Jump to content

fredbatch

Wammer
  • Posts

    122
  • Joined

  • Last visited

Reputation

107 Excellent

About fredbatch

  • Rank
    Junior Wammer
    Junior Wammer

Personal Info

  • Location
    North Wales

Wigwam Info

  • Trade Status
    I am not in the Hi-Fi trade

Recent Profile Visitors

817 profile views
  1. Hello Mike, Well, there’s a question, but restraining myself from another whimsical response, the following is genuine: Close to front wall provides the following benefits: Optimises bass reinforcement whilst limiting speaker boundary interference response (“SBIR”) which works well for speakers such as Keltiks, originally designed to be placed relatively close to front wall. Cancellation nulls move higher up the frequency range as the speakers move closer to the wall. For forward facing bass drivers, distance between front face of speaker and wall derives 1/4 wavelength cancellation frequency. In conjunction with SO, this serves to provide a fast and realistic LF response devoid of nasty LF modal peaks, for the most part. Overall, I find this produces a more natural and effortless sound compared to pulling them further into the room, as well as being more aesthetically pleasing. Arguably, imaging may be compromised, but I cannot say I have found this to outweigh the benefits of the big spacious Keltik sound, pretty much guaranteed, when separated by about 3m and set close and parallel to the wall. I don’t usually find toe-in to be beneficial either. With good quality wide dispersion tweeters I try to avoid direct HF-ear targeting. Plus, my Keltik room errs towards the live side so this further helps mitigate any overtly bright characteristics. Leaving plenty of side space (> 1m, typically in my case) delays secondary reflections off side walls enough to prevent them disrupting the perception of the primary direct sound. I have a mental image of increasing fuzziness as the speakers get closer to the side walls so this may well bias my perception a little. I have now used Keltiks in three different rooms with different floor structures, and categorically prefer hard-coupled, solid, non-suspended floors. Springy floors without some form of decoupling can be a bit of a soggy spoiler. Every time I have also found placing them on the long wall firing across the room delivers a more controlled and engaging sound. Ultimately, all these factors blend to provide a particular combination, so it is always best to listen and experiment in order to decide. The relative position of all room boundaries with respect to loudspeaker cabinet placement and driver positions will interact to cause a myriad of reflections and comb filtering effects across the whole frequency spectrum to build a unique sonic picture. When further taking account of absorption/reverberation effects, things are clearly not quite black and white.
  2. I have a feeling of déjà vu and always anticipated some yankers would come along and insist they could hear a difference in such circumstances.
  3. I’m sure others have read the recent special feature about consciousness in New Scientist. It is both fascinating and in places mind bending: “The universe could have begun as a quantum system and continued evolving (quantum) mechanically until matter first became able to integrate information (consciousness). This consciousness then started to collapse quantum reality, creating the sort of classical reality we experience today”. One of the models of consciousness mentioned is based on “Predictive Processing” “The brain is a prediction machine, meaning that what we perceive is the brain’s best guesses about the causes of its sensory input. As a result, much of conscious experience and selfhood is based on what we expect, not what is there”. I’m sure I just heard a bell ring.
  4. I managed to have another play in the last week or so with the Exakt design filters and I was pleasantly surprised to find a simple way to use SO2 with bespoke Exakt design filters due to what appears got be an interactive glitch between Konfig and the Linn account. Since this is clearly not intentional or supported I won’t say any more, but it is certainly a lot easier and faster than reverse engineering SO2 profiles by cutting and pasting xml data into corresponding SO1 profiles. Following on from a few of the other comments, the development forum, as far as I know, is not selective or elitist. With the demise of the original Linn forum, volunteers were sought to provide feedback and report on issues relating to Linn software implementation. There appears to be very little mention relating to Exakt Design in these pages. However, the DIY forum here provides at least one useful and constructive thread. Otherwise, there is quite a bit of documentation within the Exakt Design Application itself and I know Philbo has been very helpful in providing assistance where required. Personally, I have largely ploughed my own furrow, other than liaising with Finster. I’m not really sure what level of support Linn are expected to offer people using this application but given that it is something of a fringe product, combined with all the other in-house development going on, I imagine resources might be a bit tight. I will eventually get round to posting a few comparative graphs to illustrate some of the relative differences between bespoke and standard filters used with my brand of Keltiks. It all seems very quiet on the western front at the moment, apparently there’s some important football match going on.
  5. Thanks Paul, Craigas. Jail, as Andrew says, the UMIK-1 (calibrated) seems to be the most popular choice and is readily compatible with REW. It is available from miniDSP. Strictly speaking, this is not a prerequisite, but being able to make comparative measurements with REW + mic is a very useful tool. Since most of us don’t have access to an anechoic chamber and, particularly with 50kg speakers, hanging them 20 feet high on an outside pole is not an appealing prospect, we are measuring the effects of the room/speaker interaction, most of the time. Even with “gated” and nearfield measurements we are seeing the effect of dominant room reflections, particularly when interpreting the lower FR. So, caution is required when making value judgements. As far as Exakt Design goes, one of its objectives is to obtain a linear phase response and to achieve this here, it is necessary to acquire detailed information about the individual driver electromechanical properties, described by their Thiele-Small parameters. These are often quoted on spec. sheets but inevitably will vary in reality and sometimes by quite high margins. To obtain this information accurately, impedance measurements need to be made, from which corresponding T/S values can be calculated. This can be performed by making a basic circuit and measuring the impedance variation with frequency using a suitable meter or it can be more easily performed using a suitable resistor in conjunction with the REW application. Easier still, and likely more accurate and versatile, Daytona DATS 3 hardware/software can be used to measure impedance. This removes a lot of the manual calculation from the equation, as it were.
  6. So, an update on progress with the work I have been doing to create a bespoke Exakt filter for my Keltiks which use SEAS M15CH002's for mids with Hiquphon OW2 on HF duty. I have stopped waiting for SO2 to become compatible and have now spent some some time finishing the new filter. I can conclude that I happy with the results which now supercedes the performance of the native 007/038 filter I had previously adopted. I tried various values for the crossover frequencies. For bass/mid: 119 Hz, 200 Hz, having obtained different advice and information from various sources but without being able to obtain confirmation from Linn about the original specification, to date. I have subsequently settled on 200 Hz and was able to make sense of the bass impedance measurements, having considered individual, parallel and sereies figures for in-situ drivers, which led to a consistent set of results (now the expected single peak for closed box scenario). For mid to HF I have tried 2.4 kHz and 2.7 kHz, eventually settling on the former. Also, latterly I have been talking to Finster who has been through a similar scenario, designing an Exakt filter using OW1's and rolling off sooner in the bass to accomodate a subwoofer system. He is some time ahead of me, so many thanks to him for supplying me with particularly useful information regarding specific model filters for the SEAS drivers that Philbo was able to advise about. So, a great results, thanks to all. It has been fascinating to compare our methods and calculations. I have done quite a lot of comparative listening as well as checking and corroborating using REW measurements. It's possible to go a long way beyond generic solutions, accomodating personal preferences combined with speaker/room interaction. I've got too many graphs for comfort, but I will sort some out to add here, illustrating various differences related to group delay, spectral decay, spectogram, distortion, etc. One other difference between the new and standard Keltik filters is that there is an obvious gain applied to the standard filters, confirmed by Finster. I did wonder why, and apparently it was to bring them more in line with the original analogue filters. I will continue to make refinements and use Exakt Design for other home projects - already done one which works very well for a small ported speaker. Really looking forward to being able to use these filters with SO2. If anyone wants any further information or access to this data, please feel free to pm me.
  7. I quite agree that for individuals who are using Exakt Design filters, there should be a free exchange of information. However, for people like Sunbeamgls (Speakerfilters) who is building up a catalogue/portfolio for many different speaker combinations and is providing a service which requires time and effort and involves travelling costs, it seems appropriate to pay a fee to have access to these model specific filters, especially given that a trial period is provided. I am using Exakt Design only for my DIY projects (including modified Keltiks) so I don’t mind sharing information with likeminded individuals.
  8. Bit of a repeat of some of my previous posts, but here’s an summary/overview and explanation of SO, room modes, etc, - certainly no claim to be definitive: Room modal patterns of reinforcement and cancellation are created for all reflected soundwaves in an enclosed volume, across the entire frequency spectrum. It is akin to a unique fingerprint, peculiar to the precise geometry of the enclosed space. Modal frequencies transition in behaviour around the so-called Schroeder frequency, typically somewhere between 120 Hz and 200 Hz. Above the transition the behaviour of the modes is largely non-resonant, and their distribution is numerous and uniform. These modes do not generally adversely affect perceived sound quality. Conversely, below this transition region, the modes are less evenly spaced and readily discerned as resonant peaks and troughs. They are easily measured and heard. As an approximation they are often defined and calculated in terms of the order of reflections that can be defined in a cuboid space. The frequency position of the modes is an entirely dependent property of the geometry of the room and is unrelated to the speakers or speaker positions. However, the excitation of this underlying pattern of LF modes will be perceived differently from different positions and every individual “point” in space will have its own unique perspective with regard to which modes are excited and to what degree. The listening position is one such “point”. The extent of the excitation of this underlying “modal pattern” is dependent on a) the position of the LF source (primarily the bass drivers) and b) the amount of LF energy available to energise the modes. The directivity pattern of the LF source – whether it be controlled cardioid, omnidirectional full sphere, half sphere, or whatever, will influence the results (cardioid response helps remove room effects). Overall considerations are dependent not only on the position and output of the LF source, but also on the LF transmission/absorption properties of the structural boundaries of the enclosed space, which determines the amount of net LF energy available to rattle the cage. For lower frequencies, SO models this interactive behaviour and applies notch filters of suitable bandwidths and amplitudes to reduce the effect of constructive modal peaks (antinodes) below 80 Hz or so, as determined at the listening position. Destructive interference regions, as mentioned by others, cannot be addressed by applying this method in reverse. If more energy is pumped in to try and counteract a modal trough (node) it is simply cancelled by the same amount of energy reflecting back and nulling the input. It would be very wasteful of an amplifier’s power and provide no benefit to try and counteract it. SO2 uses a more advanced calculation method than SO1, encompassing a sophisticated boundary/mesh analysis technique rather than using standard equations for a regular rectangular cuboid, which just about every other calculation algorithm uses too. SO2 also applies some adjustment based on psychoacoustic preferences. Measurements and other DSP tools can be used to corroborate, validate and compare with SO. Lot’s of opinions and lot’s of choices. After the SO filters have been applied they can be tweaked. With SO1 this can be fairly gratuitous, enabling it to be used almost as a parametric equaliser if desired. I find the results of SO2 a marked improvement over SO1. Either way, the clever bit has all been done. All we have to do is make a few measurements, fanny about and then debate the consequences. Using DSP to convert minimum phase response loudspeakers to linear phase, thereby preventing phase from varying with frequency, is a separate topic. Overall Group delay is part of this phenomenon due to a combination of mechanical and electrical factors, including driver voice coil inductive reactance, the displacement of driver center positions relative to each other and the listening position (Time of Flight considerations). Conventional passive crossovers also contribute significantly to phase anomalies. TOF corrections for “measured” loudspeakers appears to be done in SO2 and/or Exakt. The latter adds further refinements to help accomplish linear phase and magnitude consistency.
  9. Problem is, extending the logic of less is more, the most clarity would be gained by disconnecting the bass drivers. I think it’s more a case that increasing LF bass extension (speaker dependent) will affect clarity levels and perceived transient response. Unless the swamping effects due to a corresponding increase in LF modal energy as a result of greater low level extension are negated, clarity will be impinged. This is what SO is good at counteracting and Optimisation Preference will vary accordingly. Prescriptive generalisations are not always applicable since every room-speaker interaction and grey matter filter are unique.
  10. Yes, I would expect it to be ok. Technically, a more difficult load, demanding more current to maintain a fixed voltage, relatively speaking, but with decent amplifiers and without a passive X-over in the way, should be plain sailing (but don’t quote me…). Besides, nominal impedance is just that, so really depends more on the overall impedance behaviour across the frequency range. Agreed, Oskar is very helpful and informative but probably not geared for regular high production batches. I’m sure he likes the results of these kind of threads.
  11. Yes, well spotted, my error. I never realised there were 4 & 8 ohm versions available.
  12. Certainly, there are various factors to consider and it is much easier to test and tweak using adjustable active filters. Especially so with digital filters, such as miniDSP or bespoke Exakt, for example. These systems allow pretty much any combination of X-over points, slopes, cuts and gains to be applied; almost too open but extremely versatile. Yet, I'd still prefer to get a good match on paper, before starting, probably +/- 1 dB as a guide, being careful to confirm that like is being compared with like as far as possible. Silk purses, and all that. I might be wrong, but I think the sensitivity you quote for SEAS W16NX001 is a bit optimistic.
  13. Steve, I think it was originally the Tonagen Electric Company? However, as you say, details are very scarce and I believe the company must now be defunct. One of the problems I originally had, trying to find a suitable high quality mid-driver with an appropriate specification for the Keltik, was finding a reference point for comparison, since no details for Tonagen were available. In the end, I used the Linn replacement Peerless (C14WG49-08) as a point of reference and went from there to decide on M15CH002 with Hiquphon OW2. I think one of the problems with the other SEAS units you mentioned as potential Kaber replacement drivers is that their sensitivities appear relatively low, certainly for the W15CH001. Good to see Finster back in the mix.
  14. Yes, agree with you there Steve, nothing really to loose by trying it as a single mid replacement in the Kaber if you can easily revert back. I'm sure more could also be done using bespoke Exakt filters to make suitable adjustments and compensations but this might be a bit OTT. I think you are correct about which driver is which in the Kaber, that would normally be preferred practice and should be easy enough to verify.
×
×
  • Create New...