Measurement Initiative - latest news Aug 2010

NEWS AUGUST 2010 - PRESS RELEASE

The following is a general press release from the Knowledge Alliance.

Last Year, the UK and RMAF shows saw the first public presentations of a new, measurement based approach to the performance of audio systems. A three-way, co-operative venture between defence contractor Acuity and audio companies Nordost and Vertex AQ, this initiative broke new ground, both in terms of its shared approach to the problem and also the results it generated. Applying sophisticated signal analysis algorithms in a conceptually simple sum and difference approach allowed us to show the impact of changes in the replay system on actual musical program material (as opposed to test tones). Perhaps most interesting of all, was the close correlation between changes in measured performance and the audible effects observed.

Now, a year on, we have been able to both verify and refine the approach, while also extending the testing protocol to embrace a complete system (as opposed to a CD player alone). Changes in the algorithms employed and a revised approach to error assessment have increased the accuracy and consistency of results, while adding extra elements to the system under test has further underlined the importance of the differences being revealed.

But perhaps most interesting and exiting of all is the way in which this has revealed the likely future path for this research - its emergence as a product lodged well and truly in the public as opposed to the solely product development domain.

We will be presenting these latest developments at special sessions in the Hockenheim room at Whittlebury Hall (5pm on Saturday 25th and10.30am on Sunday 26th) and the Humbolt Peak room at RMAF (9.30am and 5.30pm, Saturday 16th and 9.30am Sunday 10th). Seating will be limited so you are advised to get there early.

NEWS MARCH 2010

Over the last few weeks there have been 2 important meetings which are critical for the next year's measurement research.

On 25 February 2010 the board members of Vertex AQ and Nordost met at the Bristol show (over lunch of course!). At this meeting the strategic direction for the next phase of the study was agreed. Most importantly though, the funding was formally allocated which will secure the involvement of Acuity Products for another year. On 8 March 2010 Vertex AQ and Acuity Products held a planning meeting at Llandrindod Wells, Wales. This planning session reviewed the status of the progress thus far and then set out the main objectives of the next year's work.

In summary, and still subject to change until the detailed programme is agreed, the work will cover the following:

1. Continued development of the mathematics and taxonomy for the analysis and quantification of results in the next phase.

2. Expansion of the analysis framework from the measurement of a single box, to the measurement of 2 boxes (CD and amp), initially without loudspeakers, but then with.

3. Application of the expanded framework to study further the effects of systematic microphony and microphony reduction techniques.

4. Application of the expanded framework to study further the effects of differing cable impedance characteristics in a system network context.

5. Application of the expanded framework to study further the effects of systematic EMI/RFI and passive EMI/RFI reduction techniques.

6. Application of the expanded framework to study further the effects of active EMI/RFI reduction techniques.

7. Substitution of different technology solutions for the electronic components under test, particularly amplification, where valve, linear solid state and switching amps will be in included. With these changes it is expected that different system characteristics for items 3 to 6 will be seen.

Further updates will now be posted here on a regular basis. Visitors to the Park Inn show in London on 27/28 March 2010 will be able to get more background on the current progress (see the show update item on the right for room details etc).

BACKGROUND BRIEF AS AT SEPTEMBER 2009

Introduction

For many years now, we at Vertex have been busying away developing and applying technologies borne out of the defence world. Our acoustic absorption techniques and EMI/RFI filtering processes have produced a range of products that we are very proud of. However, more importantly than the individual products, we have also developed a very comprehensive strategy, or methodology if you like, that deals with the application of Vertex products in an intelligent, system-wide way. We call it the Systematic Approach.

But as we developed our products and strategies, we became increasingly aware that our, or indeed, anybody else's inability to measure the dramatic improvements we were delivering was becoming a serious problem. This deficiency leads to all sorts of difficulties of course, from the obvious ones of 'snake oil' accusations, long up-hill battles to convince those who want to believe, nervousness from the press and new dealers and so on. At a deeper level, the ability to improve our understanding for further research and product development could clearly bring massive benefits. And finally, of course, there's the question of why, oh why, do we all seem stuck in a massive void between the current techniques being used to measure our pieces of hi-fi, and the reality of what we hear?

An Alternative Battle Plan - Turning to Defence Again

Now as we have said, Vertex has a strong link with the defence industry - we have a sister company which works on many UK MoD-based research and acquisition projects. About 18 months ago we were fortunate enough to work on a collaborative project of significant technical difficulty, and one of the other participants was a company of signal processing and analysis experts called Acuity. These guys were working in the core of some of the most complex 'number crunching' projects going!

So we approached them and described our dilemma with Vertex AQ and the hi-fi business, and asked them if they could help shed some new light on the matter. They agreed to take a look at least (but were probably thinking we were nuts), so we sent them some gear. We described in detail what we thought was happening to the output of something like a CD player when you put a Kinabalu under it say, and left them to it for a few weeks.

Well imagine our surprise and delight when a few weeks later we had a call from Acuity saying 'yep, we had to use some lateral thinking based on sonar algorithm processes, but we can clearly see there are time domain changes to the output of the CD player with and without the Vertex gear'.

Time for Collaboration

Another important element to this story was our work with Roy Gregory, then editor of Hi>Fi+. At the Heathrow and Rocky Mountain shows last September/October (2008), we supported Roy in some very successful demonstrations using Vertex and other good system infrastructure equipment. And during that time, conversation often came back round to the thorny issue of measurement. Well, subsequently, Roy moved on to become VP Marketing for Nordost, but because we had now begun to build up a good joint understanding of the many systematic issues in hi-fi, there suddenly seemed to be an opportunity of ironic synergy. Two 'cable' companies, supposedly competitors, working together on a joint research project - never been heard of before, we thought, so lets do it.

So, humor aside, we now have a formal 3-way collaboration between Vertex AQ, Nordost and Acuity. We have a robust project management structure agreed between the parties and we are rigorously exploring the issues of hi-fi performance and measurement. But we should be clear here, whilst Vertex and Nordost are directing the programme, the testing and analysis work is the domain of Acuity. Indeed, at a technical level we have set up a clear air gap between the 2 hi-fi companies and Acuity, plus the Acuity staff are applying the highest level of scrutiny to the work, just as they would if they were delivering information to the Uk MoD.

Some Early Output

One of the early tasks for Acuity was to try and see what real signal output differences might be evident using a reputable mid-priced CD player with some Vertex and Nordost equipment going in and out of the equation. We knew Acuity had identified some difference already, but could we start to gain anything meaningful from the information? Well its early days yet but it seems we can.

One of the tests uses a piece of real music, rather than sine wave test signals - which already we have found do not load up the system sufficiently to truly reveal the effects that are occurring. So acuity played a sample of complex music through the player, first with it just sat on its own on the bench with standard leads, and then with the player on a Super Kinabalu and wired up with a Roraima mains lead and a Solfonn interconnect.

The graphs shown below are only a brief example of all the graphs that have already been produced. The full case studies, with all the exact detail from each of the major measurement activities, will be displayed on the Acuity website very soon. But what these graphs do show is the reduction in the amount of error between a 'Vertexed' and 'non-Vertexed' player.

This first graph is showing you a time domain representation of a real piece of music - this is the waveform as is reconstructed by the CD player from each of the digital samples, and turned into an analogue output. It is obviously only one channel, and although the scale is difficult to see, the signal is going from typically plus and minus 40 dB. Therefore the CD player's output level at this point in the piece of music is about half the player's maximum output capability. When the testing is carried out, with and without the different bits of supporting equipment, it is nearly impossible to see any differences directly, therefore there is not much point in reproducing them all here. However, when you subtract the graphs from each other, you can then clearly see the resultant difference.

In this second graph we have reproduced the difference, or error graph, between the original track and the output of our CD player when it is just sat on the test bench and connected with ordinary leads. Again the scale is not easy to see but the amplitude of this error is actually about 8 percent of the amplitude of the music signal (the scale of the error graphs has been expanded so that you can see the graph easily).

Finally in this last graph, the player-with-Vertex graph was subtracted from the original music signal graph. The scale of this graph is the same as the previous one. You can clearly see that the effect of the Vertex equipment has been to significantly reduce the error output. The error has dropped by at least 50 percent, and is probably now only about 3-4 percent of the original music signal.

We are fully aware that there are still a multitude of questions to be answered here, we have only just begun this journey. But please rest assured that these first results are meaningful - in simple terms, the Vertex equipment has reduced the error in the CD player. What we are embarking on now is a much deeper analysis of what constitutes these errors and how they relate back to the musical signal being played. We are also starting to broaden the players we are looking at, and will be progressing later to other electronics too.

The next part of our plan is to release, and demonstrate, a lot more of our progress at the UK September Show at Whittlebury Hall, and at the Rocky Mountain Audio Fest in Denver, over the first weekend in October. We will be providing more on those plans soon in the show news section.