My Pluto ASP measures with the notch at 4200Hz vice 4300Hz. (Probably some component tolerance factoring into that.)
The Aura tweeters have some variability and the mounting scheme may vary a bit from user-to-user so the best solution would be to measure the peak after construction and set the notch filter accordingly. OT: The capacitor availability issue is easily surmounted.
Anyways, this demonstrates a (very) simple application for a MiniDSP setup.
Cheers,
Dave.
The Aura tweeters have some variability and the mounting scheme may vary a bit from user-to-user so the best solution would be to measure the peak after construction and set the notch filter accordingly. OT: The capacitor availability issue is easily surmounted.
Anyways, this demonstrates a (very) simple application for a MiniDSP setup.
Cheers,
Dave.
I may be wrong but I think that cascading the low & high pass filters the way Siegfried Linkwitz recommends on his site is needed only for minimum phase filters due to their inherent phase distortion. The phase distortion for linear phase filters (FIR) is removed by delay lines. The main problem with implementing Linkwitz ASP in digital domain is shelving pass filters which are very difficult to implement in DSP without losing resolution. I believe the combined approach is the best that is to implement low & high filters, delay lines as DSP and shelving pass filters in analog domain. But this will need multichannel dac .
.All discussion here so far has been around setting up biquads (IIR not FIR) which are minimum phase filters. So, cascading is still necessary.
Are shelving filters difficult to implement as IIR (biquad) or is it just the FIR implementation that is difficult?
Taking into account relatively low crossover points in Orions (120Hz & 1440 Hz so the preringing should not be to severe) I believe the FIR filters are better suited as they allow to avoid phase distortion of IIR filters. They also reduce the needs for cascading.
Shelving pass filters filters a difficult to implement in digital domain (FIR & IIR) as any signal attenuation in digital domain results in loss of resolution. It is a lesser problem for crossovers (Hi & low pass filters) as the for high attenuation for one driver will be compensated by low attenuation for the other driver).
As soon as you switch from IIR to FIR, you are no longer doing a digital implementation of the Orion ASP. You are now implementing something very different. That "phase distortion" you refer to has been designed in to (as in, accounted for within) the design of the Orion ASP. If you just switch to FIR then the phase of the drivers within, and beyond, the crossover region will not be the same as the Orion ASP. Linkwitz has carefully designed the phase response of the Orion ASP, and if you make such as drastic change to the filter topology you will need to just as carefully design the frequency response of your system in both magnitude and phase... If you just use linear phase FIR fitlers, you are making the assumption that the drivers has no phase shift, or equal phase shift at the crossover frequency.
That’s true. It will be something different. But the fact that Linkwitz used the approach doesn’t make it inherently a better one for digital solution. He worked in analog domain and had no choice but to put up with all its benefits and drawbacks. Combining digital IIR filter as in Orion ASP will results in 8 consecutive filters for the mid-range and unavoidable precision loses. It removes of course the pretty long chain of Op Amps. What is the better solution is still an open question. I think that DSP is more expensive as it will need at least 6 channel volume control (resistive or inductive) and at least 6 channel DAC if our signal source is PCM. In case of DSD it has to be converted to PCM and if the source is a LP or a tape an additional ADC has to be used. I believe the digital approach may be a more transparent solution at least for PCM. But for hi and low pass filters I would still prefer to use FIR filters.
I've lurked some of these pages for a while... I'm curious to know what the soundstage on your Plutos is like?
I wasn't able to complete the Linkwitz asp because of obsolete components a year or so ago. However, I recently picked up some minidsps
My impression thus far is that the soundstage is generally unrealistically large. This can be good or bad depending on what you are listening to. For example, on something like Pink Floyd it’s not echoy like a simulated effect, it’s just big. I feel like I'm at the concert and completely forget about the confines of my living room. Walls, ceiling, etc.. don't seem to matter. However when listening to something like Sarah Maclachlan she sounds just as big as pink floyd… it’s like just her mouth is 3 feet tall and 5 feet wide.
I tried replicating what Linkwitz offers in the Pluto manual that you purchase on his site, and I just tried what Davey offers, but nothing really seems to change it. I don't know if my tweeters aren't physically settled into the pipe properly and I'm hearing too much direction from the speaker itself, or if its a delay issue that I've tried changing, or if these things are just meant to have such a big soundstage.
I wasn't able to complete the Linkwitz asp because of obsolete components a year or so ago. However, I recently picked up some minidsps
My impression thus far is that the soundstage is generally unrealistically large. This can be good or bad depending on what you are listening to. For example, on something like Pink Floyd it’s not echoy like a simulated effect, it’s just big. I feel like I'm at the concert and completely forget about the confines of my living room. Walls, ceiling, etc.. don't seem to matter. However when listening to something like Sarah Maclachlan she sounds just as big as pink floyd… it’s like just her mouth is 3 feet tall and 5 feet wide.
I tried replicating what Linkwitz offers in the Pluto manual that you purchase on his site, and I just tried what Davey offers, but nothing really seems to change it. I don't know if my tweeters aren't physically settled into the pipe properly and I'm hearing too much direction from the speaker itself, or if its a delay issue that I've tried changing, or if these things are just meant to have such a big soundstage.
Remember that you are listening to room + Pluto's...you might try modifying the space behind the speakers until you get a soundstage that is more to your liking. Curtains and (and even traps) are an easy experiment.
I think much depends upon the recordings. I have a few Sarah Mclachlan CD's also and this type of presentation would not be unusual for many studio, pan-potted, close-miked, etc, etc, recordings that are offered nowadays. These types of recordings can't/don't convey a performance with (what SL refers to as) a realistic auditory scene since they weren't produced in that way.
Other recordings however, that capture the ambiance of the performance venue, can be reproduced with a terrific sense of space and "being there" quality. "Soundstage," if we want to call it that, can take on a realism with the Pluto's that few speaker systems can match.
Well off topic for this thread now. We should probably keep it MiniDSP related.
Cheers,
Dave.
Other recordings however, that capture the ambiance of the performance venue, can be reproduced with a terrific sense of space and "being there" quality. "Soundstage," if we want to call it that, can take on a realism with the Pluto's that few speaker systems can match.
Well off topic for this thread now.
We should probably keep it MiniDSP related.Cheers,
Dave.
I was just wondering if there were any kind of things I could try with mini-dsp to help. I see they just partnered with REW, so maybe I will have to look into room modes. I don't find the same boomy bass issue in my room that Gainphile mentioned though.
I'm also wondering if its just the aurasounds that I hear something and not the processing.
The imaging problem doesn't matter if its Eva Cassidy, Peggy Lee, Diana Krall, Nina Simone, Ash Koley, Sarah, etc. etc. they all have the same issue.
I'm comparing to a pair of Ascend Sierras, and some in-the-works open baffles I'm trying out. The ascends have a very defined center image, and excellent width... the open baffles aren't quite as defined as the ascends unless you get them in the perfect spot, but have more of that live feel of the Plutos.
I don't know... I guess I'll keep playing around.. it will be interesting to read more on how this Orion project comes out.
I'm also wondering if its just the aurasounds that I hear something and not the processing.
The imaging problem doesn't matter if its Eva Cassidy, Peggy Lee, Diana Krall, Nina Simone, Ash Koley, Sarah, etc. etc. they all have the same issue.
I'm comparing to a pair of Ascend Sierras, and some in-the-works open baffles I'm trying out. The ascends have a very defined center image, and excellent width... the open baffles aren't quite as defined as the ascends unless you get them in the perfect spot, but have more of that live feel of the Plutos.
I don't know... I guess I'll keep playing around.. it will be interesting to read more on how this Orion project comes out.
Actually, you don't really need analog-domain volume controls, as shaped dither is quite transparent if it's applied at well over the audible range. Oversampling the signal before dither helps tremendously: I had a 6 bit signal almost indistinguishable from the original when dither/truncation was applied after resampling at 2x.
I'm not sure about oversampling in this case as it is not going to recover the lost data. Dithering helps to decorelate quantization noise introduced by lowering the resolution of the signal (it converts it essentially into noise), but it will not recover lost bits either.
You don't necessarily "lose bits" with DSP processing. I don't know about the Mini-DSP, but in some other DSPs I have used, the internal datapath is so wide that you simply will not "lose bits" due to processing. In the TI TAS3103, the internal datapath is entirely 48 bit. At the output, you can do up to 32 bit with dithering. The incoming audio is packed into the 48 bit input stream with 8 bits for overhead (allowing for gain in your EQ/filters without overflow) and nominally 40 bits for the data. That pushes the noise floor of the DSP to 20 bits (120 dB) below that of 16 bit CD, and there is an extra 48 dB of overhead. The 32 bit dithered output ensures that the noise floor of the digital signal output is well below that of any analog electronics that follow it. If you think you will "lose bits" or resolution with this system (properly used) you are quite mistaken.
Maybe someone who knows the internal details of the Mini-DSP can speak up about its resolution.
Maybe someone who knows the internal details of the Mini-DSP can speak up about its resolution.
I have been optimist in regards to the digital attenuation for level matching between channels. After all I cannot hear any degradation if I use my PC to attenuate signal (ie. headphone listening).
But when I implemented this with MiniDSP, clearly there is audible distortion. I have since put all the attenuation to maximum (0 db) and soldered pots at the output stages to control them analog-ly.
See here:
MiniDSP
Meanwhile, I'm finished my speakers and amps so I will have time soon to continue this learning effort to replicate Orion ASP.
But when I implemented this with MiniDSP, clearly there is audible distortion. I have since put all the attenuation to maximum (0 db) and soldered pots at the output stages to control them analog-ly.
See here:
MiniDSP
Meanwhile, I'm finished my speakers and amps so I will have time soon to continue this learning effort to replicate Orion ASP.
Internaly DSP can process the data with any precision. But a the and it has to downsample to number of bits a DAC can accept. And all parts of the output signal where the level is less than Odb some precision will be lost. The magnitude of the loses depends mostly on the level of attenuation of the signal.
Do you have any measurements showing said distortion? What input signal levels were you using? Any other details?
Dave.
Hi Gainphile - thanks for all this info - it is really encouraging. I'm on a similar path, and it really has its ups and downs...
I'd love to know more about the problem with the digital attenuation, as it looks like such a nice option. I've read your other two threads on it. I wonder if anyone else is having such problems...
I'd love to know more about the problem with the digital attenuation, as it looks like such a nice option. I've read your other two threads on it. I wonder if anyone else is having such problems...
Tell me how you're measuring that THD increase on your MiniDSP.
I've performed a bunch of distortion tests on my MiniDSP's with varying levels of attenuation using a 10k system volume control and/or with programmed attenuations applied to individual outputs. I don't see any increase in distortion.
Cheers,
Dave.
I've performed a bunch of distortion tests on my MiniDSP's with varying levels of attenuation using a 10k system volume control and/or with programmed attenuations applied to individual outputs. I don't see any increase in distortion.
Cheers,
Dave.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.