Well, it is extremely important to respect the standards, whether they are. This is the only way to preserve the sound oddities caused by recording and cutting technology errors as well as the hearing impairments of the recording engineers, indeed...
The problem is the same with all the digital crossovers. As soon as you get to any high slopes or corrections of any kind..the stacking of the numerical transforms causes errors in the crossover regions which creates the equivalent of acoustic mud. You get the advantage of solving some loading problems and baffle lift, etc..but the rest is, in my experience--quite detrimental.
Some people listen for some aspects in the same way that some viewers of high end home theater look for different things. Some love the colors of LCD projectors, but I find them flat and involving. Usually, people with glasses and thus a reduced capacity for contrast realization... tend to think that LCD projectors look 'great'.
er...ok.
Same sort of thing applies in audio. those of lesser capacity will..make conclusions about the state of reality that is purely based on their individuation of self realization and capacity ...instead of what is actual and real for the rest. A marked inability to think past their individual ego. Basically, psychological issues that do not involve the reality tht is there for others. Ignorance at it's finest.
I want the whole package, and due to that desire..the digital crossovers I've tried so far are all 'unacceptable' in the final analysis and end up causing me to turn my audio system off..instead of desiring to turn it on.
Lamarkism, which is the idea of genetic change in the single given life--has been shown to be real..as well as the ability to change the mind's wiring in a single lifetime. If the latter was not true, we'd never change the slightest bit in a single given life. thus the idea of having enough 'oomph', or desire, or drive..becomes critical. thus, thus who work hard enough to push the limits..actually..will do so. (human genetics has changed about 7-8% in the past 4 thousand years-- so Darwin has been shown to be a piker and a near total loss, and so on....)
If one does not understand why I'm saying all this extra stuff...well: Essentially, someone who figures out that digital crossovers have serious problems, sonically, is someone who has worked hard to train their hearing system, over time. since the human hearing system is the most complex system in the human mind/brain.... then it is entirely possible for people t learn to discern how bad digital crossovers do indeed sound. The aspect of jitter combined with inaccurate digital numerical transforms does indeed add up, especially when it is two channels that must agree with one another to utter perfection --- in the micro sense of level and timing.
We hear this micro differential in timing and level issues,a s the human ear/brain is designed to and can resolve complex harmonics both spatially and level wise, to the tune of less than 100 thousandth of a second, within the analysis of one single tone. then add the rest into make up the harmonics, which we can also resolve..and then it shows that digital crossovers end up being a disaster as they get the gross aspects correct but screw up in the micro range. The micro range of differential..where our capacity to resolve fineness is extraordinarily acute.
I can explain why and how such things are true and real but we don't have time to write books, here. Just the basic synopses.
As usual, those who do understand what I'm saying will be silent as they know the result of 'sticking out' like that..one ends up being attacked by those who do not yet know these things -attacked by those that may live their entire lives ---and never explore them. It is a self awareness and physiological positioning issue. Logical deduction/reduction along the Socratic mean.
Thus, the analysis of the results of such musing by me and any others... brings the 'surface impression' that we are in the wrong..when the exact opposite is the real truth.
Getting back to the digital crossover and using it to 'plan out' a analog crossover. In my experience that does not work out so well.
Since the ear is so sensitive to timing and signal regarding tones AND harmonics, the noise brought on by using a digital crossover vs that of using an analog passive..the two have such different noise characteristics that the comparisons can only be gross and inaccurate when trying to use the digital one to simulate an analog passive for the sake of 'testing convenience'.
The problem is that when making subtle changes in passive crossovers..everything else is connected. so when you use the digital one to plan out the analog one, the analog one is wrong from the get go (due to the sonic differentials between the two, when listening) and it must be regauged to the point that perhaps you are better off starting with analog ---in the first place.
When digital crossovers appeared on the stage of audio....I eagerly adopted digital crossovers and began to experiment. I thought I was going to have to throw away 20+ years of learning the intricacies of passive crossover design..for digital was going to make all of that fade into the mists of time as 'wasted space'.
NOT.
Analog passive crossovers are still better, even with the problems they have. Note their continued use by the best loudspeaker manufacturers in the world. There is sound reason for this. Digital simply is not up to snuff..on all the important fronts. The reasons for this are fundamental, and in the micro spaces of the signal, whihc is where 'high end' audio is centered or listening closely for-- Not in the +90% of the signal we already have, but in the ~1-2% that we do not.
In conclusion, I don't expect digital crossovers to excel any time soon, due to their need to be in a world of complex stacked (digital) numerical transforms, all in critical regions, across multiple channels ---which must act in utter perfection with one another.
Their lack of phasing issues in the crossover range is only a useful aspect when used for making subtle changes on and in recordings, at the studio mixing level. That's it. That is their whole 'area of specialty' and it is a rather useless one regarding pure musical bliss and enjoyment, as they get all of the rest wrong, when it comes to comparison to the rest and the best that the world can offer.
The most basic demands for perfection placed upon them, when 'in situ' (in use) are far greater than demands placed upon the highest of the highest end of digital audio signals, regarding recording and playback.
Last edited:
Digital crossovers don't have to have high slopes. To me the phrase 'stacking of the numerical transforms' sounds like verbal mud
Care to elucidate? From the rest of your post (after filtering out more verbal mud) it seems you're saying you've tried digital XOs and they don't sound good. No dispute with that. So how do you generalise from experience with particular examples back to the theory - to my line of thinking there does seem to be a very important step missing?you could add noise models to the DSP code if want...
no reason except expense keeps you from using SOTA flagship DAC for each driver - if any digital audio source is considered acceptable then you just use 3x or however many required by your speakers
there is actually S/N and IMD improvement from using multiple DAC with restricted frequency content, a fraction of the full audio sent to each
today Gigaflop DSP is not a problem - that's thousands of terms per sample, per driver at 192 k
experience with decades old, or corner cutting cheap hardware isn't a valid basis for dissing the concept
Last edited:
jcx said:
Thanks for the input jcx, I understand your point. I only plan to get in the ball park with the digital crossover and then fine tune it. I know of someone who did a similar design using the DCX crossover first, and then analog, and it turn out extremely well.
It would be quite difficult to implement all the possible crossover options with an analog setup. The DCX makes it very easy to compare things ... going back and forth with changes.
The crossover from the tweeter to mid will likely be first order at 5000hz. I don't plan on using steep filters, as they aren't needed in this case. But, who knows what I'll eventually do. Woofers to mids are a different story though and need a steeper filter.
I'm sure it will all work out well in the end, unless I just really screwed up somewhere. I have yet to take any measurements or anything ... I'll find out.
John
The difficulty is not so much translation from DSP based testbed to analog circuitry, but rather that you can't rely on calculations or sims to getting an analog filter right. You just need to measure your way to the desired end result.
For the rest, in my opinion the main advantage of an analog filter over a DSP is that gain structure problems are much less.
vac
For the rest, in my opinion the main advantage of an analog filter over a DSP is that gain structure problems are much less.
vac
care to quote the next phase, same sentance?
practicing your Black Bag Rhetoric again? - too many hours watching political ads?
obviously anyone insisting that any form of digital audio is unacceptable is out of the "digital XO" debate by their own free choice before it gets started
practicing your Black Bag Rhetoric again? - too many hours watching political ads?
obviously anyone insisting that any form of digital audio is unacceptable is out of the "digital XO" debate by their own free choice before it gets started
Last edited:
Well, I am working on a system where there are three basic sources. One is an announcer, the second is from a single wireless lavalier microphone and the last is from digital sources such as CD's, .wav files, AES from video sources and the all too often MP3 from folks who don't know better.
The announcer microphone will go into a vacuum tube based compressor due to the potentially enormous dynamic range, then to an A/D AES out converter.
All sources go into a digital mixing console that uses an external clock. The AES output is distributed via fiber to the four main amplifier rooms and a half dozen other small systems.
Each main power amplifier takes copper AES in and only produces an analog drive to the power amplifier stages after doing any DSP required for it's own channel.
The one external clock drives everything. Although the amplifiers only can use their own clocks they do not have an input.
All AES signals are at 96k and "24" bit, but really only around 22 bits.
Now while all this seems to me to be a common sense approach. a significant number of the parts needed to make it work are not commercially available, so I will have to build them!
The announcer microphone will go into a vacuum tube based compressor due to the potentially enormous dynamic range, then to an A/D AES out converter.
All sources go into a digital mixing console that uses an external clock. The AES output is distributed via fiber to the four main amplifier rooms and a half dozen other small systems.
Each main power amplifier takes copper AES in and only produces an analog drive to the power amplifier stages after doing any DSP required for it's own channel.
The one external clock drives everything. Although the amplifiers only can use their own clocks they do not have an input.
All AES signals are at 96k and "24" bit, but really only around 22 bits.
Now while all this seems to me to be a common sense approach. a significant number of the parts needed to make it work are not commercially available, so I will have to build them!
I think analog is coming along leaps and bounds with the renewed interest in the format over the last few years. There's a lot of good T/T's right now and no shortage of equalizers, carts and vinyl. Seems there is a CD market "suck out" taking place. A bunch of folks are headed to all digital server based systems and HQ downloads and the other is re-discovering LP's etc. let's hope the vinyl catalog keeps pace and we get more releases. Very interesting time in high end audio right now.
Well, the CES and The Show had lots of preamps, however, I did not find anything any 'better' than the Blowtorch. It would seem that all my 'best' efforts are not made anymore. The Vendetta Research phono stage (that I use in my Blowtorch) is well described, subjectively, in the latest Feb. 'Stereophile'. It is still hanging on, subjectively with some reviews, even 20 years since it was last produced.
You're dis'n everyones efforts based on what? I hope you don't say your listening impressions at the show. Way to jump start the tread.
- Status
- Not open for further replies.