<font color='#000000'>Let me say that after having read the Sony 9000/4 comparison article, I need to be a bit critical of it on several fronts.
Not only the giant error on the weight of the two units, and the actual analysis and conclusion based on that mistaken information, but other grammatical errors in the article itself lead me to believe that this article was not properly proof-read and the technical analysis is suspect. If you have the two units in house, the way to compare the weights is to get a balance and actually weigh the two units and not rely on either the manuals or some figures you got from somewhere.
The impression I received as I read the article was one that the reviewers clearly favored the 9000 series and their technology and had a clear bias against the 4/7 series right from the start. Whether that bias is justified and correct is unknown, but it certainly taints the reasoning and conclusions in the article.
The technical analysis is totally unconvincing from an engineering standpoint.
Power Supplies: in the article the ratings are stated as 15,000 milli-Farads (that is what mF means) and in the discussion earlier, they are now rated at 15,000 micro-Farads (which would be 15 milli-Farads). Proofreading? Which is correct? In any case, the purpose of the capacitors, and the power supply itself, is to deliver stable, clean power to the rest of the circuitry. IF the power supply does that properly, then the nature of the components used to do this is of no concern. Using expensive brand name capacitors compared to less expensive brand capacitors does not mean the power is cleaner at the other end. After all, a capacitor is a very simple device - it stores charge and releases it. While the quality of a capacitors ingredients and construction may affect the temperature stability and current delivering ability of the unit, if the cheapest capacitor on the market will do the job, then you gain nothing by using a more expensive capacitor other than bragging that you have more expensive capacitors in your box. It was not clear from the article the capacitors, or the winding on the transformers made any difference - in fact, the review did indicate that the quality of the audio, under a silent signal, was just as expected -silent. No hum, hiss, etc. It would seem that the power supply and the electrical isolation of the 4ES does the job. The technology of the time of the 5ES or 9000 series design may have demanded more expensive components, or those components are just overkill.
On vibration - much was made about the lack of absorbing material on many of the parts in the 4ES. I'm not familiar with the intricacies of audiophile components, however, from an electrical standpoint, I fail to see HOW a vibration in a solid-state component could in any way, translate into an acoustical signal effect on the output side. Whether the case lid dings and rings or just goes thump when you flick it with your finger really would appear to be irrelevant in a unit that is a digital signal processor where such vibrations, in my mind, have no effect on the signal quality. While turntables and other sources that require mechanical interfaces to extract/deliver sound clearly can suffer from the transfer of vibration, in digital processing systems, the data is a pure digital data stream until the point that it is converted to an analog signal. This happens in pre-amp at the outputs or on the input side, you have some analog signal until it’s converted by an ADC into digital data for processing. Even in the analog form, I cannot see how the signal is affected by vibration. Are you telling me the movement of an electrical wire will affect the signal it is transmitting? I am skeptical but very much open to any engineering information to the contrary. I may be missing some vital knowledge here, so please educate me.
On to the processors: SHARC vs. RISC. RISC is indeed a processor with a limited number of machine instructions constrained by the fact that each instruction MUST be able to be executed within a defined limit - say one processor clock cycle. This means each instruction runs as quickly as is possible on the hardware level. Any higher-level instructions missing from a CISC processor can still be accomplished by executing several RISC instructions. RISC processing is more efficient processing at the cpu level - it has distinct advantages when properly applied to the correct problems. SHARC, as described is a specialized processor architecture for special processing applications. I would conclude that SHARC is even more specialized for digital audio signal processing than the more general purpose RISC processor. Now, is one preferable to the other? Not if they both do the job. If the necessary processing of the digital signal to perform the ‘massaging’ of the data stream can be accomplished in the time frame allowed by the RISC processor, then nothing can be gained by using a more dedicated SHARC processor. It is digital data and if you take 6 and multiply it by 7, it makes absolutely no difference if I do it on my old 386 or use IBM's DEEP BLUE supercomputer, the result is still 42 (although there was that one problem on the original Pentiums from Intel (CISC processors) where that may have actually resulted in 42.0000012345
Just kidding.) However, digital signal processing in these components IS a real-time effort. The data stream comes in one end and goes out the other end at a fixed clock frequency to feed the digital-to-analog converts. Any processing of the digital stream MUST be finished on any sample or sequence of samples of the digital audio data stream within the time allotted to it. This means that a more specialized processor certainly could, and I would expect, SHOULD be able perform more processing on a digital data stream within a given time window than a general RISC processor. However, the question is - is it needed. The only way to answer if the SHARC vs. RISC solutions from Sony are different in terms of quality is to look at the source code for the algorithms that do the processing on the SHARC and RISC units in conjunction with those processor specifications. IF the RISC solution uses some short cuts or does not process the signal to the same 'resolution' as the SHARC processor, then there could be some loss in processing ability of the RISC vs. the SHARC engines. However, I'd imagine such detailed analysis was not performed (since apparently the units could not even be disassembled enough to identify the processor on the bottom of the board)- nor would I expect it to be performed. But to suggest that on the surface, the SHARC system is better than the RISC system is misleading. Any differences in the acoustic output may well come from any other number of subsystems in the unit other than the processors. Without a detailed analysis, it is just not possible to make that call. Saying that SHARC systems are better than RISC systems is not sound on an engineering basis -they may well both be performing the same data manipulation?
About 7.1- First off, let me say that every single channel is MONO (we all know that, right?). Each speaker is a 'mono channel speaker'. The 4ES's 7.1 processing provides a left and a right rear channel speaker output, where as 6.1 provides only a single rear channel output. The review states that you only get the same signal sent to the two speakers, so you are not getting discrete signal on the two rear channels. The question I've had for a while and that is not answered by the review is, what is the capacity of the system in terms of separate right and left rear channels. If fed with a 6.1 channel signal that only has one rear channel, then, of course, that same signal is sent to both rear channels on a 7.1 setup. However, is the 4ES capable of sending two different signals to the two rear channels or is the hardware wired such that you will always only get the same signal to both the rear channel speakers? IF the digital DTS audio stream comes in as a 7.1 encoded signal, carrying 2 separate rear channels, will the 4ES reproduce them correctly? After all, DTS does define a 7.1 signal (http://www.dvddemystified.com/dvdfaq.html#3.6.2). Just because no current DVD media includes it does not mean it will not show up. Of course, just because we may foolishly buy a 7.1 capable system today, does not mean we will ever be able to use that feature. There may never be DTS 7.1 media with two different rear channels. However the question remains, IF the DTS signal contains 7.1 info - that is two different rear channel signals - then will the 4ES decode them properly and thus provide two different signals on the rear outputs? This is a 'bet' people have to decide for themselves if they want to make it. Since I only buy a new unit every 10 years or so, if there has been a technological advance this is a decision I will need to make in the near future.
Connectors - I would imagine that most users would set up their systems and pretty much leave them alone. Optical digital connections are in no way inferior to coax digital. Also, the 9000ES does not provide any Component Video support (and so the lack of on screen display over component video is really a mute point). Component video switching (no processing here) may be of interest to those who will be using a component video out from their dvd players as well as from a dish receiver or HDTV tuner in the future.
About the audio quality perceived by the reviewers, I wonder how much of that difference may be related to feeling comfortable with the established system they have had for a while. It is clear that the reviewers felt there was a difference in the two systems output. Which is better, that’s a different question. If someone has listened to the sound from a 9000ES system for many years that can become familiar and any new feel or change to the sound can be considered substandard. However, different does not equate to wrong. Each person has their own tastes and as the reviewers rightly pointed out, each home setup will be different and so it is indeed a personal choice as to which system one feels happy with. Auditioning a system at HOME in the proper room where it will be used is the ideal solution - but we cannot always do that. To some, the 'correct' way to listen to something is as the original composer/mixer/studio/director/artist intended it to be. If I can get the system to put out the same signals the studios put into the source, I'm happy. Others want to tune the studio source to their specific listening preferences. So long as YOU, the listener are happy, that is all that counts.
In all, the review seemed biased from the beginning and as it approached each step of the analysis and the explanations themselves were unconvincing on a technical level. All in all I think I ended up with MORE questions in my mind than less after reading the review. I would say it was not very helpful and only serves to make owners or buyers of the 9000ES feel good. Whether that result is justified is not answered.
Finally let me say that I may have made mistakes in my assesments in this post, so if anyone finds an error, please do post about it.
For what its worth, to those who might care to know: I hold a Bachelors degree in Computer Engineering, a Masters in Electrical Engineering a PhD in Computer Science and am an licensed Professional Engineer in the field of Electrical Engineering in the United States. I am not an audiophile nor do I dig deeply into the technical details of audio or video equipment until such time as I need to make a purchasing decision - such as in the near future. I do, however, like to make very informed purchases and purchase the right equipment for the money I spend. I will buy a $50 Timex or $200 Casio watch with the latest technological features and will never spend several thousands of dollars on a Rolex or similar watch - which I'd feel I'd have to use the rest of my life. This way I can take advantage of the latest that technology has to offer.</font>