optical
Moderator
BIG STAR
Be Excellent To Eachother
Posts: 1,623 
|
Post by optical on Mar 26, 2024 11:09:16 GMT
Not great though, are they. They could be better, sure. Although is it pretty rare to find speakers with an accuracy better than +/-2.5db from 40-20000Hz. The suckout on those speakers and lift from 4-20kHz could be sorted easily with a XO tweak or a little tweeter attenuation. TBF, most manufacturers intentionally add that lift, it draws people in - many are suckers for forced presence and clarity. I'm always wary of magazine measurements too. It's very difficult to achieve great accuracy doing summed nearfield measurements. The error margin is probably 2-3db at least anyway. optical and I have been tweaking our systems/rooms for months. We generally find a deviation of a few db in Frequency can be fine. It is usually more important to minimise distortion and sort decay times in room. As Jason eludes to . . . . after a few months of learning/playing/tweaking the actual real world FR plot only tells so much of the story. For example I have a few "top-tier" measurements from my room, the ones I would submit if trying to show-off/sell a product etc. Even with 1/6th (and more so 1/3rd) smoothing, there are only small variations from 25hz to 200hz (the 'active' part of my spectrum). Using Psychoacoustic smoothing, I can get it to within 1-2 db in that area, in a fairly problematic room . . . . however, does that tuning sound the best. Nope. That 'tuning' requires me to run some fairly extreme boosts/cuts at neighbouring frequencies (always sounds bad) as well as getting creative with somewhat sharp Q factors applied to those boosts and cuts. Yes it can certainly flatten your FR plot but at the cost of distortion and general audible nasties, lowering the overall clarity of the presentation. Couple that with the bass drivers on the speakers doing an 'unequal' amount of work (doesn't matter too much below 60hz I've found due to omni-directional bass) above 60hz it can really throw things off. In my room, to make appreciable tweaks I really need to overlay the responses from both L&R speakers independently and try and level the output from both to achieve the best balance, rather than just chasing the flatness of that line. Most of these measurements have a fair amount of smoothing applied also BUT the stronger the smoothing the less detailed the information is. So for example I can have two measurements that look very similar with 'normal' smoothing (13rd/Psychoacoustic etc) but then moving to 16th and higher one measurement can easily have a 10db+ peak/dip somewhere along the curve which may only be between say 500-600hz but because there may be a peak at 450hz and 650hz, the 'smoothing' evens them out and you are presented with a visually 'flat' curve. It's not quite how your ears hear it. Thusly a curve with no smoothing applied could look the same as that but because there are no peaks at 450hz and 650hz (and indeed no suck out between 500-600hz) the two final FR plots will look practically the same . . . . Another factor is mic positioning, I can manipulate the curve significantly (3-5db+) by moving the mic an inch or two in any direction, up down, left, right etc. All these things have to be factored in. Granted my room is more problematic than most, but these factors will apply in all but the most well designed, acoustically sympathetic layouts, and they will still be present but just to a lesser extent. Like all measurements, the facts are all there in the information presented BUT it never tells the whole story, certainly not without delving into further figures relating to that FR plot. Then there is the knowledge required to interpret the further information available, which I am also hugely lacking in! My summery so far would be to not have huge boosts/cuts (10db+) anywhere on your DSP. There is no point in 'fighting' the physical limitations of the room, (at least not without proper and expensive treatment). Keep distortion as low as possible even at the cost of the 'flatness' of the curve. Pay attention to RT60 decay times and get them as even as possible. It doesn't seem to matter if they are a bit lower or a bit higher than what they should be for the room volume/size, if they are generally even (above 100hz anyway), it should sound okay. 'Some' treatment, be it furnishings, plants, bookshelves or indeed panels/rockwool, canvasses etc, can make a huge difference. Much, much bigger than people realise, i think. These differences are almost never represented in the FR plot either because it deals with echo/reverb/decay times etc as well as removing distortion in my case. Take a lot of the treatment in my room out and the FR plot is very very similar however it sounds like the speakers are playing in the room downstairs by comparison. |
|
|
|
Post by misterc on Mar 26, 2024 11:58:38 GMT
Seriously the room speaker interface is everything, I have said this for over 25 years, get this right from the outset and it will pay dividends. Also as Chris has suggested the measurement technique is AS important and the measurement itself, we have a saying here situational awareness of your signal integrity in that understanding WHY you are getting the results you are seeing on the device that is performing the measurements. Expect to see what you expect to see, if not then its either operator error in setting up the experiment, the equipment is faulty or out of calibration. You may well have a third party 'offset' that is causing you to view the results which you are not expecting, mains power fluctuation, low air pressure/temperature coefficients/emi/rf discharge/shot noise etc.
Always keep grounded and always benchmark you test equipment against a known reference before starting to perform these tests. It is also very easy to 'adjust' results to suit the manufacturers/magazines/review site's agenda. conversely how the results are interpreted by the end user the we are back to the old pre requisite of 10 chaps in a room, 5 will fell the system is amazing, four will feel it's crap and the last one will not be able to deliver an opinion until he has consulted his best mate and 16 on line communities With regard to the Wadax, the sheer amount of genuine R&D time and investment that went into that product is something else, it really up there with the best aerospace technology. Wadax is owned by a EU company that turns over billions of euro’s a year and their resources they have at their disposal is HUGE. Honestly their power supplies on those butt ugly devices are what I could achieve if I had that amount of industrial ability behind me. Respect for the sheer amount of time on producing these devices. Cost wise as with everything yes it’s a house for many people and you can’t live in or on a Wadax, its not going to keep you dry and warm lol.
That’s not the point, people will only pay for items what they feel they are worth whatever that item is so personal (taking the purchase cost out of the equation) Its that simple. In the same way why do so many wealthy people not help out the less fortunate and those in situations that genuinely would challenge anyone.
Martin if you can genuinely reproduce that exact Boyer room sound for £20K I’ll set up a Stoke on trent Branch and install you as dealer principle, plus stock it as well. |
|
Arke
Moderator
Posts: 1,259
|
Post by Arke on Mar 26, 2024 12:39:21 GMT
an object lesson in how not to do it. You wouldn't pay fifty quid for a banana because it's an obvious rip-off. £200K for a music server? Just ridiculous. What do people think they are buying? Give me same room and a £20K budget I'd equal or better that system no problem at all. Want to go second-hand? I'll do it for less than half that. 'I'd like a 4dB boost over 4KHz please' 'No problem sir that will be £29K'  Ah, found it! You've found a Kroma speaker review (same brand, but different speakers to the Boyer room): www.hifinews.com/content/kroma-atelier-stella-xtreme-loudspeakerAs Mr C and Chris say above, FR measurements have to be taken with a little pinch of salt! These measurements, in particular, will likely have a suckout in 1-4Khz region. They explicitly say they measured nearfield above the tweeter. This is a speaker with an MTM arrangement and a 'bass' driver underneath (handling under 450Hz). If you measure an MTM arrangement like this nearfield above the tweeter, the lower mid/bass (of the MTM) will be very off-axis. That mid/bass will have a drop in output above 1kHz when off-axis, therefore measuring in this manner will give a suckout and relative HF lift. IMO those FR deviations are mostly a function of the measurement technique. This is the problem with measurements (as I have highlighted on many occasions). Very few people truly understand how they were actually taken, and more importantly, one needs to understand why any particular measurements are potentially flawed. I'd imagine those speakers may measure OK if the measurements were taken in a more appropriate manner. |
|
|
|
Post by misterc on Mar 26, 2024 12:43:22 GMT
Arh if its HFN, I've seen Keith take measurments with a speaker from a microphone taped to a stand as the bracket was broke at a jaunty angle and saying well its good enough 1000% gospel, these were performed in his front room.
|
|
Arke
Moderator
Posts: 1,259
|
Post by Arke on Mar 26, 2024 12:57:11 GMT
Arh if its HFN, I've seen Keith take measurments with a speaker from a microphone taped to a stand as the bracket was broke at a jaunty angle and saying well its good enough 1000% gospel, these were performed in his front room. My point exactly Tony. Those measurements will be rubbish based on how he did it. So hard to measure a true speaker response and remove the room & measurement technique issues. I like speakers to have consistent performance in many rooms and be a universally 'very good' performance. Ultimately, you listen and measure in your room at your listening position. Optimise it there and you're on to a winner. |
|
|
|
Post by misterc on Mar 26, 2024 13:10:41 GMT
In these sutaions I feel an independant testing house would be a far more accurate and accpetable method of obtaining these results both in a chamber and a standeridsed sized living room enviroment. That way the manufacturers could be sure of having more reatic result that wouldn't call for 26 pages a year in a out moded and antiquatied form of advertsing. PM should stick to Railway Modelier and Crochet weekly they ceratinly earn him far more income than HFN and Stereobollocks do, arh but then again the Ego wouldn't be stroked  |
|
|
|
Post by macca on Mar 26, 2024 18:29:16 GMT
Martin if you can genuinely reproduce that exact Boyer room sound for £20K I’ll set up a Stoke on trent Branch and install you as dealer principle, plus stock it as well. I'm your man! |
|
|
|
Post by macca on Mar 26, 2024 18:38:09 GMT
Re HFN measurements yes they are a bit crude but where they and Stereophile have measured the same speaker they do tend to be consistent. In room measurements do not really give a true picture of 'what it will sound like' above the transition frequency. So it's important not to confuse in-room measurements with anechoic or quasi-anechoic measurements as shown in the mags or forums like ASE or on Erin's audio corner. The latter two use a Klippel scanner so you can take those to the bank. Interestingly where they measure the same speaker as Stereophile the results are pretty close. I'd say they all know what they're doing although HFW really should offer a full suite like the others do. Just showing on axis FR isn't great but as I said on the other thread it's enough to tell you what to avoid. I couldn't find any measurements of the speaker at the show, so picked another one from the same brand. Sure you could maybe sort the issue with EQ (if the off-axis is good too which we don't know) but why should you have to for that kind of cash?> I suppose if you have hearing loss above 3Khz then job's a good un.  |
|
|
|
Post by misterc on Mar 26, 2024 19:14:36 GMT
Klippel is the gold standard this is around 100K Euro's for the FULL suit sans headphones I only know of two audio companies that have this facility in the UK, although I do have access to one via a private indivdual ownership. However I don't require to use one of these personally as I'm not a speaker designer, however I'm not to shabby at making speakers work correctly in domestic enivornments. I did consider it last year, but in all honesty I'm not a speaker manufacturer.......
That outlay could be put to good use that gets me a 1/3 of the new high speed digital signal analyser, a new lab building for our new non UK residency, a couple of quality car's plus a nice 6 week cruise plus some spending cash etc. Although this leads us back to a previous post, the equipment is only as good as the person using it and understanding it's capbilities/neuonces and the ability to truly understand the results imho |
|
Arke
Moderator
Posts: 1,259
|
Post by Arke on Mar 26, 2024 19:19:04 GMT
Re HFN measurements yes they are a bit crude but where they and Stereophile have measured the same speaker they do tend to be consistent. In room measurements do not really give a true picture of 'what it will sound like' above the transition frequency. So it's important not to confuse in-room measurements with anechoic or quasi-anechoic measurements as shown in the mags or forums like ASE or on Erin's audio corner. The latter two use a Klippel scanner so you can take those to the bank. Interestingly where they measure the same speaker as Stereophile the results are pretty close. I'd say they all know what they're doing although HFW really should offer a full suite like the others do. Just showing on axis FR isn't great but as I said on the other thread it's enough to tell you what to avoid. I couldn't find any measurements of the speaker at the show, so picked another one from the same brand. Sure you could maybe sort the issue with EQ (if the off-axis is good too which we don't know) but why should you have to for that kind of cash?> I suppose if you have hearing loss above 3Khz then job's a good un. I will save other forum members the tedium of covering the same ground over and over again. If you fully understand the Stereophile and HFN (and other's) measurements for selecting speakers then that is great. You obviously understand the measuring of speakers, the limitations, bias and the multitude of potential errors. I would recommend this way of choosing (along with home demo) to anyone who understands the Frequency response, distortion, impulse and impedance measurements completely, AND fully understands the methods used and the inherent limitations and error margins. |
|
