Thank you Ryssen. I could have waited for your file before joining yahoo but it will not hurt to be there too.
I really wanted to make 3 different mods on each output (naked with just a cap, transformer and tube) to have a clear idea which way to go but now after messing up with the caps (probably unnecessarily) I am worried it may be super difficult to put it back together.
I really wanted to make 3 different mods on each output (naked with just a cap, transformer and tube) to have a clear idea which way to go but now after messing up with the caps (probably unnecessarily) I am worried it may be super difficult to put it back together.
I keep wondering about option 4, piggyback off the I2S feeding each DAC and feed it to little ES9023 boards? NO disabling the DCX AKM DACs just have a choice.
Yes, after the damage I thought of that too. But I would not advise you to go this way if your DCX works fine simply because AKM DAC is so fantastic without opamps. I own DEQ2496 with tube output done by Lampizator and it is sensational. I also have ES9023 which I compared with AK4393 playing straight through the cap and to me AKM sounds better.
Lampizator, hmm, gain stage or a cathode-follower? I would think it doesn't need gain. If Gain, do you then pad it back down? I was all set to do Broskie UNbalancers with 6n16b-v tubes [I have the PCBs] but that is a huge amount of work for a DCX. 3x ES9023 would be 1/2 a day maybe, unless using the *awesome* Curryman boards from MiniDSP which would take more PSU work [not bad]. The little DIYINHK ES9023 x 3 would be fast and leave the DCX with differential and SE outs, and caps from your single AKM DAC legs also for SE. I have not tried ES9023s in DCX, no time. I built a AK4396 DAC board fed with TE8802 board, OMG lots of rich bass -- I liked ES9023 better for sound-stage [heavily modified with regulators for the DAC, the USB-I2S, the TCXO I dropped on]. THat was a heavily modded HiFiMEDIY? TE7022/ES9023 with Minus V injected [not using ES9023 charge-pump for -V]. Caps from AKM only required if input to next SE{?} stage DOESN'T have caps. So if you know there ARE input caps in device down-wind, I short the output caps with a multi-pole switch [in my brother-in-law's BBE Sonic Maximizer for instance]. Likely I won't get to try my ideas until Winter. I have the hot-air soldering to learn on something else first; I bet DCX traces peel just looking at them.
My DEQ uses SRPP and no extra gain is required. Funny thing about gain is that DCX handles it differently than DEQ where clipping comes much earlier. DCX seems to cope way better.
You are right about the caps, I actually took signal straight from AKM chip to the input caps of my TK2050 amp to compare it with hifimediy Sabre DAC. AKM is more natural sounding to me which is evident especially when using highly sensitive speakers. But as a test it would not be too difficult to attach some small es9023 board. They can take voltage from small batteries so you can't really imagine better PSU for your DAC.
When talking about the tube output stage for DCX I saw a few PCBs on Ebay that cost less than a transformer for them. Has anyone tried them and know what they are worth?
You are right about the caps, I actually took signal straight from AKM chip to the input caps of my TK2050 amp to compare it with hifimediy Sabre DAC. AKM is more natural sounding to me which is evident especially when using highly sensitive speakers. But as a test it would not be too difficult to attach some small es9023 board. They can take voltage from small batteries so you can't really imagine better PSU for your DAC.
When talking about the tube output stage for DCX I saw a few PCBs on Ebay that cost less than a transformer for them. Has anyone tried them and know what they are worth?
Phase question about this unit:
HUGE phase question about this unit:
I was under the impression, that if we set it for a high order crossover, such as 48 DB/octave, that it would not cause anymore group delay then 1st order. IOW, I am not talking about timing between the speakers, which can be compensated for, but the phase distortion caused by the higher order filters.
I know that in analog, at 48/octave, the phase will get thrown way off. Does that STILL happen with this one? I thought that being in the digital domain, this was immune to that problem.
Can somebody help clear this up for me?
Thank you!
HUGE phase question about this unit:
I was under the impression, that if we set it for a high order crossover, such as 48 DB/octave, that it would not cause anymore group delay then 1st order. IOW, I am not talking about timing between the speakers, which can be compensated for, but the phase distortion caused by the higher order filters.
I know that in analog, at 48/octave, the phase will get thrown way off. Does that STILL happen with this one? I thought that being in the digital domain, this was immune to that problem.
Can somebody help clear this up for me?
Thank you!
HUGE phase question about this unit:
I was under the impression, that if we set it for a high order crossover, such as 48 DB/octave, that it would not cause anymore group delay then 1st order. IOW, I am not talking about timing between the speakers, which can be compensated for, but the phase distortion caused by the higher order filters.
I know that in analog, at 48/octave, the phase will get thrown way off. Does that STILL happen with this one? I thought that being in the digital domain, this was immune to that problem.
Can somebody help clear this up for me?
Thank you!
Dave,
I'm not sure where that impression came from.
Brett is correct, the phase distortion created by the IIR filters in the DCX is identical to analog counterparts. Even the separate phase control capability of the DCX implements an (additive) all-pass function equivalent to the phase distortion of the filter you have selected in the crossover.If you want zero phase distortion (or something less than an analog equivalent) you need a different DSP platform.
Something that can implement FIR filters. The miniSHARC board from miniDSP will do it and a few others. OR, another option is, to phase-distort your source with a utility like rePhase to "unwrap" the phase distortion of a crossover like the DCX2496.Cheers,
Dave.
^^ I've heard a well implemented 5 way system with a pair of DEQX driving them, with very high slope FIR, phase correction etc, and the same system (flick of a switch) with LR48 filters implemented in the DEQXs (all else the same) and the difference was very slight, and not everyone in the room agreed which they preferred. I think phase is a bit of a non issue for decently implemented xovers.
For arguments sake let's say one wants to keep the DCX (nice sounding, with linear PSU, regulators, Pilgham digital input mods, passive transformer outputs).
Using REW generated EQ filters for DCX makes a notable improvement.
Will (could) using Rephase FIR phase correction contribute further? This will of course require additional HW (like miniSHARC, or openDRC, or a spare PC....) hence the question.
I really don't know. You would have to know what the phase characteristics are of the filters you use in the DCX (that can be looked up of course) and compensate for that, can de Rephase do that?
Also whenever you change the filters in the DCX (filter type, frequency) you would have to adapt the Rephase.
I don't know the Rephase so it may be possible but I wonder if it is worth the effort.
Jan
Also whenever you change the filters in the DCX (filter type, frequency) you would have to adapt the Rephase.
I don't know the Rephase so it may be possible but I wonder if it is worth the effort.
Jan
You don't necessarily need additional hardware. If your source is a computer running Foobar2000 or JRiver or whatever, you can generate a rePhase linearization correction and have your player process it. There's already a long thread on rePhase here at DIYaudio, so queries regarding it should be placed there.
Whether linear-phase correction will improve your audible result is another question.........and one that has been debated for years.
Dave.