Just had some reflexion on my time with the D18.
First of all, I found that my test with the RCA outs was somewhat flawed.
I used a decent quality stereo RCA to XLR cable with neutrik connectors 2 meters length (cost about 25 euro), then connected that cable to decent quality balanced XLR to XLR cables also 2 meters length with neutrik connectors (cost about 55 euro if I remember correctly), which connected to the speakers.
I had it set up this way because I usually used a balanced connection so for unbalanced it was quicker to connect it to the balanced cables already connected to my speakers.
This afternoon I hooked up my old Roland UA-1EX again which is a USB DAC with RCA outs, with the cable configuration as described above and it sounded just like before (and actually somewhat similar to the way the D18 sounded from the RCA outs).
But then thought, well I don't have a balanced DAC anymore at the moment, so why not take out the balanced cables and simply connect the UA-1EX directly to the speakers with the RCA to XLR cable.
I did not expect to hear any difference. I once did a careful test with different interconnects a few years ago with a different speaker and room setup and didn't hear any difference until really cheap cables were used (though these cables were all less than 2 meters)
But right away after taking out the excess 2 meter balanced cable this time.. a very clear difference!
The "analogue" sounding top end became quite a bit cleaner and transparent. Enough to make a very real difference.
Enough in fact to make me like the sound of the UA-1EX connected this way more than I liked the sound of the D18 connected through RCA with the extra cable extension.
It simply sounds more transparent and real now, I quite like it and consider it good enough to do studio work on.. bit surprised by this. (could it be due mainly to the extra connection between cables, or does 2 meter of extra cable has this effect.. And could it even perhaps be that some output stages are more sensitive to outgoing cable sound degradation? I'm going to do more testing with this and perhaps make my own RCA to XLR cables from higher quality parts)
Since the D18 is underway to it's new owner I can't check how the RCA of the D18 would sound connected this way, but I suspect that what I wrote before about less treble through the D18 RCA outs is not true and was in fact cable induced.
Sorry about this.
I realize that I'm not a great reviewer because of several reasons.
I don't have my setup to the point where I want it yet, still tweaking things including room, and have plenty of things left to check.
Also don't have a true reference DAC (or several reference DACs) for direct A/B comparison. (and ideally one would do measured level matched blind comparisons, requiring the help of a friend)
And another reason I'm not a great reviewer is that I'm way too much of a perfectionist (I've just spent 6 years researching just intonation full time lol), and find I have only one true goal for a DAC. Simply honest and uncolored as possible reproduction of the recorded audio as it is represented in the bits. (even if this means half of all music becomes unpleasant on some level to listen to, quite different from the goal of many audiophile music lovers)
In hindsight I don't know what I was thinking getting a DAC clearly described as smooth :)
So my review was first part honeymoon (first day listening) where I heard all the things the D18 did right that for instance the DA10 did wrong, though I noticed the treble was too smooth to me. Then after more listening I found many other things that were not to my liking for my personal goals. Others may think the exact opposite of these things in their system for their listening pleasure goals.
I have found it very insightful to hear the D18 though and have also found that I find it interesting to really learn about DAC chips and power supplies and output stages etc.
Have been doing a lot of reading lately, and I don't think the Sabre chip is something for me and fitting my goals. For instance the asynchronous resampling and the way it does jitter reduction are not the purest things possible.
I can't really find a DAC that does things the way I'd like to see them done at a reasonable price.
So I'm thinking I'm going to build my own DAC, time to take the plunge into serious DIY :)
I'm thinking using the computer for upsampling in highest quality, then asynchronous USB at 32bit (well actually 24bit is plenty) 384khz to the DAC chip without further resampling and one precise clock that times both the asynchronous USB and the DAC chip bringing them in perfect sync to work as one without any tricks (I mean this is how the real benefits of asynchronous usb will truly show)
Oh and no spdif inputs at all, don't need them, only USB and only 384 or 352.8khz (only need one, computer does upsampling well enough to not matter that it's not a direct multiple, only costs more processing for the computer) . That's all for the digital side. Analogue filtering after that should leave everything under lets say 25khz untouched including phase, somewhere above that 1st order filter (and I don't care about high bandwidth above that it's inaudible to human ears, high sample rate is not about high frequencies)
Output stage as transparent as possible, power perhaps from batteries. These things I know very little about yet, but will be fun to learn.
Oh and computer for fine level control (dithered), and a switch for something like -15db and -30db analogue attenuation on the DAC output so I can connect it directly to my speakers power amp without lowering the signal too much into the output stage noise floor when listening more quietly and using computer for volume control (my speakers amp input has an attenuation pot built in but it degrades the sound a little and is best left full open)
Such a pure DAC can be cheap I'm figuring. A great designer could probably make it for 200-300 euro in parts costs or cheaper even.
I hope I can do it in a reasonable time (half a year?) with 2-3 times as much money.
Ohyeah btw I forgot to mention that I did all my listening to the D18 at 96khz, with all 44.1khz music upsampled by SOX at highest quality setting linear phase.
I found it did make an audible improvement to the D18, more of a difference than it does for for instance the Lavry DA10, much less than it does for Echo Audiofire2.
Not a big difference but perceived quality went up, it sounded slightly "deader/flatter" when feeding 44.1 directly.
Edited by slackman - 1/4/12 at 12:07am