Apologies in advance for the wall of text. There's a fairly simple question you can skip to at the end but the rest is included because it's part of a wider plan with an imperfect solution and I'm open to ideas if anyone wants to really get into it, or just hear about the project we're starting to plan.
I got the free trial of Dirac and I rather liked it. I'd like to buy it but I'm having trouble figuring out in which form.
Medium term I'm planning to build a good digital chain. It'll be composed of a modular system connected with I2S, possibly in one physical box. There will be some sort of media player module, a DAC (probably Twisted Pear), source selection, volume, Dirac, and an active crossover.
All that's going to take shape over the next year or two, for now the system is PC, USB to SPDIF interface, DAC, and preamp.
One possible complication is that I have both a digital and an analogue source. That means I need some sort of A/D conversion if I want the analogue to go through the DSP, or post digital source selection if I accept that's not going to happen.
The analogue source is my turntables. Plural because a long time ago I used to DJ and I keep the setup to play around with occasionally, with a few upgrades to make it suitable for quality single deck listening (very occasional, almost all digital these days). Occasional DJ use means on the one hand that if it goes through a digital stage it need to have a very low latency but on the other it's not the end of the world if it can't go through the DSP and gets selected after the digital chain.
What would be perfect for me is source selection, incorporating at least two digital inputs (future proofing) and an analogue input with high quality, very low latency A/D. I figure attenuation pre-DSP is a bad idea as it'll squish the dynamic range and/or raise the noise floor and I'm going for optimal quality here. That would mean separate source control and volume, the latter being analogue, after the digital chain and right before the power amp.
I should note that I work with a brilliant audio engineer who, if it's at all technically possible, can tap the I2S on the individual boards and the suchlike, even if they don't have native external interfaces for it. So it's more a question of whether things are practically technically possible than if they're built in as standard user I/O.
So, that's the plan medium term but my main concern right now what implementation of Dirac I should go for in order to ensure future compatibility with the planned system (Dirac can't be re-sold and I really don't want to pay for it twice). Dirac comes first because it's all I can afford right now, I can get it off the peg and working straight away, and I feel a desperate need for DSP to help with room acoustics (I'll be treating the room as well but I'm moving house in a bit so that has to wait).
There are three main options I'm looking at (although other suggestions are welcome) because none of them quite tick all the boxes or at least I have questions/concerns with them:
Option 1:
MiniDSP DDRC-24
This has analogue and digital inputs, which is great. However it only has analogue outputs and so would prevent me from using separate DAC(s). To be honest I don't trust at least the output stage of the DAC in this to be up there with the Twisted Pear stuff I'm planning. My engineer chum is confident he can tap the I2S in a standard 2x4HD but he's concerned that there may be some f***ery involved in the Dirac version which would prevent this, and it's a lot of money to take a chance on. This is compounded by the inability to resell Dirac products meaning I couldn't just move it on for a relatively minor loss if it's not suitable.
Second concern is how well Dirac runs on these boards. The rest of the MiniDSP Dirac embedded products seem to run on a different platform. Take the DDRC-22DA. Differences seem to be that it doesn't function as an active crossover (having only two outputs), has no analogue input, but does give balanced outputs and AES input. Yet it costs over 300 Euros more. Is this really just horribly squiffy pricing where balanced outputs and AES are considered to be worth about 300 Euros plus the loss of the analogue input and the crossover or does this platform somehow have superior processing in this context? What am I missing here?
Option 2:
MiniDSP DDRC-22D
AES I/O is nice to have as I'm slowly migrating to pro hookups but it's not that big a deal, especially as I might wind up giving it I2S I/O anyway. The attraction of this is in how little it does. Having a box with digital in and out which only handles the Dirac processing leaves all my options wide open in terms of the future modular system, but would also work with what I have right now (possibly excluding the analogue). The issue I have is that it's the best part of double the price of the other options for apparently less functionality. Just over 700 Euros for two channel Dirac only is hard to swallow. So I guess same question as above. Is there anything about this platform which justifies the significant extra cost?
Option 3:
Dirac for PC
Removes the option for a dedicated digital source, unless I pipe it through the PC, which negates the point. It would also create a similar problem for analogue in that I would have to sort out the analogue I/O at the PC which is an unwanted complication and cost if I want high quality and low latency. Advantage is that it's (relatively) cheap.
So, it seems that none of the options are quite right for me. If MiniDSP made a version of the DDRC-24 with digital outputs, if I knew that there was no compromise from the other platform beyond the lack of AES, and if I knew for sure the I2S could be tapped in the future, it'd be perfect.
The more I think about it the more I reckon I'll have to forget about processing the analogue and restrict DSP to digital sources. I've included analogue considerations in case anyone can see a way of making this happen but I think, especially considering latency issues, I can and will need to live without it. In that case I guess the obvious answer is the DDRC-22D, but can anyone tell me if anything justifies the extra cost in this unit? That's a serious question because if it does do something in the processing better then I could possibly be persuaded to save for a little bit longer.
I got the free trial of Dirac and I rather liked it. I'd like to buy it but I'm having trouble figuring out in which form.
Medium term I'm planning to build a good digital chain. It'll be composed of a modular system connected with I2S, possibly in one physical box. There will be some sort of media player module, a DAC (probably Twisted Pear), source selection, volume, Dirac, and an active crossover.
All that's going to take shape over the next year or two, for now the system is PC, USB to SPDIF interface, DAC, and preamp.
One possible complication is that I have both a digital and an analogue source. That means I need some sort of A/D conversion if I want the analogue to go through the DSP, or post digital source selection if I accept that's not going to happen.
The analogue source is my turntables. Plural because a long time ago I used to DJ and I keep the setup to play around with occasionally, with a few upgrades to make it suitable for quality single deck listening (very occasional, almost all digital these days). Occasional DJ use means on the one hand that if it goes through a digital stage it need to have a very low latency but on the other it's not the end of the world if it can't go through the DSP and gets selected after the digital chain.
What would be perfect for me is source selection, incorporating at least two digital inputs (future proofing) and an analogue input with high quality, very low latency A/D. I figure attenuation pre-DSP is a bad idea as it'll squish the dynamic range and/or raise the noise floor and I'm going for optimal quality here. That would mean separate source control and volume, the latter being analogue, after the digital chain and right before the power amp.
I should note that I work with a brilliant audio engineer who, if it's at all technically possible, can tap the I2S on the individual boards and the suchlike, even if they don't have native external interfaces for it. So it's more a question of whether things are practically technically possible than if they're built in as standard user I/O.
So, that's the plan medium term but my main concern right now what implementation of Dirac I should go for in order to ensure future compatibility with the planned system (Dirac can't be re-sold and I really don't want to pay for it twice). Dirac comes first because it's all I can afford right now, I can get it off the peg and working straight away, and I feel a desperate need for DSP to help with room acoustics (I'll be treating the room as well but I'm moving house in a bit so that has to wait).
There are three main options I'm looking at (although other suggestions are welcome) because none of them quite tick all the boxes or at least I have questions/concerns with them:
Option 1:
MiniDSP DDRC-24
This has analogue and digital inputs, which is great. However it only has analogue outputs and so would prevent me from using separate DAC(s). To be honest I don't trust at least the output stage of the DAC in this to be up there with the Twisted Pear stuff I'm planning. My engineer chum is confident he can tap the I2S in a standard 2x4HD but he's concerned that there may be some f***ery involved in the Dirac version which would prevent this, and it's a lot of money to take a chance on. This is compounded by the inability to resell Dirac products meaning I couldn't just move it on for a relatively minor loss if it's not suitable.
Second concern is how well Dirac runs on these boards. The rest of the MiniDSP Dirac embedded products seem to run on a different platform. Take the DDRC-22DA. Differences seem to be that it doesn't function as an active crossover (having only two outputs), has no analogue input, but does give balanced outputs and AES input. Yet it costs over 300 Euros more. Is this really just horribly squiffy pricing where balanced outputs and AES are considered to be worth about 300 Euros plus the loss of the analogue input and the crossover or does this platform somehow have superior processing in this context? What am I missing here?
Option 2:
MiniDSP DDRC-22D
AES I/O is nice to have as I'm slowly migrating to pro hookups but it's not that big a deal, especially as I might wind up giving it I2S I/O anyway. The attraction of this is in how little it does. Having a box with digital in and out which only handles the Dirac processing leaves all my options wide open in terms of the future modular system, but would also work with what I have right now (possibly excluding the analogue). The issue I have is that it's the best part of double the price of the other options for apparently less functionality. Just over 700 Euros for two channel Dirac only is hard to swallow. So I guess same question as above. Is there anything about this platform which justifies the significant extra cost?
Option 3:
Dirac for PC
Removes the option for a dedicated digital source, unless I pipe it through the PC, which negates the point. It would also create a similar problem for analogue in that I would have to sort out the analogue I/O at the PC which is an unwanted complication and cost if I want high quality and low latency. Advantage is that it's (relatively) cheap.
So, it seems that none of the options are quite right for me. If MiniDSP made a version of the DDRC-24 with digital outputs, if I knew that there was no compromise from the other platform beyond the lack of AES, and if I knew for sure the I2S could be tapped in the future, it'd be perfect.
The more I think about it the more I reckon I'll have to forget about processing the analogue and restrict DSP to digital sources. I've included analogue considerations in case anyone can see a way of making this happen but I think, especially considering latency issues, I can and will need to live without it. In that case I guess the obvious answer is the DDRC-22D, but can anyone tell me if anything justifies the extra cost in this unit? That's a serious question because if it does do something in the processing better then I could possibly be persuaded to save for a little bit longer.
