right but just calculate what is the RLC network that you need to correct/introduce some kind of significant modification to the overall behaviour of the DUT and if it matches the equivalent network of the cables then it is compensating for it.
As far as I know the RLC's parameter of a cable should be low wnough to not introduce significant variation.
Since I have never taken the time to measure I cable this is just my assumption.
Their measurements are anechoic , stereophiles is not , there are going to be big differences , as such many in the past would use stereophiles measuring system and technique in other to ace the test so to speak ..
Anechoic and gated measurements basically give you the same thing. The anechoic chamber removes reflections from the measurements by not producing any reflections by the nature of its design. A gated measurements on the other hand only allows a small slice in time of the data to be used for the measurement process. The gate naturally removes any reflections, as they occur in time after the initial impulse, so you set the gate to exclude them. Where you set the gate determines how much of the room you decide to let in. The disadvantage of a gated measurement is that the gate length directly affects the low frequency accuracy of the measurement. This is where a true anechoic response is very useful as there are no reflections in the first place, so the gate can be huge and you can measure down to 20Hz.
Given an appropriate setting though the results from the gated measurement and and anechoic measurement will be identical, that is up until the low frequency limit of the gated measurement is reached. In my relatively small listening room which is around 4*3 meters with high ceilings, I can get a accurate data down to around 300-500Hz depending on how much furniture I decide to reshuffle around the room.
Stereophile's John Atkinson uses the standard measurement process for someone who doesn't have an anechoic chamber - a far field measurement, to accurately capture data down to around 300Hz and then he splices in a nearfield measurement for the low frequencies. This works very well, but as the nearfield response occurs 1cm away from the cone of the bass driver it doesn't include the bafflestep transition, which is why lots of loudspeakers show a rising response towards low frequencies. Measured in room, or in an anechoic chamber, this rise wouldn't be there. It doesn't matter as per say that this rise exists and that John can't measure the speakers anechoically because if you're aware of how your measurements are affected by the conditions in which you do them, you can compensate for their effects.
That's a different topic , for many it's not about the destination .....
That's fine as long as the destination is actually reached.
Yeah but to compensate for the 3.5kHz of the B&Ws you'd need some pretty funky cables, ones that look a lot like standard crossover components. You're not going to correct for any mid band resonance with the cables.
It's not just an assumption it's the truth. Any cables should have RLC parameters low enough not to affect the loudspeakers in any significant way, otherwise they'd technically be broken. The only thing you're going to get cables to do is roll off the top octave by a tiny amount.
Besides look at it another way around. The person who's attacking the resonance with the cables - and lets say it works - will have had to introduce just the right combination of RLC to correctly compensate for the peak in the response. Now lets say you used normal cables and put in a capacitor an inductor and a resistor instead, how's this any different? Both have added the same amount of each parameter, it's just that the crossover components did it far more cheaply, with greater accuracy and most likely with far more ideal characteristics as the components would be designed specifically to do their specific job, whereas the effect in the cables is parasitic and most likely not as linear.
You are winging it 5th anechoic measurements are not the same as gated indoors. If you did the comparisons you would see such .. GP measurements outdoors is whats considering close to or better than anechoic ..
I have compared JA's method and setup to Leap/LMS /GP method and it was different . A prudent prerequisite if having speakers tested in stereophile, is to check before submission.Big players like B&W may not need to , buts that another story ..
I have compared JA's method and setup to Leap/LMS /GP method and it was different . A prudent prerequisite if having speakers tested in stereophile, is to check before submission.Big players like B&W may not need to , buts that another story ..
Last edited:
Within the limitations of gated measurements (limited low frequency resolution) they are identical to an anechoic measurement.
Do you think for some reason that you use a different type of measurement when inside an anechoic chamber? No, you use the same MLS type stimulus but in this case you can leave the gate wide open. In both cases you're excluding the reflections that would otherwise contaminate the direct sound, just because one method excludes them by the design of the room and the other does it electronically doesn't mean that the end results are going to be any different, they both end up achieving the same thing.
No it doesn't 5th and I also mentioned GP measurements , obvious to me you're Basing your suposition on theory and have not done the comparison.
In door gated measurements where the speaker is close to the floor and walls are very inaccurate vs the others , I would not trust it much below 700 hz , just watch the settling time required ...
http://www.aes.org/e-lib/browse.cfm?elib=3812
In door gated measurements where the speaker is close to the floor and walls are very inaccurate vs the others , I would not trust it much below 700 hz , just watch the settling time required ...
http://www.aes.org/e-lib/browse.cfm?elib=3812
Interesting debate that goes beyond my knowledge.
I am thinking though could this be the reason why stereophile was unable to match the extremely flat response publish by B&W?
However if I am not understanding wrong these type of measurements shouldn't impact midband thus in my eyes there isn't a reasonable explanation to this incoherent result!
Please let me know your thoughts!
Last edited:
A.wayne seems to be only half reading my posts where I've clearly stated that indoor gated measurements are compromised in the low frequencies. How far down you can accurately get indoors depends on the room dimensions and how the loudspeaker is positioned in it.
As said before I can get accurate data down to around 300-500Hz in my room.
I didn't ignore that fact that you mentioned ground plane measurements, I just chose not to comment on them as they are lumped into the other two. All three methods use different ways to remove reflections from the measurement process, the ground plane does it by creative positioning, the anechoic chamber does it by creating no reflections in the first place and the indoor gated measurement does it by mathematically excluding the data that contains the room reflections.
If used within their limitations they will ALL give the same results at high frequencies, which is the band that we are talking about here. Ground plane measurements also can have issues at higher frequencies too.
And yes Stafanoo you are correct, the gated in door measurement will give exactly the same results as an anechoic measurement at the frequencies we are discussing with regards to the B&W loudspeaker.
It depends usually about 1-1.5 meters. Obviously higher mitigates the floor reflection and allows for a longer gate. Usually I splice at around 500Hz, but if I've rearranged the room accordingly I can get good data down to around 300Hz, these do tend to allow in a touch of the room, but a tiny amount of smoothing generally sorts that out.
Here though is a measurement taken with optimal conditions, this is raw without any smoothing.
So yes, a little bit of the room gets in, a small amount of smoothing easily sorts that out, but the measurements are perfectly adequate for work down to around 300Hz. Raise this by an octave to 600Hz and there are no problems whatsoever, so in terms of resolving the issues the B&W has between 1-10kHz, John Atkinson's gated in room measurements should be more then enough and should not differ much to an anechoic measurement of the same thing.
Here though is a measurement taken with optimal conditions, this is raw without any smoothing.
So yes, a little bit of the room gets in, a small amount of smoothing easily sorts that out, but the measurements are perfectly adequate for work down to around 300Hz. Raise this by an octave to 600Hz and there are no problems whatsoever, so in terms of resolving the issues the B&W has between 1-10kHz, John Atkinson's gated in room measurements should be more then enough and should not differ much to an anechoic measurement of the same thing.
Attachments
No idea what spl the measurement was taken at as this doesn't really matter. Probably around 75-80dB, normal listening levels for that speaker.
If you say it's only accurate down to 1.2kHz then how come all measurement programs, that show you the accuracy cut off for the given gate length, say different? ARTA says this is accurate down to just below 300Hz providing I've set the gate accordingly, as does Holm. They ain't gonna lie about something which is presumably calculated mathematically.
Besides what makes you choose 1.2kHz?
If you say it's only accurate down to 1.2kHz then how come all measurement programs, that show you the accuracy cut off for the given gate length, say different? ARTA says this is accurate down to just below 300Hz providing I've set the gate accordingly, as does Holm. They ain't gonna lie about something which is presumably calculated mathematically.
Besides what makes you choose 1.2kHz?
I spent about 9 years working with Meyer's SIM2 and SIM3 measurement systems doing repair for a Broadway rental house. The truth is you can get some funky traces depending on the environment you are measuring in. The beauty of those particular system's is you have a visual indication of where the software isn't sure of the result- for SIM it's the Coherence trace. Many other pieces of measurement software do this, but SIM always seemed to perform a bit better. Then you can adjust the measurement window to suit the environment and look at your relative phase trace to get the best result trace.
What we ended up doing to have the best chance indoors was use a bunch of acoustic absorption panels. We used Sorber panels made by ClearSonics which are commonly used for vocal and drum booths for live concerts and shows. We made a sort of portable booth surrounding the SUT with 10 feet or so of panels on the floor in front of the SUT and a couple behind the measurement mic (usually a DPA 4007).
This was in early 1900's converted railway warehouses with finished plywood floors and parallel brick walls.
We found that compared to manufacturers traces done in their own anechoic chamber, we could get very close. Certainly close enough to confirm a within spec window of +/- 3 dB.
What we ended up doing to have the best chance indoors was use a bunch of acoustic absorption panels. We used Sorber panels made by ClearSonics which are commonly used for vocal and drum booths for live concerts and shows. We made a sort of portable booth surrounding the SUT with 10 feet or so of panels on the floor in front of the SUT and a couple behind the measurement mic (usually a DPA 4007).
This was in early 1900's converted railway warehouses with finished plywood floors and parallel brick walls.
We found that compared to manufacturers traces done in their own anechoic chamber, we could get very close. Certainly close enough to confirm a within spec window of +/- 3 dB.
I suppose back then things we're quite a bit different indeed. ARTA lets you view the impulse response directly, whilst setting the gate. This gives you a very good visual representation of what the captured response (room included) is like and you can see the response captured in time, seeing when each of the reflections arrive at the mic.
The issue at the moment, at least from my point of view, is that the data captured is accurate and the gate has been set accordingly. With those conditions ARTA says the data is accurate down to around 300Hz, but A.wayne say's it's only accurate to 1200Hz, what I want to know is why he thinks that and why then is ARTA being misleading.
The issue at the moment, at least from my point of view, is that the data captured is accurate and the gate has been set accordingly. With those conditions ARTA says the data is accurate down to around 300Hz, but A.wayne say's it's only accurate to 1200Hz, what I want to know is why he thinks that and why then is ARTA being misleading.
and that is what I thought: could have the process meaasurement have been done wrong on stereophile's hand?
Could this cause in your opinion a significant discrepancy with the manufacturer's trace?
checkin out
Without knowing what software was used (I went back a bit to find the article in the thread but no luck) all I can say is it's possible to interpret the results incorrectly.
However, the software that I have used- SIM, SMAART, and Spectrafoo all show you when to question the measurements. If the phase trace looks good, the coherence trace looks good, your mic is good, and you've reasonably treated the area for spurious reflections if indoors you have an excellent foundation for good data.
5th,
I'm pretty sure the all of the above software does impulse response. SIM doesn't do the pretty spectral maps, but SMAART and Foo certainly do.
However, the software that I have used- SIM, SMAART, and Spectrafoo all show you when to question the measurements. If the phase trace looks good, the coherence trace looks good, your mic is good, and you've reasonably treated the area for spurious reflections if indoors you have an excellent foundation for good data.
5th,
I'm pretty sure the all of the above software does impulse response. SIM doesn't do the pretty spectral maps, but SMAART and Foo certainly do.
the measurements discussed are made by stereophile, here is the link
http://www.google.com/url?sa=t&rct=...8ICQDw&usg=AFQjCNFtwjoDUIYIsVhcvDs6g2l1CqEhFw
http://www.google.com/url?sa=t&rct=...8ICQDw&usg=AFQjCNFtwjoDUIYIsVhcvDs6g2l1CqEhFw
Well with either measurement technique what you're doing is effectively isolating the direct sound from any reflected sound. An anechoic chamber doesn't have any reflections and the gated measurement only looks at the first few milliseconds of sound and so isolates it that way. If everything is set up correctly, the only way that the gated measurement is going to become compromised is if you set the gate incorrectly and allow too many reflections in. This looks quite horrible on the frequency response. You can see a little of this in the response I posted above, superimposed over the frequency response you've got small wiggles, up-down-up-down, like passband ripple in a DACs brick wall filter (it's effectively comb filtering in this case). If you let more reflections in these get progressively worse to the point where you don't know if its coming from the loudspeaker or if its a reflection. One can counter this to a degree by using smoothing, but smoothing will obviously smooth out any peaks that are actually there, so anything beyond 1/12/octave is usually shunned, at least by me. John Atkinson's measurements don't show telltale signs of being heavily smoothed so he must be gating his measurements correctly and using an appropriate hardware set up to allow a reasonably large gate.
Even so, the shortfall of gated measurements is their low frequency accuracy, measuring 3.5kHz accurately is extremely easy.
As you can appreciate B&W could easily apply smoothing to their curves to lessen the impact of this peak and if they wanted they could simply photo-shop it out, it wouldn't be the first time that manufactures have done this.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.