Re: Other options
Anyway I have an XM44 (in progress) and according to Phil Marchand it is possible to implement a delay with the frequency boards (of course it will always be less flexible than going digital).
Btw. Nobody familiar with that DEQX stuff??
http://www.clarityeq.com/
http://www.e-speakers.com/products/deqx-new.html
Why the grin?Ken L said:
Anyway I have an XM44 (in progress) and according to Phil Marchand it is possible to implement a delay with the frequency boards (of course it will always be less flexible than going digital).
Btw. Nobody familiar with that DEQX stuff??
http://www.clarityeq.com/
http://www.e-speakers.com/products/deqx-new.html
Oh yes, I own a DEQX unit for about 8 months now. I was using an analog tube based crossover before that. The DEQX simply sounds better. Hard to imagine for all the analog freaks out here, but even after AD/DSP/DA conversion with a lot of opamps in the signal chain it does sound better than my tube based crossover ,wich sounded a lot better than my previous opamp based crossover, a Thel diy 24db/oct crossover.
Talking about bad solutions this is really the worst crossover mony can buy, trust me on that one. If you are interested in a completely flat and dull soundstage, buy one
The DEQX is a very neutral piece of equipment, any losses will be compensated with the extra benefit of the DSP power.
Don't be afraid of digital solutions, hear it before you make any judgements on them. Every time I hear the same arguments about all the extra electronics, the extra ad/da conversion etc etc, but they all forget the benefits wich are much much bigger. Again, Listen to it yourself before coming to any conclusions otherwise I won't take your answer serious.
I have never owned a behringer but I cannot imagine that it would sound worse than the Thel crossover wich cost me about the same with all the so called High End options.
People don't trust the Behringer because it is cheap. I can tell you a litle nice story on that one. When I was trying to sell my previous diy speakers I had them advertised for 500 euro (about 1/3 that they had actually cost me) Nobody was interested, some people cam over and listened, found them sounding good but did not trust it because it was so cheap. A couple of months later I had them advertised for 2000 euro, just for fun to see what happened. and I sold them within two days, there was even a bidding competition going on. People are sometimes really really stupid.
Talking about bad solutions this is really the worst crossover mony can buy, trust me on that one. If you are interested in a completely flat and dull soundstage, buy one
The DEQX is a very neutral piece of equipment, any losses will be compensated with the extra benefit of the DSP power.
Don't be afraid of digital solutions, hear it before you make any judgements on them. Every time I hear the same arguments about all the extra electronics, the extra ad/da conversion etc etc, but they all forget the benefits wich are much much bigger. Again, Listen to it yourself before coming to any conclusions otherwise I won't take your answer serious.
I have never owned a behringer but I cannot imagine that it would sound worse than the Thel crossover wich cost me about the same with all the so called High End options.
People don't trust the Behringer because it is cheap. I can tell you a litle nice story on that one. When I was trying to sell my previous diy speakers I had them advertised for 500 euro (about 1/3 that they had actually cost me) Nobody was interested, some people cam over and listened, found them sounding good but did not trust it because it was so cheap. A couple of months later I had them advertised for 2000 euro, just for fun to see what happened. and I sold them within two days, there was even a bidding competition going on. People are sometimes really really stupid.
mhelin said:
Delay was the reason I went with the DCX2496.
I am running Azurahorns that sit on top of NHT1259 subs. The arrangement physically places the Azurahorn drivers 17" (450mm) or so to the rear of the woofers.
In my configuration, setting the delay correctly was essential to integration.
Regards
Ken L
Sticking a slow sampling “cheap” AD converter a processor and DAC after your CD, DAC and preamp will not get you into realm of high-resolution music. The more stuff in the signal path the worse the sound is going to get, it is like noise. The more amp stages you have the worse the noise.
When you put another ADC in the signal path, you have to add an anti-aliasing filter, which reduces the bandwidth and increases the phase distortions and add more delays and signal error. Now you have to process this signal with a DSP more disproportional delays and phase errors will be added. Now on top of these problems you still most convert the signal back to the analog domain and add more phase and time errors.
I can understand that people might need such a device to equalize a poor room acoustics, however room treatment are a much better solution. The Behringer is more of low end sound reinforcement and recoding product for garage bands device than it is a stereo product. Yes, it might be fun to play with and if you have a home recoding studio it could be fun.
The fact is that a quality passive crossover with biamping is a better solution for sound quality. The problem is crossovers in most speakers’ crossover are built with cheap components, even in some expensive systems. The fact is, if your goal is higher quality sound replace the cheap caps in your speakers with some high quality polypropylenes foil capacitors and good inductors, this will make a bigger difference than a Behringer ever could for sound quality. Just try it. On the other hand, if you’re looking for a gismo to play with this may do the trick.
When you put another ADC in the signal path, you have to add an anti-aliasing filter, which reduces the bandwidth and increases the phase distortions and add more delays and signal error. Now you have to process this signal with a DSP more disproportional delays and phase errors will be added. Now on top of these problems you still most convert the signal back to the analog domain and add more phase and time errors.
I can understand that people might need such a device to equalize a poor room acoustics, however room treatment are a much better solution. The Behringer is more of low end sound reinforcement and recoding product for garage bands device than it is a stereo product. Yes, it might be fun to play with and if you have a home recoding studio it could be fun.
The fact is that a quality passive crossover with biamping is a better solution for sound quality. The problem is crossovers in most speakers’ crossover are built with cheap components, even in some expensive systems. The fact is, if your goal is higher quality sound replace the cheap caps in your speakers with some high quality polypropylenes foil capacitors and good inductors, this will make a bigger difference than a Behringer ever could for sound quality. Just try it. On the other hand, if you’re looking for a gismo to play with this may do the trick.
oehlrich -
thanks for your detailed post. it's great to see the logical order and thorough evaluation you went through.
jewilson -
have you heard anything with a digital crossover? if so i'd be curious to what you listened to. How is a 24 bit, 96 kHz dac slow sampling? there seems to be a lot of people out here that would dispute your "fact" that passive is better.
- Robert
thanks for your detailed post. it's great to see the logical order and thorough evaluation you went through.
jewilson -
have you heard anything with a digital crossover? if so i'd be curious to what you listened to. How is a 24 bit, 96 kHz dac slow sampling? there seems to be a lot of people out here that would dispute your "fact" that passive is better.
- Robert
dezzz said:
Her er et godt link/ Here is a good link:
in danish
http://www.lcaudio.dk/filters.htm
something in english
http://www.lcaudio.dk/com/xover.htm
You can buy an assembled filter, but you can build it yourself, and it is not very expensive if you diy.
I'm planning on using the Behringer to start with, fed digitally at 96/24. Ideally I'd like a crossover that supports a World-Clock input, but it's my feeling that good DACs are cheaper than good ADCs.
I wouldn't trust the Behringer's ADC units for anything but subwoofer applications.
The nice part is that since I intend to go 5.1 with digital actives, I can always recycle the Behringer for my rear channels
and use something better for the front 3.1
I wouldn't trust the Behringer's ADC units for anything but subwoofer applications.
The nice part is that since I intend to go 5.1 with digital actives, I can always recycle the Behringer for my rear channels
and use something better for the front 3.1
On the 96k ADC, you would have to look at the specification of that converter. I may not be that if it can truly covert all 24bit at that rate, however most cheap ADC drop bit at those conversion speeds. More than likely you have a converter that maybe good for 18bits. In addition, how may bit you bit is a function to the noise performance of the design. Again, you have to have a anti-aliasing filter in front of the ADC or you will have a problem with fold over states Nyquist. Anyway, most cheap ADC is much worse than cheap DAC's.
If you look at the majority of class A components ones review in Stereoophile, the over welling majority are using passive crossovers. While it is true you can build a more feature rich active or digital filter, however what penalties. For a good one such as the PASS unit it cost and more amplifiers.
As for active crossovers, they need quality parts to and good design to make them fly. It is just much easer to get where you want for a stereo solution going a quality passive direction.
That a very good point.
If you look at the majority of class A components ones review in Stereoophile, the over welling majority are using passive crossovers. While it is true you can build a more feature rich active or digital filter, however what penalties. For a good one such as the PASS unit it cost and more amplifiers.
As for active crossovers, they need quality parts to and good design to make them fly. It is just much easer to get where you want for a stereo solution going a quality passive direction.
That a very good point.
Not quite that simple:
There is a significant cost for the original R&D + fixed overhead. A lower volume unit has to have more "profit"/unit built in (margin) in order to overcome the initial R&D development. A higher volume unit can have less profit/unit and still outrun the R&D and fixed overhead costs.
Consider this scenario:
Lower selling price target -> less margin/unit -> but higher volume -> greater total margin contibution -> subtract fixed engineering R&D cost = more profit.
Sure a really crumby DAC or ADC is well ....crumby.
But keep in mind how really good a "good" 24bit ADC or DAC is today compared to a "really good" ADC or DAC 3 years ago.
Just because component prices have come down and new chip technology and design improves specs does not mean that a implementation is now "bad" because it costs less.
I thought I'd also add that I wouldn't feed the Behringer with anything less than it's full digital data rate, as the internal SRC could be another weak point. Good standalone Sample Rate Convertors can sell for upto $1000, so why should the one in the DCX unit be any good?
The filters probably aren't great, good enough springs to mind, and the DACs are probably as good as anything you'll find fitted to any 96/24 product available for under $1500, if only because they have to operate over very narrow ranges when used in a 3-way configuration.
The filters probably aren't great, good enough springs to mind, and the DACs are probably as good as anything you'll find fitted to any 96/24 product available for under $1500, if only because they have to operate over very narrow ranges when used in a 3-way configuration.
It cetainly could be a weak point, but I am not sure that the price is a suficient indicator.
If I was going to develop a seperate sample rate converter product with case, power supply, electronics etc. and sell it at low volume, I would price it quite high.
If I was adding the capability to a design where almost all of the H/W, R&D etc. was there already, I could add it for relatively low cost.
So the proof really will be in the result. The price compared to standalone is not sufficient.
What I meant was that the Behringer probably uses a simple bit mapping model, whereas an external unit will try to upsample. With one you will get a 96/24 version of a 44 or 48 16 bit sample, whereas the other (invariably better sounding) will try to represent the original signal using 24 bits sampled 96 thousand times a second.
OK, makes sense. And your assumption may very well be correct.
Still I think it would require deeper investigation to determine how they do the conversion. The DCX2496 has a 32bit DSP with 40 bit floating point, presumably clocked pretty fast. There are a lot of good things one can do with that.
I would still claim that price vs. low volume standalone is not a suficient indicator.
I’ve used the DCX2496 for about 8 month now. During this time I managed to form an opinion about this unit. I initially purchased both the DCX2496 along with the DEQ2496 and used them with my 3-way dipole stereo system. I must say that I am very impressed with the capabilities of the DCX. Its THD is on the order of 17 bits un-weighted, which makes it as transparent as anything else on the market today. The BSS, DBX or Rane hardware is not any better and in fact worst in some cases then the DCX2496.
I would like to address the quality of the D/A and the A/D conversion process along with a sample rate converter (SRC) used in DCX. The SRC is a software implementation and it is very good indeed, since its contribution to the THD numbers is non-existent in my measurements.
Also, I think it is important to realize that the sample rate conversion process does not change the integrity of data. With this in mind, it is unequivocal that the up-sampling from 16 bits to 24 bits does not create a 24 bit signal. You cannot create something from nothing, - and so, the interpolation algorithm is trying to ‘guess’ the ‘missing’ samples but inevitably it is wrong. It is only possible with discrete sine waves, but not with real program material, since music is non-deterministic in nature.
The A/D and the D/A converters in the DCX do have a 24-bit internal architecture, but you cannot get any better then 17 bits out of this box. By the way, these 17 bits is as good as it gets today regardless if how much money you are willing to spend. In general some may find this shocking, but the fact remains that the converters used in the DCX and the DEQ products are top notch relative to other much more expensive units available today.
Now, I can see the hesitation that some may have regarding the use of an extra A/D and the D/A stage in the signal path. Well, it is only problematic if the Nyquist criterion is not followed properly. However, is it followed properly in any availably products today?
Let’s see, - we have 96 kHz sampling rate. That would mean that if we consider 20 kHz a usable audio bandwidth there is about 76 kHz between the data and the image. We need to satisfy 24-bit signal depth requirement with the anti-aliasing filter, - that puts the filter rejection at about 144 dB. Well, even with a 10th order Chebushev filter with 5 dB ripple it is impossible to satisfy the necessary S/N requirement. This filter, by the way, is the worst for audio and should never be used. It is, however, very steep and I am using it here to illustrate the point.
So the DCX designers were forced to ‘live’ with a much higher S/N. I am sure they, like many others, figured that a 16 bit requirement for the filter rejection band was adequate. Well, that may be, but such engineering decisions simply ‘kill’ any hope of 24-bit resolution.
If I for a moment disregard the noise coming from external sources, the power supply noise and passive component nose and simply look at what is possible as far as 24-bit A/D converter is concerned; I must say that perhaps if I had a 192 kHz sampling rate, then I can build an anti-aliasing filter that will do justice to 24-bit signal. With 96 kHz it is impossible in my view, as the filter requirements, such as the order, will simply introduce more noise and distortion.
I do not know what exactly the DCX design uses for a filter, but my measurements indicate that at least a 16-17 bit signal can be recovered. This unfortunate reality is also present in any other digital boxes today. It is impossible to do much better and my own very exotic DIY D/A converters could never do better then 18 bits.
So, considering all of the above techno-babble, I submit that the DCX2496 is virtually imperceptible in a signal chain due to its honest THD numbers in the area of 16-17 bits. It is also as good if not better then the BSS box. No amount of internal modification will make the DCX much better. The internal mods may raise the THD up to 18 bits, but very unlikely more then that. However, if anyone is brave enough to bring the digital signals from the D/A outside and redesign the any-imaging filters along with super low noise power supplies, - that along may do wonders for the S/N. I would also like to see the same treatment for the input chain. Those mods will create a true monster out of a small box that the DCX is today.
Well, in the end I ended up selling my DCX, but I kept the DEQ. Unfortunately, my dipoles need more headroom then the DCX can give me. But I have absolutely no problems with keeping the DEQ2496 that has identical A/D and D/A converters and associated electronics.
Vadim
I would like to address the quality of the D/A and the A/D conversion process along with a sample rate converter (SRC) used in DCX. The SRC is a software implementation and it is very good indeed, since its contribution to the THD numbers is non-existent in my measurements.
Also, I think it is important to realize that the sample rate conversion process does not change the integrity of data. With this in mind, it is unequivocal that the up-sampling from 16 bits to 24 bits does not create a 24 bit signal. You cannot create something from nothing, - and so, the interpolation algorithm is trying to ‘guess’ the ‘missing’ samples but inevitably it is wrong. It is only possible with discrete sine waves, but not with real program material, since music is non-deterministic in nature.
The A/D and the D/A converters in the DCX do have a 24-bit internal architecture, but you cannot get any better then 17 bits out of this box. By the way, these 17 bits is as good as it gets today regardless if how much money you are willing to spend. In general some may find this shocking, but the fact remains that the converters used in the DCX and the DEQ products are top notch relative to other much more expensive units available today.
Now, I can see the hesitation that some may have regarding the use of an extra A/D and the D/A stage in the signal path. Well, it is only problematic if the Nyquist criterion is not followed properly. However, is it followed properly in any availably products today?
Let’s see, - we have 96 kHz sampling rate. That would mean that if we consider 20 kHz a usable audio bandwidth there is about 76 kHz between the data and the image. We need to satisfy 24-bit signal depth requirement with the anti-aliasing filter, - that puts the filter rejection at about 144 dB. Well, even with a 10th order Chebushev filter with 5 dB ripple it is impossible to satisfy the necessary S/N requirement. This filter, by the way, is the worst for audio and should never be used. It is, however, very steep and I am using it here to illustrate the point.
So the DCX designers were forced to ‘live’ with a much higher S/N. I am sure they, like many others, figured that a 16 bit requirement for the filter rejection band was adequate. Well, that may be, but such engineering decisions simply ‘kill’ any hope of 24-bit resolution.
If I for a moment disregard the noise coming from external sources, the power supply noise and passive component nose and simply look at what is possible as far as 24-bit A/D converter is concerned; I must say that perhaps if I had a 192 kHz sampling rate, then I can build an anti-aliasing filter that will do justice to 24-bit signal. With 96 kHz it is impossible in my view, as the filter requirements, such as the order, will simply introduce more noise and distortion.
I do not know what exactly the DCX design uses for a filter, but my measurements indicate that at least a 16-17 bit signal can be recovered. This unfortunate reality is also present in any other digital boxes today. It is impossible to do much better and my own very exotic DIY D/A converters could never do better then 18 bits.
So, considering all of the above techno-babble, I submit that the DCX2496 is virtually imperceptible in a signal chain due to its honest THD numbers in the area of 16-17 bits. It is also as good if not better then the BSS box. No amount of internal modification will make the DCX much better. The internal mods may raise the THD up to 18 bits, but very unlikely more then that. However, if anyone is brave enough to bring the digital signals from the D/A outside and redesign the any-imaging filters along with super low noise power supplies, - that along may do wonders for the S/N. I would also like to see the same treatment for the input chain. Those mods will create a true monster out of a small box that the DCX is today.
Well, in the end I ended up selling my DCX, but I kept the DEQ. Unfortunately, my dipoles need more headroom then the DCX can give me. But I have absolutely no problems with keeping the DEQ2496 that has identical A/D and D/A converters and associated electronics.
Vadim
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.