The Wandering Image

22 Feb

One of the issues that thrown against matrix systems was that you had to sit in the ‘Sweet Spot’ and when you turned your head the image moved. This unfortunately was more relevant to SQ than the other systems based on ‘RM’, but it did still happen.

This was one of the areas i looked into while developing ‘Phoenix’ and i feel i’ve mostly eliminated the problem. Whilst listening to a ‘Pheonix’ decode looking forwards then turning your head either left or right by 90 degrees the image should still sound as if it’s coming from the same direction as before.

Of course, if your system has speakers that are not all the same or your listening room is an odd shape, etc, the effect may be diluted, but do give it a try, and post your views, positive, negative or indifferent, i’m interested on how well it’s heard out there in the real world.

Try it on a ‘Sphinx’ decode too…..


23 Responses to “The Wandering Image”

  1. crispin14 February 24, 2015 at 19:52 #

    First, OD, I’m intrigued by your reference, “Try it on a ‘Sphinx’ decode too…..” What is that?

    Anyway back to the question, I just compared the first couple of minutes of the recent “Music for Trumpet and Orchestra” in both the older and new Phoenix version. The Phoenix image is spectacularly more focused than the older decode. I don’t have a lot of problem keeping track of instrument placement when I turn my head on either version, but certainly everything is more stable in the new Phoenix version.

    It might be fun if you could put together a “test file” of a few fragments of recent Phoenix releases offering ABAB comparisons without one having to go through the rigamarole of switching disks or files. That of course would just be for fun.

    Over the years a bigger problem that I’ve tried to solve both with stereo and with quad (and modern cinema surround) is getting good sound to 3 or 4 people sitting on a couch. The usual complaint is that the person on the right says she only hears the right speaker.

    Surround sound, and your decodes in particular have mostly alleviated this effect, because a person the the right is more or less served by the complete sound field.

    I’m using 4 Vintage Dhalquist DQ-10 speakers and 3 subwoofers. The two rear DQ10s are connected passively, each to an Audio Concepts subwoofer I assembled: each has 2 8″ drivers mounted on a single angled panel facing in, so they work as an expanding and contracting sphere in the reflex cabinet. The Front soundstage is supported by a SpeakerLab sub with 2 front facing 12″ woofers. All the woofer units are mounted on points to isolate them from the room and limit my exposure to neighbour complaints. The front subs are powered by a 1000 Watt Dayton Subwoofer Amplifier.

    I rely on a Yamaha’s (RX A-3000 YPAQ system to set the timing on the speakers.

    The one important discovery I made was that the “sofa balance problems” could be solved by toeing in both the front and rear speakers so that the left speaker points directly at the person sitting on the far right and visa-versa. The DQ 10 speakers are pretty directional, so there may be a minor alteration in the frequency curve for the person sitting closest to the left speaker, but this is far superior to his not hearing the right speaker.

    OD, your classical release give me a good deal of pleasure, and I download and contribute even if the music is old for my ears (pieces I enjoyed when I was young and now find a bit too familiar). These recording do take on a new life in clearly imaged surround sound.


    • oxforddickie February 24, 2015 at 21:18 #

      Thanks for that feedback, and by the way ‘Sphinx’ is QS.

      I guess the problem why there’s not much drift between the two versions is because SQ/II had some correction already built in to help minimize it, i suppose a comparison between a hardware decode and ‘Phoenix’ would show it better.

      You’ve got a powerful system there, do you invite the neighbours in when you have a party?

      Speakers and their interaction with the room can make or break many aspects of ‘Hi-Fi’ and slight tweaks like the one you described can often pay large dividends.

      I’m now of the belief that classical music really does need to be heard in surround, that’s how it’s heard when you go to a concert. The acoustics of the room plays an important part of the experience, so it makes sense to try and reproduce that at home. And i’m sure i once read that quad was first promoted for listening to classical music at home.

      • crispin14 February 25, 2015 at 01:20 #

        My place is too small and has speakers spread all over, so it is not a good venue for parties.

        I do live in a condo and although when I first moved in I did some work sealing electrical boxes and putting down some heavy carpet underlay and a in a few vulnerable places sheets of lead on the floor (no one is above me). I also removed one wall that has the cable and heating coming up through holes and sealed off the openings, then put in double plaster board, because that is where most the the sound was leading.

        Putting speakers on points makes a huge difference being a good neighbour. These isolation points may be the most important thing one can do to improve sound quality and imaging. I was astonished by the improvement.

        My invention here was to get a piece of 3/4 inch board into which I screwed the points (making that a bit of a speaker stand). Subwoofers sit on those boards. Since the points would have pierced the carpets and underlay, I got some Chinese restaurant sauce holders (the small cups they pour soy or sweet and sour sauce into). I place these on the carpet and then let the points rest on cup. The cups are inverted to spread any vibration over a larger area of carpet.

        Finally, as most of us know, Quad provides about a 4db increase in “apparent” sound pressure compared to stereo. Thus when the sound in quad is quite loud and you walk into another room it does not sound nearly as loud as the same apparent volume would in stereo.

        Why would ANYONE ever not want as much quad or surround as possible?

  2. prefabfan February 22, 2015 at 19:04 #

    Inthe 80s a radio show by an audio engineer said it is impossible to get 4 distict channels from a matrix quad. So you’ve already accomplished the imposdible.if you nevermake any other contributions to music listening you’ve done enough for any one person. You have done audio alchemy.. If you get a solid SQ center image you have done the impossible AGAIN,

    and that thing people pop open at the front of the car is called the “HOOD”!

    • oxforddickie February 22, 2015 at 20:21 #

      I’m rather pleased with Phoenix’s imagining properties. It’s the first time i can sit and listen to an SQ decode without thinking there’s something wrong.

    • sailorpaul February 23, 2015 at 04:41 #

      For the record, I think the audio engineer was correct in that you cant get 4 distinct channels from a matrix system (discreet is a better word). I don’t think OD’s process does this either. If you listen to the rears alone there is always some degree of bleed through from the fronts. Its just that the bleed is so far down in level it is difficult to discern. I don’t think he has done the impossible as much as he has taken what is possible to its extreme. Now if he were to apply his knowledge to the original encoders….

      • oxforddickie February 23, 2015 at 05:53 #

        I totally agree, it is impossible to get truly discrete from any matrix system, like it’s impossible to get truly discrete reproduction from CD-4 or even FM stereo.

        If my memory serves me right, it was once stated that obtaining a separation of 20db or more was considered the minimum required to satisfy the term ‘discrete’ audio reproduction. I think JVC used that as proof that their CD-4 system was deemed to be considered as discrete. So when it comes to the phase matrix quadraphonic systems, i believe that the improvements i’ve implemented easily exceed that requirement, but i will not say that the results are discrete.

        What i suspect he means is that i’ve taken SQ decoding to a level that has been impossible until now. SQ is a complex matrix which causes problems when it comes to decoding it. Unlike the ‘RM’ family (QM/QX/QS/BMX/Matrix H/HJ) it was designed to have full stereo width when played back in stereo. That means the front channels are not encoded in any way, with just the rears being encoded, which has to be more complex than any other matrix.

        When it comes to decoding back to four channels, the decoder is only designed to process the rears, the fronts are left completely alone. What this means is that when played back the fronts still contain all of the rear channel information. I’m sure you can see that this causes many problems like blurring of images, reduced separation etc.

        Phoenix is the first ever attempt to remove this information, and within the limitations of today’s available technology to me, i believe it has taken SQ decoding further than previously thought possible, with the obvious audible improvements.

        Personally i don’t like the word discrete, i’ve never tried to create a discrete decode process, even though the separation is well over 30db, the main objective has been to make listening to an SQ decode as pleasant an experience as possible.

        A little note about the encoders. The problem isn’t there, there’s nothing that could have been done to improve matters once the ‘London Box’ was brought into use.

        Again, sorry for the ramble….

  3. circularvibes February 22, 2015 at 18:48 #

    I look forward to checking out this effect. I have a question though, what is a “Skippy box”?

    • oxforddickie February 22, 2015 at 20:17 #

      The reason i call it that (and there are a few who know what it means) is because i refuse to discuss it.

    • Owen Smith February 22, 2015 at 20:28 #

      It’s the Surround Master by Involve Audio. An Australian company hence “Skippy” box. Search QQ if you want to know more about it, it is OD’s wishes that it is not discussed here.

      Note to OD: feel free to edit or delete this post if you wish, but I think people deserve to know what we’re talking about. People have the right to make their own minds up.

      • oxforddickie February 22, 2015 at 20:35 #

        I must state that it is advertised as being able to decode everything and can’t. Of course it is entirely up to the individual whether they believe the advertising hype, but i suggest you find out the people that have complained about the issue of instrumentation appearing in places where they don’t belong.

        As Owen say’s i will not allow any discussion of this unit here, it’s NOT capable of decoding SQ or DY/EV-4 or (and this is the major joke) UHJ/Ambisonics.

        • circularvibes February 22, 2015 at 23:05 #

          OK, thanks OD and Owen. I had never heard the term and won’t mention it again.

  4. Beechwoods (@Beechwoods_) February 22, 2015 at 18:10 #

    Imaging for me is rock solid. I’m listening to the Phoenix decode of ‘Exposed’ at the moment and I often have to move around the image is completely consistent. I have slightly larger speakers at the front (Rogers JR150’s) but quite well matched rears (largish Gale bookshelves) but they match tonally and I have them carefully level balanced for my usual listening position. I’ll have to compare with the effects on the older decodes, but maybe I’ve been spoilt… 😀

  5. bayards February 22, 2015 at 15:46 #

    In the day I had CD-4, SQ and QS. The latter within an all-in-1 Sansui system. I preferred QS to SQ even though the latter had more releases. I was fairly happy but CD-4 led the way – even forking out for a Shibata stylus.
    The Phoenix releases are a sheer revelation. Excellent separation and no “wandering” as far as I detect. I have big floor standers for the fronts and more modest floor standers for the rears. Listening to the opening track of the Chase album is wonderful.
    If only all that matrix info had been extractable back then…

  6. Owen Smith February 22, 2015 at 12:23 #

    I have to say I’ve never found this “wandering image” problem. I often turn my head to localise sounds better, apparently this is natural for humans. Even with SQ*Final I was never aware of the image turning with me. Of course I have very different speakers front (floorstanding Castle Harlechs) and rear (Castle Richmond 3 on wall brackets firing up at ceiling) so maybe the different tonal characteristics of the speakers help to keep the image in place for me.

    • oxforddickie February 22, 2015 at 12:43 #

      There were small level of correction in the earlier process’s but as ‘Phoenix’ is to be the last process i do for SQ i figured i’d nail it on the head.

      The problem is caused by the high level of unwanted phase information left over once decoded, and SQ leaves a ton of the stuff behind, and it’s this high level of out of phase information that causes some people to not to be able to tolerate listening to an SQ decode, normally just saying “i don’t like SQ”.

      All hardware decoders ,including the Skippy box, are incapable of dealing with this issue, all they do is pump the wanted and wanted information up and down, causing further blurriness.

    • Owen Smith February 22, 2015 at 12:55 #

      My Castle Harlechs are unusual speakers. As well as the normal bass and treble drivers firing forwards, they have another bass driver in the top that fires straight up at the ceiling. The shape of my room dictates that my front pair are a long way apart, too far really. Every other speaker I tried had a hole in the middle of the stereo image, it was only the Harlechs that managed to bridge this gap. I’ve never been sure if it was because they’re better speakers (they were certainly the most expensive I tried) or because the upwards drivers plug the gap. They also have two quarter wave folded horns (one for each bass driver) of different lengths for bass re-enforcement instead of ports, which really works.

      I’ve never been sure what the forwards and upwards firing bass drivers are doing for phase issues even in stereo. But the results are great and stereo imaging is superb. Combine that with my purely upwards firing rear speakers (I don’t have any space behind my sofa and chais) and I have a poor man’s version of Dolby Atmos nearly!

      And the ceiling is artexed. I’ve often wondered if the reflections from a flat ceiling would sound different, but getting rid of artex is such a pain.

      • sailorpaul February 23, 2015 at 04:45 #

        Can you elaborate for the ignorant among us on this side of the pond exactly what artex is?

        • oxforddickie February 23, 2015 at 06:02 #

          Oh dear Owen, what have you started… LOL

        • sacdtodvda February 23, 2015 at 08:51 #

          Think of it as lumpy plaster spread on top of your dry wall.

          Except that pre-1990 it had asbestos in it.

        • Owen Smith February 23, 2015 at 09:33 #

          My artex dates from 1987 when the house was built. Regulations in the UK at the time mean it shouldn’t have any asbestos in it, but the concern that it might is one of the issues with getting rid of it. Just painting it is so much safer.

          Luckily I only have artex on the ceilings. No wood chip wallpaper in sight (that other bane of UK housing from the 1980s).

        • Owen Smith February 23, 2015 at 09:34 #

          Links aren’t normally allowed on here, but maybe a link to artex is OK? The picture on this sight looks EXACTLY like my artex:

        • Owen Smith February 23, 2015 at 09:36 #

          Each of the circular swirls is about 8 inches in diameter on my ceiling, should have said since there’s no scale on the picture.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: