Site Links
Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Categories
Please review the site Rules, Terms of Service, and Privacy Policy at your convenience. Rules, TOS, Privacy
Get familiar with the reaction system: Introducing the Reaction System
MiniDSP inverts output polarity with USB input?
I just find this appalling. Apparently the analog input is faithful, but the USB is not?
https://www.audiosciencereview.com/forum/index.php?threads/introducing-directiva-an-asr-open-source-platform-speaker-project.20407/page-17#post-715932
I'm obviously not up on the gear as I don't own one, but this would really upset me had I been using one.
Comments
Seems like an oversight, someone obviously wasn't paying attention.
2x4HD specifically being discussed here. Can you even use the RCA and USB input simultaneously? I think not. So the entire processing chain is inverted, no different than many amplifiers, so what's the audible error? Oversight maybe, hardly appalling IMO. I would hope that anyone using these products is in combination with acoustic measurements, so just use the invert button in the DSP program as necessary.
At least miniDSP can calculate a text book filter transfer function properly (looking at you, Dayton).
You can only use one input at a time with the 2x4HD.
Ron
I spend the greater part of the weekend trying to figure out why my latest and greatest speaker, which goes to 11, sounds distant and weak in the lower mids.
Turns out the output on channel 1 is inverted!!! 😡But not channel 2, 3 or 4.
So my woofer was in out of phase with my midrange. Argh!
this is on optical input.
(And yes, Wolf, the USB inverts all outputs.
I believe you can invert per channel in the settings. If what you see is true for all mini dsp there would be a lot of upset customers.
I just invert the outs on my 4x10 until it sounds right - took me a while but it works.
But I wasn't aware that the outs are inverted inconsistently across all channels . . . I need to take a listen again tonight.
Yup.
The MiniDSP-4x10 unit has at least channel 2 and 6 out of phase with each other and it affects the stereo mix/balance.
Channel 1 and 5 are OK, not sure about 3 and 7 or 4 and 8 yet . . .
Not good.
I looked on the Mini DSP site but see nothing related to this issue.
https://www.minidsp.com/forum/software-support/12546-180-degrees-phase-inverting
Ken, That link doesn't seem to actually address the issue raised here as far as I can tell.
If the XO's are the same for the L/R individual speaker drivers, should the polarity/phase not be the same for each rather than 180* out?
Thanks for finding something, though.
The main issue here is the output channels are not in phase with each other. I’m not talking about absolute phase, I’m talking about relative phase.
So when one is designing the crossover in an app like VituixCAD and exporting the settings/biquads to the miniDSP you expect it to be translate. You know, if your woofer or tweeter was in/out of phase with one another in the design (eg. acoustic LR2 absolutely requires this) then when you setup your miniDSP you expect the settings to translate.
Currently you have to measured your system and check. But it’s not: so you click on phase invert on the miniDSP output channel and it’s fixed. Except on the next boot or reboot or firmware update, the phase on that channel is flipped again! What the?!
It’s takes an experienced listening to tell that you have a deep null at your crossover frequency, all of a sudden, one day.
The fact that ALL channels are not in phase is a terrible problem when dealing with multi channel
I understand the downside. From what I gather the issue is in the software not the hardware. Has anyone contacted MiniDSP? This seems too simple of an issue for a company that well regarded.
TK, are you SAVING the current config before you disconnect the DSP unit?
I find that the invert button stays put once set if I SAVE the config before disconnecting.
Aside: I am not technically conversant yet on the subject to contact MiniDSP on this subject.
I just ordered (directly from the company) a 2x4hd to upgrade from my older 2x4. I've been watching the delivery status and see that it got shipped from Hong Kong to Cincinnati in a day. The next day it made it to Denver, where it will be sitting for multiple days before delivery. Guessing Denver has an extended 420...
>
At the moment I am permanently connected to my MiniDSP via USB because I use it to prototype crossovers.
I will check this save/disconnected function I’m at the moment I’m never disconnected so I don’t think that is the issue. I want to troubleshoot further before I Submit a helpdesk request from MiniDSP; I’m sure the last lots of questions and to be useful I Need to understand more
Thanks for the favor of your reply, TK.
I was trying to EQ a set of cheap speakers and XO to a sub tonight and had trouble with the MiniDSP 4x10 unit being unresponsive and found that if I disconnected it (shut down the connection & application) and relaunched/reconnected it that the issue went away and all was well.
Please try shutting down the connection and relaunching DSP 4x10 for your preset/config again before pulling your hair out.
Channel 2 and 6 are DEFINITELY out of phase with each other and it transfers to multiple speakers/configs - gotta invert one or the other in order to get a proper stereo balance.
Aside: It is annoyingly easy with this DSP unit to make cheap speakers and crappy XO's sound fantastic - yet you cannot subsequently move the speaker to another system without the DSP unit and amps while keeping the work you have done . . .
Upgraded the firmware to the latest tonight and it fixed the phase issue with one speaker preset (2 way) but not the 3 way XO speaker preset - problem remains with that one.
I suspect that the XO slopes/types have something to do with it like it does with passive XO's but don't understand why it isn't equally distributed between the same FR range drivers on the L/R channels.
Something is wrong for sure . . .
I’ve submitted a bug report and referred miniDSP to this thread.
I wonder whether this is an isolated event; or whether it has been under-detected for awhile.
To be clear; my woofer is attached to output channel 1. Midrange is channel 2 and tweeter channel 3. Channel 4: no output.
Unlike a reverse null at mid to tweeter crossover frequencies eg. at typical 2-4Khz that is more easily audible
When adeep reverse null exists at my Woofer to mid crossover, It was not detected until I system a fully system test at 1m.
At the listening positioning the sweep doesn’t show a deep null as clearly because it is swamped by the room interaction. But it is audible as a “distant” lower mids.
You’re probably thinking what the big deal? Just click on invert; and problem solved, right?
We’ll that’s true; but how can one release a crossover file for public use if one cannot even be sure it’s done correctly.
It’s the equivalent of instructing someone to wire up their passive crossover out of phase…
Deleted - Mods; remove at your discretion - Thanks.
What happened Steve?
I, for one, hope it’s a user error. You know, something boneheaded like my wires because wired incorrectly. Or the crossovers I’m designing in the software saying I should be out of phase; and this I need to flip the polarity, when I don’t. Or vice versa.
I’ve had a few responses from miniDSP and of course I have to go back and troubleshoot some more and send them more evidence of it being out of phase…
I deleted the content of a post concerning an additional issue with the 4 x 10 unit that was in fact a boneheaded oversight on my part - I cannot delete my posts though . . .
At least you are getting the attention required from MiniDSP - that is good news.
TK - Lets get on the same page for testing and figuring this out.
What is your source audio in every respect? (Audio track, delivery system, etc).
I am judging my findings mostly by a TIDAL.com account in HiFi mode with PC system controlled audio into the MiniDSP 4x10 via a usb>optical converter and playing the track by Toto-I Will Remember-Album=Tambu.
Channel 1 & 5 are sent to a powered 12"/100W Subwoofer via RCA.
Channel 2 & 6 are sent to a Class A/B Stereo 105 WPC amp > Eminence Beta 10CX speakers.
Channel 3 & 7 are sent to a Class A/B Stereo 75 WPC amp > JBL Selenium 202t Compression Drivers.
I am coming to the conclusion that DSP XO's are NOT immune from the effects of phase shifting induced by XO types and slopes.
Phase relationship of IIR filters are the same as passive filters, this is DSP 101. FIR filters are required to control amplitude and phase independently.
Verifying channel polarity is a very simple task, just run loop back measurement using any sound card or audio interface and REW or ARTA.
Thanks for the confirmation/edification, dcibel!
(you can read or be instructed all day long but never understand until you experience 1st hand . . .)
This is that time.
First I thought I was losing my mind, but the ears hear it; and measurements confirm it.
I’m not more saying much more; it may not be productive; whilst I await miniDSP’s tech team response.
When the inversions are all set = off there is nothing in the stereo mix in the center - everything is spread between the left and right channels.
When the midwoofers inversion (channels 2 & 6) are set to 2 = ON and 6 = OFF, there is a centered mix and separation to the left and right like a normal stereo mix.
My wires are not crossed-up.
It sounds fine, it just isn't as expected . . . when I play my edrums with the bass drum and snare centered in the source mix they don't sound like they are coming from the center - they sound like they are coming from the left and right speakers separately - so there is some sort of delay between the channels which I call phase issues.
Agreed Steve.
I did a factory default reset and can confirm all my outputs 1-4 are inverted when USB and coaxial are selected.
But not the stereo analog
As decibel commented the best way is just record the impulse response into an audio interface
Know if that was all I probably wouldn’t be too bothered, because as others have stated absolute polarity doesn’t make a difference.
My issue is when I have my configuration loaded for channels 2 + 3 are not inverted, but channels 1&4 are. Again yes I can just flip the polarity and all is well.
But what happens if I release my XML crossover for public consumption. How do I know that people setting up their three-way system that I’ve designed will be setting it up correctly?
This thing is definitely ready for prime time or OEMs.
Well, if they use your .XML file as you configured it then there will be no work/setup on the end users part to mess up.
Does that help?
So whilst waiting for miniDSP to respond, I pull out my old MiniDSP nanoDigi, and set it up; loaded the same biquads for equalisation, baffle step, same crossover.
And no phase inversion for nano Digi! Hmmm
That just ain't right.