What???
So called room correction software is not particularly challenging, you need software developers, not mathematicians, the algorithms and methods are available to programmers proficient with DSP programming and engineering… however equally important is the hardware to run it… that needs careful digital electronics engineering design so as to provide the optimum SQ performance, and the resources in the streamers using the Analog Devices SHARC processor are unlikely to be able to run it without an unacceptable impact to SQ if at all possible with the current Naim product architecture.
The DSP in the current Naim streamers is relatively very straight forward using a zero value over sampler, and a single IIR Butterworth low pass filter with added poles (steepness) to reduce time smearing in the audio band… it’s apparently only a few lines of assembler long, the SHARC processor does most of the heavy lifting, and those lines are carefully sequenced and timed so as to provide the best SQ performance in terms of minimum added aberrations from the filter processing,
(If not familiar conceptually think of a SHARC processor as equivalent to a purpose built DSP FPGA processor).
My view is that if Naim ever toy with more involved DSP such as EQ and Room / Speaker correction then they will most likely use a processor separate from the DAC, with a high sample rate link between the processor and DAC or streamer… perhaps call it Uniti Core DSP or something,… kind of not dis similar to the approach Roon usually take.
You made that specific suggestion a couple of weeks ago… What is it you know of the state of employment of mathematicians (specifically, ones from Oxbridge)?
Re DSP, I can understand if Naim’s view is that negative effects of DSP outweigh the benefits given that with due care in room setup you can usually fix most problems. But that is ignoring the fact that a significant proportion of customers or potential customers also want their rooms to be normal living rooms (or as near normal as semi-industrial looking massed ranks of black boxes can be!). On that basis, it would make every sense to offer it in their digital boxes, though that may mean waiting for other models with the added necessary processing capacity, or offer another black box. And the latter perhaps could also double as an active crossover to update the Snaxo range, making it feasible to activate a much wider range of speakers. Unless, of course, it would cost more than something like the Trinnov, and Naim judge the latter is as good as they could do and so not worth competing…
But remember, of course, DSP still cannot fix everything, and aside from the possibly relatively rare problem of cancellations, early reflections in particular may be worth fixing if your room design/layout create them.
@Innocent_Bystander, Of course, I was just joking about the the unemployment state of the OxBridge postgraduates. Otherwise, I would not get your attention.
@Simon-in-Suffolk, I think the Linn SO algorithm is much more sophisticated than what you think it is, although I do not know exactly how they do it. But by the look of it, there are quite a lot of variables, it even has some of the environmental variables such as room temperature, humidity, window, wall materials, etc. So it cannot be implemented by a couple of assembly lines, I would have thought.
I think Simon meant that the DSP already in Naim’s streamers is only a few lines of code, but to implement good room equalisation would be much more complex, and so could not suggest the added to the existing products!
Which is true and as far I know the crunching for SO2 is done in the cloud. It is much more sophisticated than simple room-correction code. SO2 and room-correction is not the same thing.
Only the result of this is run on the device and it has the potential to improve music listening to a certain level.
But there will always be boutique-audio which together with user-knowledge will be able to go beyond this.
Naims big problem is they seem stuck in a 1960’s model of a stereo where you have a gramophone, a tuner and a tape-deck and need a traditional preamp that can handle these plus a CD-player. Naim is right about the gain staging needed. They are wrong in their apparent belief that it can only be implemented on top of a traditional preamplifier.
No matter how sophisticated the processing it can’t get around the mathematical limitations of digital filters.
Using software to apply even simple room correction algorithms to the data in the file in non-realtime processing before they even start to be sent to the player, still cause a loss of quality in the signal, even in parts of the frequency spectrum that are supposedly not significantly affected by the filter(s).
The gains from any DSP processing have to exceed the negative effects of this mathematical degradation for it to be beneficial to the sound quality.
I don’t know the degree of ‘sophistication’ although I did discover two years ago it wasn’t quite as advanced as I thought when I was talking to Linn about it.
I was referring to the Naim designs with the SHARC assembler… Naim keep it simple by design so as to keep processing noise to a minimum.
Which bit don’t you understand? The DSP is performed by a SHARC DSP processor running DSP functions with Naim
Other manufacturer designs use FPGA DSP processors to do similar functions. Conceptually they do the similar things … sorry I don’t understand why you don’t understand that…
Now it’s my turn to say, what are you on about…
You have some rather strong assertions there… you might not agree with how Naim design and build their Hi-Fi… but good thing is there are plenty of other manufacturers out there … however can’t quite understand why you hang around a Naim social media site if you think their approach and ‘beliefs’ are wrong as well as their ‘big problem’ is that they are rooted to 1960 stereo systems…
Good… I might need to enquire into it again to see how they have improved their modelling and response systems… next time Linn have an open day at my dealers I might wander along.
@Simon-in-Suffolk, since you are the most open-minded, knowledgeable, techie guy here (AFAIK), it would be very interesting to hear your feedback on it.
In fact, it took me a very long time, and after several trials and errors before I fully accepted the Linn SO.
Ok let’s hold this over until then… I am not a great fan of the Linn sound so I might not be the best judge… but I am certainly interested to hear about what they do now.
Thanks for the compliment… I do think many of the ‘techie’ regulars here are equally open minded…
My first reaction to Urika II was why on earth would anybody want to digitise vinyl replay then I heard it. No fiddly air plugs, no external box, uses same quiet PS as LP12 and connection to DSM is a single Blue Jeans Cables Cat 5e. I can dial in the precise gain, impedance and capacitance and sound quality is wonderful and I get the benefits of Space Op V2.
I do think Linn has stolen a march on the traditionalists. I especially like that I no longer need a seperate pre-amp.
Indeed. In 2009, Linn were the first to discontinue CD players. Today, Naim have discontinued all but one CD player too. Making an excellent CD player is still an art, but a futile one now. Streamers deliver better results for less money.
In 2015, Linn discontinued their preamps. Their vision is to get the digital signal as close to the speakers as possible. Someday Naim might probably have to do the same.
To me the rather big flaw in Linn’s approach that may come back to haunt them is that they are so dependent now in their chain on the digital to audio reconstruction. Naim give you the option to swap out here if you want.
Digital to audio reconstruction is not exact, both technically as well as theoretically and can only ever be an approximation in practical terms. Therefore different designers, vendors, chipsets and manufacturers tend to produce ‘in house’ approaches and by definition which compromises to choose.
With Linn… it’s an all or nothing … there is no analogue audio preamp / controller… you need to buy into the Linn reconstruction compromises for their entire chain if I understand their products correctly.
With Naim I can swap out… and for me that is important… I personally don’t like the Naim chosen reconstruction compromises so I swap out and then back in within the Naim audio chain,
Digital to analogue reconstruction is relatively fast evolving area after having been fairly static for years… with the development of better FIR filters and windowing techniques with more powerful less current hungry cores, we can expect FIR filtering techniques and hardware architectures to evolve further over the next few years. This is as much to do, if not more, so with hardware as it is software.
So I can enjoy my Naim amplification, and digital transports, and swap out my DAC reconstruction as it improves and evolves as it certainly has done over recent years… now to me whether intentionally or not, Naim have got this balance right.
I very much agree with this being hardware driven.
The more capable hardware has permitted to more complex mathematical solution to be implemented practically, and this has lead to the ongoing surge in development. The mathematical principles have been well known for quite some time, it’s the practical application that hasn’t been realistically achievable until recently.
With a Linn DS/DSM you can also choose to use the SPDIF/TOSLINK outputs to attach your own DAC. Whether this makes sense or not is the question. You would need a volume control too, the digital output signal is line level. I believe you can still use the other digital features like SO over SPDIF/TOSLINK. I’ve never tried that, I like the Linn DACs.
Regarding Exakt, you are correct. This is a closed system, because AFAIK the Exakt Link between DSM and Exaktbox/Exakt speaker is not an open specification. It would be nice if other manufacturers could make their own Exakt DACs.
