to do a proper blind test (Honours BSc in Statistical Mathematics).
dave
I like your post! and concur: I have a 70 watts rated hifi solid state amplifier which has very much more distortion than a KT88 tube pp mono-block feed with 800vac 250ma transformer.
Yes, well said, this is what says also a Thorten Loesch for instance when he wrotes : why I would make an output stages with transistors for my CD players when I can achieve a better quality more easily with tubes !
But I 'd like to listen to the first watt of the unique Jfets (only with) amp Papa made !
But I 'd like to listen to the first watt of the unique Jfets (only with) amp Papa made !
Last edited:
I am aware what ABX testing is, so were all of the test subjects. There WERE flaws in our testing, and these were known to us, but we did it anyway....because we could, and we wanted to.
There were only 6 test subjects. All were male. All had "HiFi" systems. Three of them were Motorola engineers who had completed the same "six sigma" training that I had since it was mandatory. Two others were from technical or scientific backgrounds. There should have been more, and more varied test subjects.
There was a delay between A and B and X of about 10 seconds. We shut the amp off to flip the OPT switches to avoid a large pop in the $8000 speakers! Each set of A,B and X ran 4 times in random order with different music each time. The transition should have been instantaneous.
Only one set of speakers were used and they were extremely efficient (106 db) Lowther based horns. The results might have been totally different on inefficient speakers with different characteristics. There was another set of speakers in the room, but we wouldn't have finished testing with another variable. We should have used a more "neutral" speaker.
Each listener has their own biases as to what they like, and listen for. These biases can be accentuated, or attenuated by the music choice, the speakers, the volume level, and the alertness of the test subjects. Two OPT's with different frequency response curves could be determined to be identical if there is little music in the frequency range in question. DCR can influence the dynamics, and speaker damping. Some test subjects may pick this up, some may not. Some may not like Lowther speakers either, hence they are single driver horns. These particular speakers have a built in subwoofer that receives their input AFTER the OPT, hence accentuating the bass resonant peak in impedance. We set the sub volume on minimum.
The only way to remove personal preferences from ANY testing is to choose a random set of test subjects with respect to the testing being done, and to have a large enough set of test subjects so that the few remaining biases wash out. All other variables and sources of bias should be removed. Hence taking a box full of unmarked phones to the shopping mall. The usual test was a comparison between two different vocoder software in the DSP, but there should be only ONE variable.
Yes, we tested five pairs of OPT's using the ABX method, and kept the test subjects in the dark as to which OPT was being heard. There were three OPT's that were similar in size (about 4 pounds each) specs and price (at the time). These were One Electron, Electra Print (seen in the Lexan amp in pic#3), and Transcendar (pic#1). There were two OPT's that were physically larger, the 9 pound Edcor, and the 11 pound Hammond. They were rated for higher power and cost more. The One Electron OPT's had a DCR that was about twice that of the other two small ones.
It was concluded that the test subjects could not RELIABLY prove that there was any difference between the three small OPT's. We chose this criteria. For 6 test subjects at least 4 of them should be able to pick X as either A or B 3 out of the 4 times the test was run. That did not happen although a trend did emerge. One listener picked the Electra Print as different EVERY time, and three listeners picked the One Electron as different most of the time.
The Edcor and Hammond could reliably be identified against any of the three small OPT's and against each other. 6 out of 6 and 4 out of 4 for the Hammond against anything.
We then broke for lunch and discussion. During this time I revealed the names of the OPT's, what to listen for and when on each music selection, and why. "Did you hear the bass guitar solo on......." How about the drum transients......The triangle..... Was Norah Jones in the room singing?
Everyone agreed that one OPT stood out as having big bass down low, limited highs and no "life" or "realism". That was the big Hammond. No one liked it. It was determined that there was another OPT that sounded good, but different from the others. It had big bass like the Hammond, didn't have the rolled off highs, and had realism, but WAS different. It was determined to be the Edcor. Based on our discussion, the other three were determined to be subjectively "better" than the Edcor or Hammond, but more testing was needed to "pick a winner".
After lunch we took 3 OPT's and 3 different musical selections, and predetermined notions of what to listen for back to the listening room. We zipped through another round of ABX testing, with a similar but stronger trend, but no reliable proof that the three OPT's were indeed different, although we knew that they were, so......
We then started single blind testing (I knew which OPT was tested, since I was hooking them up in the next room) of OPT's two at a time where each test subjects was asked to put a check mark on a score card as to which was "better" A, B or neither (both perceived to be the same). I chose some tests where the two OPT's WERE indeed the same to see if they were identified as such, and random combinations in other tests.
We did this for about an hour, then I revealed which sets of OPT's were used in each test, and we added up the "better" scores. Two people preferred the Electra Prints, which were the most expensive of the bunch. I prefer them as well. One person consistently did not like them. He said they blurred the reality...or something like that. The Transcendar was number two, and the One Electron number three, but it was determined that dulled transients and muddied bass was the reason. I had told the reviewers to listen for these clues.
I had initially brought the breadboarded SSE amp with the Hammonds installed because I liked the way they sounded on my system. I agree that they sounded gross on Jose's system. Why? My system used Yamaha NS10 studio monitors with an 87db sensitivity. I normally listen with the volume control between half way up and full throttle. Jose's 106 db speakers required 1/100 of the power for equal volume. That amp was turned almost all the way down. The less than linear magnetic losses in 11 pounds of iron was likely half the amp's output and worse at the high frequencies. As stated previously the results of this entire day of testing would likely have turned out totally different with a different set of speakers.
It not about hi-fi, but about subjectively that called "sound good". "Sound good" is about taste. If you do not know how the real music instruments sound, you can not judge a sound system is accurate or not, it just sound good or not.
Given how bad & compromised (in an absolute sense) even the best hifis are there are many equally valid ways for the output to resemble the input.
dave
Amp, preamp, DAC and even speakers can be evaluated on the fidelity level by comparing the measurements of input signal to output signal.
If you are talking about the measurements, that's only part of the evaluation. Listening is another part of evaluation, especially for speakers.
As for my question to you on post #49?
You can really only evaluate the sound of a hifi as a system, including the transducer at the end.
For instance i have seen distortion measures where the low measured amp + speakers had more than a much higher distortion amp + the same speakers. An instance showing that a system has to be evaluated as a whole.
dave
- Status
- Not open for further replies.