GR Research LGK 2.0 Speaker Review (A Joke)

[Mark_A]

There is a issue with having a power handling test in that it can be a destructive test. What @amirm does is NDT (non-destructive testing.). To use destructive tests is to need a repair workbench, more instrumentation, spare parts, many tools, the space for that and money for parts to repair peoples' gear that has been damaged during testing. That's why he does not heat/power soak gear in testing it. To avoid break-down issues.

I was not suggesting that anyone develop a speaker power handling test. That is the most irrelevant and abused spec in all of audio, because if a speaker is efficient enough, it doesn't require a lot of power.

I was referring to a test of two (or more) frequencies played at once, and measuring the distortion, as described in this post below, and in the YouTube video that Amir did that I provided a link for. But instead of listening to music with two or more tones played at once, develop a signal test that outputs 2 or frequencies at once and measures the speaker distortion.

Buchardt S400 Speaker Review

I didn't say that we can always listen to speakers or other audio products before we buy them, I said that is the ideal situation, especially for speakers. For example, the S400 is only available via direct order from the manufacturer in Denmark. Did you read any listening test reviews before...

www.audiosciencereview.com

 

[Doodski]

I was referring to a test of two (or more) frequencies played at once, and measuring the distortion, as described in this post below, and in the YouTube video that Amir did that I provided a link for. But instead of listening to music with two or more tones played at once, develop a signal test that outputs 2 or frequencies at once and measures the speaker distortion.

That would be handy for sure. A speaker multitone test.

 

[Doodski]

@Mark_A I just remembered that @amirm is already doing distortion tests for speakers. It's all in the graphs.

 

[amirm]

I was referring to a test of two (or more) frequencies played at once, and measuring the distortion, as described in this post below, and in the YouTube video that Amir did that I provided a link for. But instead of listening to music with two or more tones played at once, develop a signal test that outputs 2 or frequencies at once and measures the speaker distortion.

It is not a practical test for speakers. The amplitude of each tone will vary based on frequency response. As a result, so will the intermodulation distortion. What this means is that it will be impossible to compare the results of one speaker to another. I have run such tests and you get a bunch of sidebands that are very hard to interpret.

 

[Mark_A]

It is not a practical test for speakers. The amplitude of each tone will vary based on frequency response. As a result, so will the intermodulation distortion. What this means is that it will be impossible to compare the results of one speaker to another. I have run such tests and you get a bunch of sidebands that are very hard to interpret.

I was responding to a claim from someone on this forum that said all differences in audio equipment can be measured (as opposed to doing a listening test). So it sounds like for speakers, a listening test is still useful.

 

[amirm]

Yeh, the last 10 to 30% is not quantifiable with measurements or at least, not easy to translate into audible effect. This gets to be less of an issue as you go up in price and compromises vanish.

 

[Mark_A]

EDIT: So measuring speakers with electronics is more than a capable system.

See Amir's comment here:

GR Research LGK 2.0 Speaker Review (A Joke)

There is a issue with having a power handling test in that it can be a destructive test. What @amirm does is NDT (non-destructive testing.). To use destructive tests is to need a repair workbench, more instrumentation, spare parts, many tools, the space for that and money for parts to repair...

www.audiosciencereview.com

 

[YSC]

Yeh, the last 10 to 30% is not quantifiable with measurements or at least, not easy to translate into audible effect. This gets to be less of an issue as you go up in price and compromises vanish.

May I add my personal 2 cents? so far from my impression, the current measurement set characterize the speaker pretty good enough.
But since for speakers room modes play such a huge effect, with speaker placement also contributing a lot onf even higher frequency reflection issues, so in OUR living space the same speaker can vary a lot, plus personal taste of balance listening test contribute to the final 10-30% IRL. But I do think when expectations are not met it will be a good way to do some in room measurement to see if we can find a problem.

And that the listening need means that a great measuring speaker might not sound good in your placement, but basically a broken speaker won't sound good in any room when comparison can be done

 

[mhardy6647]

There is a issue with having a power handling test in that it can be a destructive test. What @amirm does is NDT (non-destructive testing.). To use destructive tests is to need a repair workbench, more instrumentation, spare parts, many tools, the space for that and money for parts to repair peoples' gear that has been damaged during testing. That's why he does not heat/power soak gear in testing it. To avoid break-down issues.

Unlike the way we establish weight limits on bridges.

 

[Robbo99999]

It is not a practical test for speakers. The amplitude of each tone will vary based on frequency response. As a result, so will the intermodulation distortion. What this means is that it will be impossible to compare the results of one speaker to another. I have run such tests and you get a bunch of sidebands that are very hard to interpret.

I mentioned a potential solution (if it is one) for this before in another thread, but you didn't respond...... you could use your Klippel anechoic results to EQ each speaker to the same relative dB for each of the two tones you'd play together. So in other words for the two frequencies that you'd play together you'd EQ each frequency point to a consistent point on a target curve of your choice. That way you're taking frequency response differences out of the equation and would be showing the "IMD Performance Potential" of the speaker (when EQ'd to for example Anechoic Flat).

EDIT: from an educational point of view it would be useful if you would deign to respond, as myself & others could well learn if proposals have drawbacks or aren't relevant.

 

[Mark_A]

I mentioned a potential solution (if it is one) for this before in another thread, but you didn't respond...... you could use your Klippel anechoic results to EQ each speaker to the same relative dB for each of the two tones you'd play together. So in other words for the two frequencies that you'd play together you'd EQ each frequency point to a consistent point on a target curve of your choice. That way you're taking frequency response differences out of the equation and would be showing the "IMD Performance Potential" of the speaker (when EQ'd to for example Anechoic Flat).

EDIT: from an educational point of view it would be useful if you would deign to respond, as myself & others could well learn if proposals have drawbacks or aren't relevant.

It's not a frequency response problem. It is a distortion problem when trying to play low bass and upper-mids on the same driver at the same time (that he noticed when playing a music track). Amir describes it the YouTube video review of the LGK 2.0 (starting at the relevant part of the video).

 

[Mark_A]

May I add my personal 2 cents? so far from my impression, the current measurement set characterize the speaker pretty good enough.
But since for speakers room modes play such a huge effect, with speaker placement also contributing a lot onf even higher frequency reflection issues, so in OUR living space the same speaker can vary a lot, plus personal taste of balance listening test contribute to the final 10-30% IRL. But I do think when expectations are not met it will be a good way to do some in room measurement to see if we can find a problem.

And that the listening need means that a great measuring speaker might not sound good in your placement, but basically a broken speaker won't sound good in any room when comparison can be done

The problem that Amir noticed (that he found in a listening test) is not related to frequency response, so room boundaries and speaker placement don't matter in this particular case. It was a driver distortion problem when a music track was played with bass frequencies and a female voice at the exact same time, and since the system was a single driver speaker, that one driver had particularly bad distortion playing both frequencies at once, versus one frequency at a time that is typically done via some technical measurements.

 

[rubley00]

No problem. I don't know about Copernicus, but Galileo antagonized a lot of people.

There is a theoretical physicist named named Sabine Hossenfelder who has antagonized a lot of her peers by saying:

I love Sabine, and agree with her on all points.

 

[rubley00]

May I add my personal 2 cents? so far from my impression, the current measurement set characterize the speaker pretty good enough.
But since for speakers room modes play such a huge effect, with speaker placement also contributing a lot onf even higher frequency reflection issues, so in OUR living space the same speaker can vary a lot, plus personal taste of balance listening test contribute to the final 10-30% IRL. But I do think when expectations are not met it will be a good way to do some in room measurement to see if we can find a problem.

And that the listening need means that a great measuring speaker might not sound good in your placement, but basically a broken speaker won't sound good in any room when comparison can be done

Its impossible to come up with an "average" room, there's no point in trying. Which is why no one measures speakers in a room.
It would be the opposite of helpful. It would obfuscate the true performance of the speaker.

 

[rubley00]

I was responding to a claim from someone on this forum that said all differences in audio equipment can be measured (as opposed to doing a listening test). So it sounds like for speakers, a listening test is still useful.

All differences could be measured. There is nothing that can be heard that can't be measured. If a speaker doesn't live up to the expectations from testing, then that points to a gap in the testing regime. It may be hard to come up with enough tests to reveal every possible problem, but theoretically its possible.

 

[YSC]

Its impossible to come up with an "average" room, there's no point in trying. Which is why no one measures speakers in a room.
It would be the opposite of helpful. It would obfuscate the true performance of the speaker.

No I don’t mean get an average room and so review with one, I mean when trying a speaker in our own room, the in room measurement will be more useful see what is the actual problem if we don’t like

 

[Robbo99999]

I was not suggesting that anyone develop a speaker power handling test. That is the most irrelevant and abused spec in all of audio, because if a speaker is efficient enough, it doesn't require a lot of power.

I was referring to a test of two (or more) frequencies played at once, and measuring the distortion, as described in this post below, and in the YouTube video that Amir did that I provided a link for. But instead of listening to music with two or more tones played at once, develop a signal test that outputs 2 or frequencies at once and measures the speaker distortion.

Buchardt S400 Speaker Review

I didn't say that we can always listen to speakers or other audio products before we buy them, I said that is the ideal situation, especially for speakers. For example, the S400 is only available via direct order from the manufacturer in Denmark. Did you read any listening test reviews before...

www.audiosciencereview.com

It is not a practical test for speakers. The amplitude of each tone will vary based on frequency response. As a result, so will the intermodulation distortion. What this means is that it will be impossible to compare the results of one speaker to another. I have run such tests and you get a bunch of sidebands that are very hard to interpret.

I mentioned a potential solution (if it is one) for this before in another thread, but you didn't respond...... you could use your Klippel anechoic results to EQ each speaker to the same relative dB for each of the two tones you'd play together. So in other words for the two frequencies that you'd play together you'd EQ each frequency point to a consistent point on a target curve of your choice. That way you're taking frequency response differences out of the equation and would be showing the "IMD Performance Potential" of the speaker (when EQ'd to for example Anechoic Flat).

EDIT: from an educational point of view it would be useful if you would deign to respond, as myself & others could well learn if proposals have drawbacks or aren't relevant.

It's not a frequency response problem. It is a distortion problem when trying to play low bass and upper-mids on the same driver at the same time (that he noticed when playing a music track). Amir describes it the YouTube video review of the LGK 2.0 (starting at the relevant part of the video).

@Mark_A (in response to your last quote above) All those quotes above is the flow of the discussion, you can see in Amir's post (the second quote) that he is indeed saying it's a frequency response problem, in terms of not being able to do a reliable IMD test when 2 different frequencies are being played at the same time. And as you can see I countered that by proposing that he EQ's the frequency response at the two points to the same curve for each speaker he would test for IMD, thereby taking the frequency response differences of the speakers out of the equation - which was Amir's argument for not doing the test. So you can see I'm actually trying to support your suggestion, or at least suggest a way for Amir to get round the problems that he explained were preventing him from doing the IMD test. He's mentioned such a reason for not doing them before, in another thread a few weeks ago perhaps, and I proposed to him the same potential solution re EQ'ing the two points to the same curve, and he didn't respond then & he didn't respond yet to me now either - to be honest I'm a bit annoyed by it, as I'm only trying to offer a solution, and if it's not a solution then from an educational point of view for myself (& others reading this forum) it would be useful if Amir responded with a "Yes, that would work", or "No, we can't do that because of X, Y, or Z"......I think it's pretty bad he's just ignored my idea. (Instead, if he keeps ignoring my idea for the solution, then it just makes me think that he knows the solution would work, but would just create more work for himself....that's what I conclude (& perhaps other forum readers too) unless he deigns to answer me.)

 

[amirm]

And as you can see I countered that by proposing that he EQ's the frequency response at the two points to the same curve for each speaker he would test for IMD, thereby taking the frequency response differences of the speakers out of the equation - which was Amir's argument for not doing the test.

What two frequency? How do you pick them and keep them the same across many speakers?

I also don't know what you mean by EQ. Are you asking for an ultra-sharp (high-Q) filter to boost the one frequency? How does that not change the behavior of the speaker?

 

[amirm]

.I think it's pretty bad he's just ignored my idea. (Instead, if he keeps ignoring my idea for the solution, then it just makes me think that he knows the solution would work, but would just create more work for himself....that's what I conclude (& perhaps other forum readers too) unless he deigns to answer me.)

You haven't remotely addressed the problem at hand. See my previous post.

 

[Robbo99999]

What two frequency? How do you pick them and keep them the same across many speakers?

I also don't know what you mean by EQ. Are you asking for an ultra-sharp (high-Q) filter to boost the one frequency? How does that not change the behavior of the speaker?

Hey, well thankyou for finally replying to one of my posts where I asked you a question. I'm only barbed about it because recently you've not been replying to any of my questions, and it's not like I spam you with questions all the time, so I was starting to take it personally. If we take all the emotion out of it though, you'll see that my intial post to you was a simple suggestion of a solution, I never said I knew it was a solution - it's called a discussion.

Regarding the Q of the filters, you wouldn't have to use a sharp Q because you're only testing at two frequency points, so you could use a broad 1.5 to 2.5 Q if you wanted (because no need to EQ the whole frequency response to the target curve). It wouldn't matter which frequency of the two you boost, you'd just make sure that both of your chosen frequencies were sitting on the same target curve - could be an anechoic flat curve if you liked. And of course it wouldn't matter if the other points on your frequency response that weren't being tested weren't EQ'd to the target curve, because they're not being tested of course.....it would be a simple matter of just making sure your two tested frequency response points are on the target curve.

Yeah, regarding choosing two frequencies, I don't know, how would you choose which two frequencies........I suppose you're concerned that just picking two frequencies would mean there's a chance it wouldn't describe the overall IMD behaviour of the speaker, as you might just get lucky or unlucky with where those frequencies would land which might build an unrepresentative picture of the speaker. I suppose one way of getting round that would be to have a regime where you test more than just one pair of frequencies, you might test 5 frequency pairs across the whole frequency response. So maybe you'd do 50Hz & 2000Hz, 100Hz & 3000Hz, 150Hz & 4000Hz, 200Hz & 5000Hz, etc. Or perhaps if we want to just focus on the woofer, which might be more relevant if we're doing it to see if distortion in bass can influence the rest of the frequency range then maybe you'd narrow the gap down to testing say 50Hz & 1000Hz, 55Hz & 900Hz, 60Hz & 800Hz, 70Hz & 700Hz. I don't know, but this is part of the discussion right, a back & forth, arriving at and taking the best of what the participants in the discussion can bring as well as stimulating ideas in other people.

(EDIT: in headphone testing you could widen the gap between frequencies and they'd still be playing on the same driver, as just one driver in a headphone.)