IPB

Welcome Guest ( Log In | Register )

2 Pages V   1 2 >  
Reply to this topicStart new topic
How do you reliably ABX equipment that needs to be disconnected, and how can TOS #8 be satisfied in such situations?
BearcatSandor
post Nov 2 2010, 03:23
Post #1





Group: Members
Posts: 149
Joined: 18-May 10
From: Montana, USA
Member No.: 80732



(I was not sure where to put this. My emphasis is really on the testing methodology, not TOS #8 so i decided this would be the best area to put it in)

I'm still new here and trying to learn about proper ABXing and how to avoid the threat of a dreaded TOS 8 hammer (unsuccessfully it seems). I was asked about whether i had done an ABX test between 2 amplifiers. That got me thinking about how one would be able to do that in the first place in their own home.

Suppose i am ABXing a similar amplifier against the one i already have. My system is: Sound card ---> interconnects ---> amplifier (2 channel/1 input) ---> speaker cable ----> speakers
New amp would be the same: Soundcard ---> interconnects ---> amplifier (2 channel/1 input) ---> speaker cable ----> speakers
(a third one is needed? How does that work with hardware?)

To switch between the amps my assistant would need to move the interconnects back and forth between amps and the sound card outputs. That involves making sure that the amp 1 is shut down, removing the cables from amp 1 and plugging them into amp 2 and turning on amp 2. There would be significant lag in doing so, wouldn't there? The same problem would be there for cable ABXing as well.

Doesn't lag play a part in the accuracy of these tests?

How do you folks satisfy an ABX request in such a situation? Do you commonly have some sort of switching device in your home? How would i provide a sample for the community as outlined in TOS #8 (making it unsatisfiable)? Provide duplicates of the system used?

I looked around the site for information on how to ABX components but what i found seemed to relate to sound files and not hardware.

Could you folks help me understand this?

Thanks


--------------------
Music lover and recovering high end audiophile
Go to the top of the page
+Quote Post
kiit
post Nov 2 2010, 04:12
Post #2





Group: Members
Posts: 125
Joined: 9-October 03
From: Washington D.C.
Member No.: 9229



TOS 8 is usually mentioned when someone has stated their opinion about something when all known facts point to the exact opposite conclusion..

Yes ABX for amps is pretty much impossible in the home by yourself. But it is not impossible to do, and has been done, and the results always seem that the cheap off the shelf amp being is as good or more preferred than the expensive flashy amp. As for tubes over chips, that comes down to personal preference and do you like harmonic distortion or not.

If you want to post about how xxx amp is better than yyy amp, I think you are on the wrong message boards and would find a happier time of it at a place that doesn't insist on scientific and objective testing. In the 8 years I've been coming here not once have I seen a happy OP on those topics.

<edit> P.S. you don't get booted for TOS #8, not even warned, so don't worry about it.. you have to be pretty mean and/or obstinate just to get a warning. Though if you ignored that and people complained, then you could worry. I imagine anyone trolling that hard wouldn't mind much as they probably didn't want to be Here in the first place. (And I think doing ABX on an amp would almost require someone else to help you with it using a switch box.)

This post has been edited by kiit: Nov 2 2010, 04:18
Go to the top of the page
+Quote Post
saratoga
post Nov 2 2010, 04:22
Post #3





Group: Members
Posts: 4962
Joined: 2-September 02
Member No.: 3264



Get a really good ADC and record both outputs, then ABX the WAVs.
Go to the top of the page
+Quote Post
analog scott
post Nov 2 2010, 04:54
Post #4





Group: Members
Posts: 332
Joined: 26-July 09
Member No.: 71796



QUOTE (BearcatSandor @ Nov 2 2010, 04:23) *
(I was not sure where to put this. My emphasis is really on the testing methodology, not TOS #8 so i decided this would be the best area to put it in)

I'm still new here and trying to learn about proper ABXing and how to avoid the threat of a dreaded TOS 8 hammer (unsuccessfully it seems). I was asked about whether i had done an ABX test between 2 amplifiers. That got me thinking about how one would be able to do that in the first place in their own home.

Suppose i am ABXing a similar amplifier against the one i already have. My system is: Sound card ---> interconnects ---> amplifier (2 channel/1 input) ---> speaker cable ----> speakers
New amp would be the same: Soundcard ---> interconnects ---> amplifier (2 channel/1 input) ---> speaker cable ----> speakers
(a third one is needed? How does that work with hardware?)

To switch between the amps my assistant would need to move the interconnects back and forth between amps and the sound card outputs. That involves making sure that the amp 1 is shut down, removing the cables from amp 1 and plugging them into amp 2 and turning on amp 2. There would be significant lag in doing so, wouldn't there? The same problem would be there for cable ABXing as well.

Doesn't lag play a part in the accuracy of these tests?

How do you folks satisfy an ABX request in such a situation? Do you commonly have some sort of switching device in your home? How would i provide a sample for the community as outlined in TOS #8 (making it unsatisfiable)? Provide duplicates of the system used?

I looked around the site for information on how to ABX components but what i found seemed to relate to sound files and not hardware.

Could you folks help me understand this?

Thanks

You can do it with an ABX box. Not sure where you would find one but someone else may know. That was pretty much built for comparing things like amps.
Go to the top of the page
+Quote Post
BearcatSandor
post Nov 2 2010, 05:15
Post #5





Group: Members
Posts: 149
Joined: 18-May 10
From: Montana, USA
Member No.: 80732



QUOTE (saratoga @ Nov 1 2010, 21:22) *
Get a really good ADC and record both outputs, then ABX the WAVs.


Ohh. Thank you. That's a very good idea.


--------------------
Music lover and recovering high end audiophile
Go to the top of the page
+Quote Post
BearcatSandor
post Nov 2 2010, 06:27
Post #6





Group: Members
Posts: 149
Joined: 18-May 10
From: Montana, USA
Member No.: 80732



QUOTE (analog scott @ Nov 1 2010, 21:54) *
You can do it with an ABX box. Not sure where you would find one but someone else may know. That was pretty much built for comparing things like amps.

Is that any different from a 3-way switch box (i donno why didn't think of that)


(i apologize for the double posting. I tried to edit these into one post but i too much time has passed to do so. Blame the D20 game my mate was running)


--------------------
Music lover and recovering high end audiophile
Go to the top of the page
+Quote Post
analog scott
post Nov 2 2010, 06:33
Post #7





Group: Members
Posts: 332
Joined: 26-July 09
Member No.: 71796



QUOTE (BearcatSandor @ Nov 2 2010, 07:27) *
QUOTE (analog scott @ Nov 1 2010, 21:54) *
You can do it with an ABX box. Not sure where you would find one but someone else may know. That was pretty much built for comparing things like amps.

Is that any different from a 3-way switch box (i donno why didn't think of that)


(i apologize for the double posting. I tried to edit these into one post but i too much time has passed to do so. Blame the D20 game my mate was running)


Yes it is quite different. It allows for double blindness.
Go to the top of the page
+Quote Post
krabapple
post Nov 2 2010, 07:26
Post #8





Group: Members
Posts: 2271
Joined: 18-December 03
Member No.: 10538



QUOTE (analog scott @ Nov 2 2010, 01:33) *
QUOTE (BearcatSandor @ Nov 2 2010, 07:27) *
QUOTE (analog scott @ Nov 1 2010, 21:54) *
You can do it with an ABX box. Not sure where you would find one but someone else may know. That was pretty much built for comparing things like amps.

Is that any different from a 3-way switch box (i donno why didn't think of that)


(i apologize for the double posting. I tried to edit these into one post but i too much time has passed to do so. Blame the D20 game my mate was running)


Yes it is quite different. It allows for double blindness.


The QSC ABX Comparator did even more than that...it also allowed precise level matching .

http://www.ackthud.net/shawnfogg/pics/temp/ABX_manual.pdf


Unfortunately the devices have been discontinued for some years now.
Go to the top of the page
+Quote Post
BearcatSandor
post Nov 2 2010, 07:44
Post #9





Group: Members
Posts: 149
Joined: 18-May 10
From: Montana, USA
Member No.: 80732



QUOTE (saratoga @ Nov 1 2010, 21:22) *
Get a really good ADC and record both outputs, then ABX the WAVs.

I've been thinking about this. Unless i am misunderstanding something that i've always taken for granted, that is that the inputs of an amp are electrically different than the speaker outputs, how could you do this with an amp? I assumed you were only speaking of source components.

I'm not an electrical engineer.


--------------------
Music lover and recovering high end audiophile
Go to the top of the page
+Quote Post
greynol
post Nov 2 2010, 09:24
Post #10





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (kiit @ Nov 1 2010, 20:12) *
you don't get booted for TOS #8, not even warned, so don't worry about it.

...and just how would you know this? It certainly isn't true!

Before anyone decides to come forth to corroborate this post of mine, be aware that doing so is considered a violation of our TOS, can lead to a warning and, depending on your history on this forum, might very well result in banishment.

This post has been edited by greynol: Nov 2 2010, 09:48


--------------------
I should publish a list of forum idiots.
Go to the top of the page
+Quote Post
greynol
post Nov 2 2010, 09:54
Post #11





Group: Super Moderator
Posts: 10000
Joined: 1-April 04
From: San Francisco
Member No.: 13167



QUOTE (saratoga @ Nov 1 2010, 20:22) *
Get a really good ADC and record both outputs, then ABX the WAVs.

Do you need to make sure to take into account the loading of the amplifier like what is often needed when working with headphones? If so then will the signal feeding the ADC be adequate without pre-amplification?


--------------------
I should publish a list of forum idiots.
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Nov 2 2010, 14:21
Post #12





Group: Members
Posts: 3797
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (BearcatSandor @ Nov 1 2010, 22:23) *
(I was not sure where to put this. My emphasis is really on the testing methodology, not TOS #8 so i decided this would be the best area to put it in)

I'm still new here and trying to learn about proper ABXing and how to avoid the threat of a dreaded TOS 8 hammer (unsuccessfully it seems). I was asked about whether i had done an ABX test between 2 amplifiers. That got me thinking about how one would be able to do that in the first place in their own home.

Suppose i am ABXing a similar amplifier against the one i already have. My system is: Sound card ---> interconnects ---> amplifier (2 channel/1 input) ---> speaker cable ----> speakers
New amp would be the same: Soundcard ---> interconnects ---> amplifier (2 channel/1 input) ---> speaker cable ----> speakers
(a third one is needed? How does that work with hardware?)

To switch between the amps my assistant would need to move the interconnects back and forth between amps and the sound card outputs. That involves making sure that the amp 1 is shut down, removing the cables from amp 1 and plugging them into amp 2 and turning on amp 2. There would be significant lag in doing so, wouldn't there? The same problem would be there for cable ABXing as well.

Doesn't lag play a part in the accuracy of these tests?


All good points.

ABX evolved in a time (1970s) and place (a bunch of experienced audio techs) where people had a lot of technical resources that are probably a lot less common today.

For example, the first pre-ABX amplifier tests were done using a remote-controlled amplifier/speaker switch box that was operated by someone who had a clear view of everything. This solved the cable-swapping problem. The first amplifiers that were compared were power amps that had input level controls, so level-matching was a slam dunk.

The first refinement was developing a blinded randomizing controller that operated the remote controlled amplifier/speaker switch box. Once we got the basic ABX procedure working, we incorporated speaker-level switching into the randomizer. When we moved on to comparing line-level items like LP playback, we devised matched preamplifier's that had switching relays built-in. They were slaved to the speaker-level switching in the ABX Comparator. When we moved to CD players, we built a dedicated line-level swticher that was also slaved.

When we started comparing equipment that lacked appropriate built-in volume controls, we developed external volume controls with fine adjustments.

In most cases the "we" was "I". ;-)


QUOTE
How do you folks satisfy an ABX request in such a situation? Do you commonly have some sort of switching device in your home? How would i provide a sample for the community as outlined in TOS #8 (making it unsatisfiable)? Provide duplicates of the system used?

I looked around the site for information on how to ABX components but what i found seemed to relate to sound files and not hardware.

Could you folks help me understand this?


Eventually a purpose-engineered system for hardware ABX was developed. a partnership of 6 people (still friends to this day) was formed to develop, produce, and sell it. You can find out more about it here:

ABX Comparator Hardware Web Site

This hardware was sold to a number of equipment and R&D shops over a period of maybe 5 years back in the 1980s. More than 50 systems were sold. There are no central records of who currently owns them or how operational they are.

In the mid-1990s ABX and similar or improved blind listening tests became very interesting to people who developed perceptual coders. This kind of testing is well-served by file comparison software.

In the year 1000 a methodology I called PC ABX was developed and popularized via a web site called www.pcabx.com. It was based on playing files on a computer using purpose-written software.

The files being compared can be produced by recording the outputs of audio equipment. ADC and DAC technology has progressed to the point where the best ADCs and DACs are comparable or superior to most audio equipment,. Using them to record and play the output of audio gear does not significantly degrade sound quality.


Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Nov 2 2010, 14:29
Post #13





Group: Members
Posts: 3797
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (greynol @ Nov 2 2010, 04:54) *
QUOTE (saratoga @ Nov 1 2010, 20:22) *
Get a really good ADC and record both outputs, then ABX the WAVs.

Do you need to make sure to take into account the loading of the amplifier like what is often needed when working with headphones?


Yes. I studied a ton of loudspeaker test reports and developed a "reasonable worst case" simulated speaker load.

QUOTE
If so then will the signal feeding the ADC be adequate without pre-amplification?


Depends on the power level you want to test the amplifier at. In general, ADCs are sensitive enough that one problem is reducing the output of the power amp enough so that it does not clip or damage the ADC. I have a resistive and a complex dummy load, each with built-in attenuation controls.

I have damaged ADCs and audio interfaces while doing amplifier tests, so due care is in order.

I've done tests where additional amplification was needed. I've been pretty successful using high quality mic preamps for this purpose. Good mic preamps can be very clean when operated at the relatively low gains that are usually required. Obviously, the first test is to loop the mic preamp and do technical and subjective tests and adjust as needed to produce adequate results showing sonic transparency. Ditto for the audio interface or AD/DACs that you use.
Go to the top of the page
+Quote Post
pbelkner
post Nov 2 2010, 14:30
Post #14





Group: Members
Posts: 412
Joined: 13-June 10
Member No.: 81467



QUOTE (Arnold B. Krueger @ Nov 2 2010, 15:21) *
Using them to record and play the output of audio gear does not significantly degrade sound quality.

How do you ABX this one?

This post has been edited by pbelkner: Nov 2 2010, 14:32
Go to the top of the page
+Quote Post
Arnold B. Kruege...
post Nov 2 2010, 16:17
Post #15





Group: Members
Posts: 3797
Joined: 29-October 08
From: USA, 48236
Member No.: 61311



QUOTE (pbelkner @ Nov 2 2010, 09:30) *
QUOTE (Arnold B. Krueger @ Nov 2 2010, 15:21) *
Using them to record and play the output of audio gear does not significantly degrade sound quality.

How do you ABX this one?


Hook the output of the ADC to a DAC and compare the output of the DAC to the input of the ADC using hardware ABX.

With really good equipment, pairs of converters can be cascaded 20 times or more and still escape detection.

I think that Ethan Winer recently tried this with a $25 SoundBlaster card and even this was undetectable, when cascaded 5 or 10 times.

Modern converters are that good!

Back around Y2K the new SoundBlaster Live! could be detected in just one pass with a tiny bit of work. Before that, detection of SoundBlaster converters just one pass was very easy.

Converters in professional gear and good mid-fi CD players has always been much better than that, even back then.

The converters in some of the first generation CD players were just barely detectible, but things improved rapidly in the next generation. The best second generation players had sonically transparent converters.

The converters in any regular iPod or even a Sansa Clip are sonically transparent.
Go to the top of the page
+Quote Post
saratoga
post Nov 2 2010, 16:39
Post #16





Group: Members
Posts: 4962
Joined: 2-September 02
Member No.: 3264



QUOTE (BearcatSandor @ Nov 2 2010, 02:44) *
QUOTE (saratoga @ Nov 1 2010, 21:22) *
Get a really good ADC and record both outputs, then ABX the WAVs.

I've been thinking about this. Unless i am misunderstanding something that i've always taken for granted, that is that the inputs of an amp are electrically different than the speaker outputs, how could you do this with an amp? I assumed you were only speaking of source components.

I'm not an electrical engineer.


I assumed you wanted to test the amp and not the speakers, so you could hook the amp up to the speakers and then record the output of the amp driving the speakers using an ADC. Since the impedance of the speakers is low (few ohms) and the ADC very high (many kohms), the ADC records the signal essentially without changing it. If you actually want to test the speakers, then theres no good way to do this.
Go to the top of the page
+Quote Post
BearcatSandor
post Nov 3 2010, 01:46
Post #17





Group: Members
Posts: 149
Joined: 18-May 10
From: Montana, USA
Member No.: 80732



QUOTE (saratoga @ Nov 2 2010, 09:39) *
QUOTE (BearcatSandor @ Nov 2 2010, 02:44) *
QUOTE (saratoga @ Nov 1 2010, 21:22) *
Get a really good ADC and record both outputs, then ABX the WAVs.

I've been thinking about this. Unless i am misunderstanding something that i've always taken for granted, that is that the inputs of an amp are electrically different than the speaker outputs, how could you do this with an amp? I assumed you were only speaking of source components.

I'm not an electrical engineer.


I assumed you wanted to test the amp and not the speakers, so you could hook the amp up to the speakers and then record the output of the amp driving the speakers using an ADC. Since the impedance of the speakers is low (few ohms) and the ADC very high (many kohms), the ADC records the signal essentially without changing it. If you actually want to test the speakers, then theres no good way to do this.

Please excuse my ignorance. I'm going to be asking for clarifications to new ideas and i'm not near the level of you folks. In fact i'm a recovering audiophile i guess. A lot of the things that i *think* i know are going to be proven false over the next few weeks i bet. Feel free to point me to links rather than explain.

Nope, i was speaking of testing the amplifiers themselves. You and Arnold clearly explained a way to do it. The idea of hooking up an amplifier to an ADC is really something i never would have thought of..

This is fascinating.

Thank you

EDIT: typo and my spelling checker suggested that i had meant to say "..hooking up an amplifier to AC/DC is something i never would have thought of" Frankly, i don't think i'd much like the sound of that band doing an Unplugged album. :")

This post has been edited by BearcatSandor: Nov 3 2010, 01:52


--------------------
Music lover and recovering high end audiophile
Go to the top of the page
+Quote Post
LocrianGroove
post Dec 29 2010, 05:57
Post #18





Group: Members
Posts: 107
Joined: 28-February 07
Member No.: 41051



I'd like to propose an ABX test method, so that when I present results, my results will be taken seriously. My goal is to ABX a laptop soundcard and an iPod with IEM's.

1) I would have someone assist in switching connections while I'm out of the listening room, using a randomized sequence, 8 X's long. That person would leave the listening room between listening passes. I would get to listen to A and B as many times as I need to, just as in a software ABX test.

2) I would let each device play back a looped segment for the duration of the test, so we can press play on each of the players just once. The players won't be looping, but rather the files played back will consist of loops.

3) I would achieve level matching by choosing a reasonable listening level, then recording the output of each device and adjusting the outputs until I could make a recording of each device playing the same segment with the same RG value to within 0.5 dB.

Would the results of such a test be taken seriously? If not, what else might I try?
Go to the top of the page
+Quote Post
saratoga
post Dec 29 2010, 06:21
Post #19





Group: Members
Posts: 4962
Joined: 2-September 02
Member No.: 3264



QUOTE (LocrianGroove @ Dec 28 2010, 23:57) *
Would the results of such a test be taken seriously? If not, what else might I try?


Thats probably fine from a blindness perspective, however unless the differences are quite large I think the long time it'll take to switch samples will make it impossible for you to detect a difference. So your test will be blind, but also probably give you a negative result smile.gif

Assuming you have access to a reasonably good sound card, I would use the method I suggested above so that you can quickly change samples in a double blind manner. Given that your source material will be 16 bit, transparently recording it should be quite easy.
Go to the top of the page
+Quote Post
knutinh
post Dec 29 2010, 14:45
Post #20





Group: Members
Posts: 569
Joined: 1-November 06
Member No.: 37047



QUOTE (saratoga @ Dec 29 2010, 06:21) *
Thats probably fine from a blindness perspective, however unless the differences are quite large I think the long time it'll take to switch samples will make it impossible for you to detect a difference. So your test will be blind, but also probably give you a negative result smile.gif


If one cannot remember the audible impression from sound gadget #1 over a time-span of 5 minutes well enough to reliably compare it to sound gadget #2, what does that say about the benefit I would get in my home stereo if I bought one instead of the other? After all, if I purchase a new amplifier, I do not expect to rapidly switch back and forth between it and the older one for the products lifetime.

-k
Go to the top of the page
+Quote Post
MerlinWerks
post Dec 29 2010, 14:56
Post #21





Group: Members
Posts: 11
Joined: 6-April 02
Member No.: 1709



QUOTE (BearcatSandor @ Nov 1 2010, 22:23) *
I looked around the site for information on how to ABX components but what i found seemed to relate to sound files and not hardware.

Could you folks help me understand this?

Thanks


Perhaps you can find some inspiration HERE and HERE.

I've built the A/B switch box and although it certainly is not an ABX test, after a few random pushes on the switch it's virtually impossible to know which output is active. I used DPDT relays instead of the SPDT relays to avoid having a common ground between amp channels which could be destructive if you wanted to test various class D or "T" amps.
Go to the top of the page
+Quote Post
DonP
post Dec 29 2010, 14:58
Post #22





Group: Members (Donating)
Posts: 1471
Joined: 11-February 03
From: Vermont
Member No.: 4955



QUOTE (LocrianGroove @ Dec 28 2010, 23:57) *
3) I would achieve level matching by choosing a reasonable listening level, then recording the output of each device and adjusting the outputs until I could make a recording of each device playing the same segment with the same RG value to within 0.5 dB.

Would the results of such a test be taken seriously? If not, what else might I try?


I think half a dB is in the prime range where the louder (otherwise identical) music sounds better without being perceived as louder.

0.1 dB I think is the usual target.
Go to the top of the page
+Quote Post
DonP
post Dec 29 2010, 15:12
Post #23





Group: Members (Donating)
Posts: 1471
Joined: 11-February 03
From: Vermont
Member No.: 4955



QUOTE (BearcatSandor @ Nov 1 2010, 21:23) *
To switch between the amps my assistant would need to move the interconnects back and forth between amps and the sound card outputs. That involves making sure that the amp 1 is shut down, removing the cables from amp 1 and plugging them into amp 2 and turning on amp 2. There would be significant lag in doing so, wouldn't there? The same problem would be there for cable ABXing as well.

Doesn't lag play a part in the accuracy of these tests?



Even without the suggested switching devices, consider that a call for blind testing has often been in response to "day vs night" claims sometimes observed between system components in different buildings on different days. IF the amps in question have such glaring differences, they should still be very obvious when tested in the same room a few minutes apart, and no need for repeated switching in the same trial. That makes things a lot easier.
Go to the top of the page
+Quote Post
LocrianGroove
post Dec 30 2010, 01:11
Post #24





Group: Members
Posts: 107
Joined: 28-February 07
Member No.: 41051



I think 0.1 dB accuracy would be hard to attain by adjusting output levels, but I could compensate by changing RG tags. Thanks for the feedback.
Go to the top of the page
+Quote Post
LocrianGroove
post Dec 30 2010, 05:51
Post #25





Group: Members
Posts: 107
Joined: 28-February 07
Member No.: 41051



QUOTE (saratoga @ Dec 29 2010, 00:21) *
Assuming you have access to a reasonably good sound card, I would use the method I suggested above so that you can quickly change samples in a double blind manner. Given that your source material will be 16 bit, transparently recording it should be quite easy.


Thanks for your feedback as well. I could go the route you've suggested, but I'm going to see what kind of results I get first by just switching devices. If that test is inconclusive, then I'll do a second test by recording each device. Either way, I'll post the results of the initial test, which I plan to do Friday or Saturday.
Go to the top of the page
+Quote Post

2 Pages V   1 2 >
Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 18th September 2014 - 07:58