I spend an afternoon playing with the nanoDIGI. Nice piece of equipment. Definitely a step up from the minDSP.
First, the setup: Files on a laptop played through Foobar2000. This computer only has HDMI, USB and headphone output, and the nanoDIGI has only SPDIF input (either coax of TosLink). I have a USB>>SPDIF converter to feed the nanoDIGI. I don’t hear any difference between the coax and the TosLink connection. The output of the nanoDIGI goes to a pair of Topping D-20 DAC’s. The bass goes to a Crown XLS 1500 and on to a pair of Dayton RSS315HF-4 subs in 50l sealed boxes. The treble goes to a Topping TP-60 and a pair of Alpair A10P’s in 16l BR’s.
The miniDSPis a nice piece of equipment and I recommend it as an introduction to DSP and active XO. It has a couple of limitations that bother me. First, the miniDSP is analog input. It then converts to digital, does its thing and then converts back to analog. This means that you only need one DAC to feed the minDSP, and the miniDSP contains DAC’s to drive the amp. The output of the miniDSP is only 0.9v. This is not enough to voltage to drive my amps to full power. Using the nanoDIGI, my DAC’s are providing a full 2v to amps and they can be driven to clipping. The second problem is psychological – the extra DAC>>ADC step is just wrong. Yes, I hear a difference between the minDSP driving the amps and the Topping DAC’s driving the amps.
The FR plots below were created with the microphone at the listening position 2.5m from the speakers 2.5m apart. Therefore they contain all of the room interaction. The plots are 1/6 octave smoothed. The black trace is both channels running. It’s not bad. Does leave some room for more tweaking. The red trace is the right channel only and the green the left. Note how bad each channel is by itself, although the combination is quite acceptable. That’s what you get when the room is not symmetrical, and who has a symmetrical listening room. You must resist the urge to EQ each channel separately. When that is done, the combined result is a disaster. Individual features can be flattened, but the effect must be seen in the combined output. See the peak in the right channel at 180Hz. BTW the XO is a 200. Some work needs to be done with that.
My impression of this EQ is that is somewhat bass shy and a bit flat sounding. The lack of bass may be due to the depression at 70Hz, but this is the area of maximum room interaction, and there is not much that can be done about it. The overall sound is great for instrumentals, but too uninvolved for female voice. A bump up just below 1kHz would fix that and a bit of tip up at the top would add “Life”. So many possibilities.
Bob
First, the setup: Files on a laptop played through Foobar2000. This computer only has HDMI, USB and headphone output, and the nanoDIGI has only SPDIF input (either coax of TosLink). I have a USB>>SPDIF converter to feed the nanoDIGI. I don’t hear any difference between the coax and the TosLink connection. The output of the nanoDIGI goes to a pair of Topping D-20 DAC’s. The bass goes to a Crown XLS 1500 and on to a pair of Dayton RSS315HF-4 subs in 50l sealed boxes. The treble goes to a Topping TP-60 and a pair of Alpair A10P’s in 16l BR’s.
The miniDSPis a nice piece of equipment and I recommend it as an introduction to DSP and active XO. It has a couple of limitations that bother me. First, the miniDSP is analog input. It then converts to digital, does its thing and then converts back to analog. This means that you only need one DAC to feed the minDSP, and the miniDSP contains DAC’s to drive the amp. The output of the miniDSP is only 0.9v. This is not enough to voltage to drive my amps to full power. Using the nanoDIGI, my DAC’s are providing a full 2v to amps and they can be driven to clipping. The second problem is psychological – the extra DAC>>ADC step is just wrong. Yes, I hear a difference between the minDSP driving the amps and the Topping DAC’s driving the amps.
The FR plots below were created with the microphone at the listening position 2.5m from the speakers 2.5m apart. Therefore they contain all of the room interaction. The plots are 1/6 octave smoothed. The black trace is both channels running. It’s not bad. Does leave some room for more tweaking. The red trace is the right channel only and the green the left. Note how bad each channel is by itself, although the combination is quite acceptable. That’s what you get when the room is not symmetrical, and who has a symmetrical listening room. You must resist the urge to EQ each channel separately. When that is done, the combined result is a disaster. Individual features can be flattened, but the effect must be seen in the combined output. See the peak in the right channel at 180Hz. BTW the XO is a 200. Some work needs to be done with that.
My impression of this EQ is that is somewhat bass shy and a bit flat sounding. The lack of bass may be due to the depression at 70Hz, but this is the area of maximum room interaction, and there is not much that can be done about it. The overall sound is great for instrumentals, but too uninvolved for female voice. A bump up just below 1kHz would fix that and a bit of tip up at the top would add “Life”. So many possibilities.
Bob
Attachments
Great topic and great timing as we are discussing virtues and evils of EQ'ing in DSP in Nautaloss thread. I have the newer miniDSP and it has 2 volt drive capability but the input impedance must be pretty low because my normal headphone level amp does not drive it to clipping. There is also a $40 digital input option for the miniDSP that snaps on as a daughter board to give you spdif and I think optical input.
Your freq response in room looks great btw. You have to be careful with EQ'ing at the 1khz range as the EQ in nanoDIGI is IIR and will present phase distortions in critical vocal range but as Anthonybisset pointed out, use EQ'ing in wide broad strokes and reduce peaks rather than fill valleys and this minimizes noticeable phase effects.
Your freq response in room looks great btw. You have to be careful with EQ'ing at the 1khz range as the EQ in nanoDIGI is IIR and will present phase distortions in critical vocal range but as Anthonybisset pointed out, use EQ'ing in wide broad strokes and reduce peaks rather than fill valleys and this minimizes noticeable phase effects.
I realize that there are other solutions in building miniDSP stacks, but I don't do electronics, so I have the in-a-box versions.
Yes, one must be careful when applying filters. Order is important, and applying a +ve peak filter after a few others can give truly bizarre results. This why I have a hole around the main room mode at 70Hz. No way to EQ it out. Just have to live with it. Of course, moving the system around the room might help, but as often is the case, that is not an option.
Bob
Yes, one must be careful when applying filters. Order is important, and applying a +ve peak filter after a few others can give truly bizarre results. This why I have a hole around the main room mode at 70Hz. No way to EQ it out. Just have to live with it. Of course, moving the system around the room might help, but as often is the case, that is not an option.
Bob
Bob,
Thank you very much for posting your impressions so far.
In the long term, I may choose to do something similar with a nanoDIGI and I am glad I didn't pull the trigger on the miniDSP recently. I have recently come to realize that the final DAC stage (there really only should be one) is more important than I thought it was. For some time I was very happy running a HotAudio DAC-WOW. It was far better than the DAC in my netbook (no big surprise there). About a year ago I built my own DAC, by choosing separate USB->SPDIF (CM6631A) board and DAC (WM8740) board and wiring them into an enclosure. I chose the parts (and opamps) for their specific qualities. When I first heard the rig, I was quite impressed. I rolled through another handful of opamps (the usual suspects) over an number of months and have been really enjoying the AD8599 lately. Over time, I have grown accustomed to how good it sounds. Yesterday, I plugged in the old DAC-WOW, just out of curiosity. Warm, boring, small soundstage, almost no imaging at all. WOW is right. I was impressed by that at some point. Hard to believe, but true.
So, your concern about the final DAC stage in the miniDSP is not falling on deaf ears.
Thank you very much for posting your impressions so far.
In the long term, I may choose to do something similar with a nanoDIGI and I am glad I didn't pull the trigger on the miniDSP recently. I have recently come to realize that the final DAC stage (there really only should be one) is more important than I thought it was. For some time I was very happy running a HotAudio DAC-WOW. It was far better than the DAC in my netbook (no big surprise there). About a year ago I built my own DAC, by choosing separate USB->SPDIF (CM6631A) board and DAC (WM8740) board and wiring them into an enclosure. I chose the parts (and opamps) for their specific qualities. When I first heard the rig, I was quite impressed. I rolled through another handful of opamps (the usual suspects) over an number of months and have been really enjoying the AD8599 lately. Over time, I have grown accustomed to how good it sounds. Yesterday, I plugged in the old DAC-WOW, just out of curiosity. Warm, boring, small soundstage, almost no imaging at all. WOW is right. I was impressed by that at some point. Hard to believe, but true.
So, your concern about the final DAC stage in the miniDSP is not falling on deaf ears.
I have the 2x4 and am considering a second, but the miniSHARC combined with choice of DACs tagged off the I²S buss looks like a very nice option
miniSHARC Kit | MiniDSP
Bob ask for help on the electronics. I'm 100% positive there's help for the electronic anemic
miniSHARC Kit | MiniDSP
Bob ask for help on the electronics. I'm 100% positive there's help for the electronic anemic
You might be interested in my little study of how the miniDSP affects sound quality.
http://www.diyaudio.com/forums/full-range/247598-nautaloss-ref-monitor-26.html#post3808837
FWIW, as inexpensive as the miniDSP is, its DAC is quite good - I could not hear a difference when driven straight from a CD player to an amp vs going through the miniDSP, nor could I measure the difference.
I would like to try a separate DAC via the I2S bus one of these days but I suspect that I won't be able to hear a difference.
Having played with a miniDSP for a month now, I am pretty happy with how well it works. The flexibility one has in making XO's with a bi-amp is just amazing. No way will I go back to passive XO's.
http://www.diyaudio.com/forums/full-range/247598-nautaloss-ref-monitor-26.html#post3808837
FWIW, as inexpensive as the miniDSP is, its DAC is quite good - I could not hear a difference when driven straight from a CD player to an amp vs going through the miniDSP, nor could I measure the difference.
I would like to try a separate DAC via the I2S bus one of these days but I suspect that I won't be able to hear a difference.
Having played with a miniDSP for a month now, I am pretty happy with how well it works. The flexibility one has in making XO's with a bi-amp is just amazing. No way will I go back to passive XO's.
Last edited:
The 2x4 doesn't do FIR, has better sampling rate 192 vs 48 and has alot more taps, alot more. The 2x4 does low frequencies well, but FIR is the way to fix the critical mid and up. Far more flexible. Have run into the limits of what a 2x4 can do already (OB)
Want an OpenDRC or similar for preprocessing. UE can do but am leery of windoze based solutions running the show.
xrk, you should really look into DRC minimal phase filtering for your fullranges.
Want an OpenDRC or similar for preprocessing. UE can do but am leery of windoze based solutions running the show.
xrk, you should really look into DRC minimal phase filtering for your fullranges.
I am looking into linear phase and minimum phase right now. That is the discussion on the Nautaloss thread for past couple of days. I want to get either the miniShARC or openDRC but the miniDSP website is just terrible at showing you what is needed to go with what. I still can't figure out what components are needed to get a 2x4 or 2x8 FiR setup going with analog in and outs.
What I have learned in past couple of days is that FIR is not all its cracked up to be. I have been warned that it has its own evils of pre-ringing that is used to achieve linear phase. This pre-ringing shows up and sounds "electronic" for lack of a better term. I have tried and tested out the minimum phase capabability of miniDSP if one uses EQ sparingly with following guideline: cut peaks rather than boost valleys, use broad strokes, and use sparingly. I was able to get rid of a peak, fix baffle step, and boost the treble falloff while maintaining flat phase throughout full range of 200 Hz to 20kHz. See example here http://www.diyaudio.com/forums/full-range/247598-nautaloss-ref-monitor-32.html#post3813588
What I have learned in past couple of days is that FIR is not all its cracked up to be. I have been warned that it has its own evils of pre-ringing that is used to achieve linear phase. This pre-ringing shows up and sounds "electronic" for lack of a better term. I have tried and tested out the minimum phase capabability of miniDSP if one uses EQ sparingly with following guideline: cut peaks rather than boost valleys, use broad strokes, and use sparingly. I was able to get rid of a peak, fix baffle step, and boost the treble falloff while maintaining flat phase throughout full range of 200 Hz to 20kHz. See example here http://www.diyaudio.com/forums/full-range/247598-nautaloss-ref-monitor-32.html#post3813588
Last edited:
I will try to do a test with the minidsp in the chain and without with headphones so im even more susceptible to changes with headphoens and little loss of details.
I use my mini dsp in a weird way. I have two pair of output out of my dac. The first pair goes directly into my amps to drive my main speakers that do 80 hz to 22 khz. so the mini dsp is not in the critical range at all. I have my second pair of outputs that goes to the mini dsp for the 80hz 4th order low pass. the mini dsp outputs goes into my amps driving my subwoofer.
This setup sounds perfect and the mini dsp proves to be a really good way to have a subwoofer system easily without any compromises. I will try to have the mains also go thrue the mini dsp to see if theres any loss.
I use my mini dsp in a weird way. I have two pair of output out of my dac. The first pair goes directly into my amps to drive my main speakers that do 80 hz to 22 khz. so the mini dsp is not in the critical range at all. I have my second pair of outputs that goes to the mini dsp for the 80hz 4th order low pass. the mini dsp outputs goes into my amps driving my subwoofer.
This setup sounds perfect and the mini dsp proves to be a really good way to have a subwoofer system easily without any compromises. I will try to have the mains also go thrue the mini dsp to see if theres any loss.
Well I would not trust my ears but the frequency response measurement doesn't show any significant difference other than an additional single pulse overshoot on the impulse response introduced by the miniDSP. Certainly, sensitive audiophile-grade ears can hear a 50 microsecond long pressure overshoot - but if someone has objective measurements that show the stock DAC is not as good as a boutique DAC I would love to see it. I am sure it is possible, just not sure if my ears have enough resolution to hear it.
This is a little OT but related to the question of whether or not a stock DAC chip as used in the base miniDSP can be as good as a more expensive variety. There is a headphone amplifier called the O2 (Objective 2) that is based on a couple of $0.56 (56 cents each!) op amp IC's. The total cost of the parts to build this amp is about $40, and to buy fully assembled with nice case is $130. This amp was designed, built and tested (with published test results) that show that it beats many kilobuck$ audiophile grade headphone amps out there that use discrete components and fancy parts - you can see the design and test results here: NwAvGuy: O2 Headphone Amp. This headphone amp would also make a great high quality pre-amp as it can drive just about any line load. I am thinking of getting one...
The amount of engineering that goes into a high quality IC such as a DAC is staggering. Just because a component is inexpensive does not equate to inferior sound quality. A similar case can also be seen in the latest class-D amps from Texas Instruments (TPA31xxD2 series). They cost about <$1 ea in quantity, and are sold $15 to $30 fully assembled and the sound is very very good - rivaling many much more expensive amps.
I have this amp and love it.
The amount of engineering that goes into a high quality IC such as a DAC is staggering. Just because a component is inexpensive does not equate to inferior sound quality. A similar case can also be seen in the latest class-D amps from Texas Instruments (TPA31xxD2 series). They cost about <$1 ea in quantity, and are sold $15 to $30 fully assembled and the sound is very very good - rivaling many much more expensive amps.
I have this amp and love it.
I agree about cost. The DAC that I cobbled together cost me about $85 (including a handful of opamps) and it sounds fantastic. It uses the same components as much, much more expensive DACs.
I do have my doubts about the miniDSP DAC stage though. Does anyone know what chip it uses? Being "as good as a CD player" does not encourage me to buy. I have heard very few really good CD players, though.
I do have my doubts about the miniDSP DAC stage though. Does anyone know what chip it uses? Being "as good as a CD player" does not encourage me to buy. I have heard very few really good CD players, though.
P.S. Like xrk971, I am suspicious of boutique DACs (especially ones marketed by companies that also sell thousand-dollar speaker cables and interconnects). For quite some time, in fact, I was pretty sure the DAC itself really didn't matter that much. My recent (and very brief) switch back to my old DAC is what has caused the paradigm shift. The DAC that I built is so much better it is obvious not only to me, but to 3 others in my house (not audiophiles, but all 3 almost certainly have better hearing than I).
I agree about class d amp, but you have to be careful about that. If its outputs more then 20W, any class d/t will have too much switching noise.
About the odac, it sounds ok at best, but its not a miraculous dac, there much much better dac out there for like 200$.
so you dont just liten with your ears to see which one sounds better? Measurments will tell you nothing about how the two dac sounds. Thats like saying that two amps that measure the same will sound the same.Well I would not trust my ears but the frequency response measurement doesn't show any significant difference other than an additional single pulse overshoot on the impulse response introduced by the miniDSP. Certainly, sensitive audiophile-grade ears can hear a 50 microsecond long pressure overshoot - but if someone has objective measurements that show the stock DAC is not as good as a boutique DAC I would love to see it. I am sure it is possible, just not sure if my ears have enough resolution to hear it.
About the odac, it sounds ok at best, but its not a miraculous dac, there much much better dac out there for like 200$.
I answered my own question from above; the miniDSP (2x4) employs the ADAU1701 which includes 4 built-in sigma-delta DACs. Based on the datasheet, the DACs' specs appear to be "below average" (and I am being nice). The datasheet shows both passive and active output filter options, and suggests that the passive filter is "sufficient" but the active filter (including an AD8606 opamp) may provide better audio performance. The miniDSP uses the "sufficient" filter. Perhaps "good enough"?
The miniDSP 2x8, on the other hand, uses a completely different set of chips and architecture. ADAU1445 DSP --> CirrusLogic CS42528 Codec (with 8 built-in DACs) --> NJM2068 opamp.
The DACs in the CS42528 have better specs than the ones integrated into the ADAU1701, but still fall short of (even somewhat average) dedicated DAC chips. The choice of opamp is most certainly based on electrical compatibility and cost ($0.72 each) rather than on performance. Compare the NJM2068 to something like OPA2134 or AD8599 and it is like comparing a Toyota Yaris to a Porsche or a BMW.
The proof is in the pudding, however, and I would embrace the opportunity to hear either one of these units. I do remain highly skeptical, though. Perhaps more so now that I know what is inside them.
The miniDSP 2x8, on the other hand, uses a completely different set of chips and architecture. ADAU1445 DSP --> CirrusLogic CS42528 Codec (with 8 built-in DACs) --> NJM2068 opamp.
The DACs in the CS42528 have better specs than the ones integrated into the ADAU1701, but still fall short of (even somewhat average) dedicated DAC chips. The choice of opamp is most certainly based on electrical compatibility and cost ($0.72 each) rather than on performance. Compare the NJM2068 to something like OPA2134 or AD8599 and it is like comparing a Toyota Yaris to a Porsche or a BMW.
The proof is in the pudding, however, and I would embrace the opportunity to hear either one of these units. I do remain highly skeptical, though. Perhaps more so now that I know what is inside them.
Last edited:
Cogitech,
Thanks for chasing down the chip inside these things. Can you tell me how the specs compare to a "good" DAC? What figures do you look at and how important are they? I think you used to have a signature that said something along the lines of "I am glad I don't know what jitter is..." I assume you can hear what jitter is now?
For $85, the miniDSP cannot of course have high end components, but for the flexibility and power of tuning and tweaking on the fly active crossovers - I do not think anyone can fault it for sound quality. It has an I2S interface and allows plugging in a $40 Curryman DAC (ES9023). Is this a good DAC? Might be worth the upgrade.
Curryman DAC (ES9023) | MiniDSP
Thanks for chasing down the chip inside these things. Can you tell me how the specs compare to a "good" DAC? What figures do you look at and how important are they? I think you used to have a signature that said something along the lines of "I am glad I don't know what jitter is..." I assume you can hear what jitter is now?
For $85, the miniDSP cannot of course have high end components, but for the flexibility and power of tuning and tweaking on the fly active crossovers - I do not think anyone can fault it for sound quality. It has an I2S interface and allows plugging in a $40 Curryman DAC (ES9023). Is this a good DAC? Might be worth the upgrade.
Curryman DAC (ES9023) | MiniDSP
For DACs, the two specs I mostly consider are SNR and THD + N. A DAC is more than just the chip, of course, so I also consider the circuit that they are in, including output filter, opamps (are they swappable?), power supply requirements, etc.
Regarding specs:
ADAU1701: SNR = 104dB and THD+N = -90dB
WM8740: SNR = 120dB and THD+N = -104dB
A few things to consider:
1) The differences between 104dB and 120dB may seem negligible until you remember the logarithmic nature of the dB scale. You know how much louder 120dB is than 104dB. Simply apply the same principle to SNR and THD.
2) The Wolfson WM8740 can be considered a "pretty good" DAC. There are better ones, though.
3) My "old" DAC uses the WM8524, which has a built-in charge pump putting out 2v RMS, which negates the need for an opamp, which means all you're hearing is the DAC chip. It does 106dB SNR and -89dB THD. It should sound pretty good, and it does compared to a netbook DAC, but the WM8740 + AD8599 just blows it away. Honestly, it reminds me of the difference between "regular" cassettes and "chrome/metal" cassettes. It is that obvious.
4) Perhaps the DACs in ADAU1701 sound about as good as WM8524 (specs are pretty close). I was happy with the WM8524 for quite some time, so the ADAU1701 can't be terrible. Still, after hearing a better DAC, I could never go back to the WM8524.
No, I don't think I have ever heard jitter. I have heard jitter simulations, but my audio systems have never sounded like that. Perhaps my choice of quality components and software over the years has shielded me from the terrible monster named "jitter". I don't know. However, I can hear a lot of the "immeasurable" sound differences between components such as tubes and opamps. These differences become more and more apparent as I go through iterations of other components, especially drivers. Where I could hear subtle (but meaningful) differences between opamps in my DAC with the Alpair 10.2, Alpair 10P makes the differences much more obvious. With EL70 (as much as I love them) I may not be able to tell the difference between DACs, much less opamps. I should give that I try.I think you used to have a signature that said something along the lines of "I am glad I don't know what jitter is..." I assume you can hear what jitter is now?
I agree that the price for the 2x4 is good, for what you get. The convenience and tweakability maybe outweighs the sound quality issue. I also think that for less money someone could DIY something better. The ES9023 option, based on a quick look at the datasheet, would seem to be well worth $40.
Last edited:
For what it is worth, I'd be willing to bet that Bob's Topping D-20 DACs (PCM1793 + OPA2134) sound pretty darned good compared to the DACs built into the ADAU1701. What would make the Topping even better would be a proper USB interface chip (it is limited to 48kHz) and an opamp socket (rather than soldered directly to the board). Although, in Bob's setup, the USB interface is unused so that's moot for him, and there's not a whole lot to hate about the OPA2134. I guess they figured if they were going to solder in an opamp, they'd make it a good one.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.