Interesting thread. It's a bit apples/oranges to compare analog/DSP because each have their advantages/disadvantages. The Orion is a bit of special case because of the largish EQ requirements and possible clipping concerns with some of the DSP-based crossovers like the DCX. (DSP boxes make the addition of room correction simpler though.)
Steve Dodds uses the DCX2496 with his "Bob" system and has good results. http://www.doddsy.net/steve6_009.htm
However, the EQ requirements for the woofers are not as great as the Orion so it's a better situation.
"however it is distinctly unaudiophile to have 10 opamp stages between signal source" What a ridiculous statement.
More than likely the signal has already passed through dozens of op-amps before ever being mastered onto a CD. A few more...if they accomplish a beneficial goal and are well designed...are certainly not anything to worry about.
Cheers,
Davey.
Steve Dodds uses the DCX2496 with his "Bob" system and has good results. http://www.doddsy.net/steve6_009.htm
However, the EQ requirements for the woofers are not as great as the Orion so it's a better situation.
"however it is distinctly unaudiophile to have 10 opamp stages between signal source" What a ridiculous statement.
More than likely the signal has already passed through dozens of op-amps before ever being mastered onto a CD. A few more...if they accomplish a beneficial goal and are well designed...are certainly not anything to worry about.Cheers,
Davey.
Thanks for all your replies. Some interesting points
There are a significant number of op-amp stages in the Orion XO which must contribute something to the sound - a simple experiment with a opamp follower circuit will show you that. So the question arises how to minimise the effect of these additional stages. DSP would be one route but the gain problem arises with the Orion, which introduces a raft of additional complexity, due to the need for downstream volume control and/or addtional offboard opamp stages. Furthermore, I am not convinced that the D/A converters and analog output stages in the cheaper off the shelf DSP boxes are as good as they could be due to cost constraints plus the issue of additional A/D conversion or the effect of a non-optimal SPDIF / AES reciever circuit. The cost of optimising these issues is sizeable (optimised D/A and output stages, remote 6 channel volume control), suggesting that a computer based XO (e.g. Brutefir, LYNX AES16 soundcard) could provide a solution.
The benefits of DSP appear to be the ability to introduce room correction and phase linear XOs. Something like the DEQX. I remember Mac (a diyaudio contributor) said the DEQX was a definite improvement over previous analogue efforts. Any1 else have any experience?
So other solutions could be appropriate - biasing op-amps into class A, optimising grounding, rolling discrete op-amps....
There are a significant number of op-amp stages in the Orion XO which must contribute something to the sound - a simple experiment with a opamp follower circuit will show you that. So the question arises how to minimise the effect of these additional stages. DSP would be one route but the gain problem arises with the Orion, which introduces a raft of additional complexity, due to the need for downstream volume control and/or addtional offboard opamp stages. Furthermore, I am not convinced that the D/A converters and analog output stages in the cheaper off the shelf DSP boxes are as good as they could be due to cost constraints plus the issue of additional A/D conversion or the effect of a non-optimal SPDIF / AES reciever circuit. The cost of optimising these issues is sizeable (optimised D/A and output stages, remote 6 channel volume control), suggesting that a computer based XO (e.g. Brutefir, LYNX AES16 soundcard) could provide a solution.
The benefits of DSP appear to be the ability to introduce room correction and phase linear XOs. Something like the DEQX. I remember Mac (a diyaudio contributor) said the DEQX was a definite improvement over previous analogue efforts. Any1 else have any experience?
So other solutions could be appropriate - biasing op-amps into class A, optimising grounding, rolling discrete op-amps....
Big Dada said:
Hey Big Dada
The more I think about this and read others opinions and commercial efforts in this area (Emotive Excellence, Overkill Audio, NHT, Omen Ra), makes me think that the ONLY real way to go if DSP is to be pursued is to use a computer to store music, do phase linear XOs and room correction and feed out multiple digital signals to dacs with volume control after the dacs. I just need someone to build me a PC with ability to use phase linear XOs and room correction....(and yes, I have been following others progress in this area, it seems like a significant amount of work and cost to get this set up working!)
Big Dada said:Hey Nanu
I agree with your assessment of PC based storage, player & DSP, time and cost.
And it is right there on my to-do list.
My guess is it will all be oldhat in 5years time. So you could just play some music and chill out till then.
If it is oldhat in 5 years time it may just be coming to the top of your to do list
Huh? A subtractive crossover can easily be built with exactly the same slopes and overlap as an LR-4....or just about any other crossover configuration.
I think the reason you don't see subtractive crossovers used more is they're not any less complex than a typical Sallen-Key or State-Variable topology since they need a differential amplifier and all-pass filter circuit to create the complement to the actual filter section. It's about the same number of active stages any way you proceed.
Also, subtractive crossovers are similar to the State-variable in that asymetrical slopes and independent xover frequencies are not supported.
Cheers,
Davey.
I think the reason you don't see subtractive crossovers used more is they're not any less complex than a typical Sallen-Key or State-Variable topology since they need a differential amplifier and all-pass filter circuit to create the complement to the actual filter section. It's about the same number of active stages any way you proceed.
Also, subtractive crossovers are similar to the State-variable in that asymetrical slopes and independent xover frequencies are not supported.
Cheers,
Davey.
phase_accurate said:
There seems to be a misunderstanding. I never talk about those subtractive crossovers that mimic an LR for instance since I see no reason for using them at all.
When I say subtractive crossover then I mean the constant-voltage and transient-perfect type of subtractive crossover.
And those have asymmetric slopes (at least in their simplest form) and they also have large overlap and humps.
Regards
Charles
When I say subtractive crossover then I mean the constant-voltage and transient-perfect type of subtractive crossover.
And those have asymmetric slopes (at least in their simplest form) and they also have large overlap and humps.
Regards
Charles
Hi,
I don't listen to the subtractive filter I bought.
When first wired in it sounded pretty poor, no worse than that.
I then learned how to measure it's outputs and discovered that the xover slopes and F3db points were all over the place and not even coincident.
Later found a mistake in the circuit board that mixed up one of the subtractive phases on both channels. Threw it in the cupboard & went back to passive xover.
Getting ready to go active again after a gap of 15 or so years.
I don't listen to the subtractive filter I bought.
When first wired in it sounded pretty poor, no worse than that.
I then learned how to measure it's outputs and discovered that the xover slopes and F3db points were all over the place and not even coincident.
Later found a mistake in the circuit board that mixed up one of the subtractive phases on both channels. Threw it in the cupboard & went back to passive xover.
Getting ready to go active again after a gap of 15 or so years.
Big Dada said:
The drummer was lurking in the shape of a Benchmark DAC1 that I picked up as a reference for my 2nd all out non-os implementation. It is more transparent and dynamic than my old dac, showing that the Orions are very sensitive to changes upstream....but does not have as much musicality or tonal colour as the non-os.
After I've finished the Class Ds (one is nearly ready) I will start on the all out non-os implementation, I will be interested to see how close I can get it to the Benchmark....
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.