miniDSP Tide16 - Holy Grail with 16 Channel Atmos/DTS:X, high SINAD

[Ultrasonic]

- let's say a +6 dB gain PEQ will not budge such a dip of -6 dB at 90 Hz

A 6 dB boost will fill in a - 6dB dip.

Where talk of boosting making no difference comes from is more of a theoretical situation where you get complete cancellation of direct and reflected sound waves. Generally in-room you won't get such complete cancellation though, as you get many reflected signals contributing, of varying amplitude. Provided there isn't actually complete cancellation, any boost applied will reduce the dip.

Whether it's wise to boost is instead a question of system headroom and distortion levels. But as you alluded to, boosting by 3 dB is exactly like increasing the volume level by 3 dB at the relevant frequency.

 

[Ultrasonic]

I would guess that it’s because the dip won’t be a fixed 6db at every output level.

It would be. The acoustics that causes the dips is essentially a linear process.

 

[Keith_W]

The depth of the dip very much depends on the smoothing used when displaying. Due to this, is appera as if the dip moves with the gain, but in reality, it does not.

@peng please read this again. I saw that you replied, but I don't think you understood what @voodooless was saying. I don't know if I can put it any clearer, but i'll try: the depth of the dip is an artefact of the smoothing algorithm. If you measure at a louder SPL and choose the same smoothing, the dip will be at the same depth relative to the rest of the signal, but at an apparent higher level relative to 0dB.

Also consider that not all dips are full nulls. But I would not try to boost them anyway, unless they are like 3db down without much smoothing.

And this is another important point. A true dip, or a zero outside the unit circle, is caused by two sound sources that are perfectly 180deg out-of-phase. Inversion of this dip produces a pole - i.e. infinite power is required. But if two sound sources are 150deg out-of-phase, it will still produce a dip. But because this is close to the zero but not at the location of the zero, inversion does not produce a pole. The closer the location is to the zero, the closer the inversion will approach a pole. In other words, a heck of a lot of power is still required, but not infinite power.

 

[peng]

@peng please read this again. I saw that you replied, but I don't think you understood what @voodooless was saying. I don't know if I can put it any clearer, but i'll try: the depth of the dip is an artefact of the smoothing algorithm. If you measure at a louder SPL and choose the same smoothing, the dip will be at the same depth relative to the rest of the signal, but at an apparent higher level relative to 0dB.

I meant to stop sidetracking this thread but may be just one more time to clarify, I think I understand his point as it was in simple English, but in this case it has nothing to do with smoothing, if as you said the dip would be at the same depth relative to the rest of the signal but at an apparent higher level relative to 0 dB, then the dip would still be as deep as before, that's not what I saw in my graphs in which no smoothing was applied anyway.

And this is another important point. A true dip, or a zero outside the unit circle, is caused by two sound sources that are perfectly 180deg out-of-phase. Inversion of this dip produces a pole - i.e. infinite power is required. But if two sound sources are 150deg out-of-phase, it will still produce a dip. But because this is close to the zero but not at the location of the zero, inversion does not produce a pole. The closer the location is to the zero, the closer the inversion will approach a pole. In other words, a heck of a lot of power is still required, but not infinite power.

Also was not the case in the example I used, I knew those dips were room mode related, true enough that they were likely not 180 deg out "poles", nevertheless they had no to little effects if I simply boost those points at the specific frequency points. I better really stop now before the mod gets mad, thanks again for trying, as I said in the beginning, it was my stupid question, so I may be too stupid in this case to understand something lol..

 

[thcdru2k]

I apologize for opening a can of worms. Boosting a single driver is fine if done conservatively. I personally avoid it because I am very conservative, so if I gave the impression it should always be avoided that was not my intention. I think this thread can be steered back toward the Tide16. It is important to highlight how something like Tide16 or any reasonably flexible DSP can be used together with REW to get a proper acoustic timing reference. Once you do that and look at REW spectrograms the key line really jumps out: “The dashed line is the Peak Energy Time trace… an ideal PET would be a straight line with the same time value for all frequencies.” Check out the 3rd paragraph here: https://www.roomeqwizard.com/help/help_en-GB/html/graph_spectrogram.html

Before that I used to EQ things to make a pretty SPL graph, and when I went back and looked at my saved mdat files I was shocked at my PET line. When I did auto calibration and then measured different room correction schemes with a UMIK and full range sweeps of multiple drivers and subwoofers at the same time, I noticed that the PET line was not behaving the way REW describes in the handbook.

I would be careful with room correction workflows that never really establish a clean timing baseline for PET in the first place. In those cases you can still use PET within a single measurement, but it is harder to trust it as a precise guide for alignment or to compare between configurations, so people fall back to looking only at SPL. Boosting a single driver in isolation is one thing, but once you are measuring multiple drivers and subs together, adding boost without a reliable timing reference can easily change the vector sum and arrival time in ways you do not intend. Tide16 together with REW makes it much easier to treat time alignment as a first class target instead of an accidental side effect of EQ.

This is as straight as I have been able to get the PET line with a Flex HTx. All I did was follow the REW guide here https://www.roomeqwizard.com/help/help_en-GB/html/makingmeasurements.html and use the delta T information to adjust delay in the miniDSP device console. Edit: Sorry I forgot I also had to align subs different because acoustic timing reference doesn't really work for subwooofers and speakers.

 

[Mort]

may be just one more time to clarify

It's an important point and one worth it's own thread because there is subtlety. There are a cadre of 'never boosters' that underestimate the efficacy. GLM added small boosts to it's 5.0 software according to their materials. They too appear to have converted.

I'm not suggesting one can simply eq boost out of room nulls, only that it's more nuanced.

 

[thcdru2k]

It's an important point and one worth it's own thread because there is subtlety. There are a cadre of 'never boosters' that underestimate the efficacy. GLM added small boosts to it's 5.0 software according to their materials. They too appear to have converted.

I'm not suggesting one can simply eq boost out of room nulls, only that it's more nuanced.

I may be wrong but GLM is also extremely high quality build and expensive to boot..not sure if other speakers / drivers / subs would hold up to something like that.

 

[EWL5]

I apologize for opening a can of worms. Boosting a single driver is fine if done conservatively. I personally avoid it because I am very conservative, so if I gave the impression it should always be avoided that was not my intention. I think this thread can be steered back toward the Tide16. It is important to highlight how something like Tide16 or any reasonably flexible DSP can be used together with REW to get a proper acoustic timing reference. Once you do that and look at REW spectrograms the key line really jumps out: “The dashed line is the Peak Energy Time trace… an ideal PET would be a straight line with the same time value for all frequencies.” Check out the 3rd paragraph here: https://www.roomeqwizard.com/help/help_en-GB/html/graph_spectrogram.html

Before that I used to EQ things to make a pretty SPL graph, and when I went back and looked at my saved mdat files I was shocked at my PET line. When I did auto calibration and then measured different room correction schemes with a UMIK and full range sweeps of multiple drivers and subwoofers at the same time, I noticed that the PET line was not behaving the way REW describes in the handbook.

I would be careful with room correction workflows that never really establish a clean timing baseline for PET in the first place. In those cases you can still use PET within a single measurement, but it is harder to trust it as a precise guide for alignment or to compare between configurations, so people fall back to looking only at SPL. Boosting a single driver in isolation is one thing, but once you are measuring multiple drivers and subs together, adding boost without a reliable timing reference can easily change the vector sum and arrival time in ways you do not intend. Tide16 together with REW makes it much easier to treat time alignment as a first class target instead of an accidental side effect of EQ.

This is as straight as I have been able to get the PET line with a Flex HTx. All I did was follow the REW guide here https://www.roomeqwizard.com/help/help_en-GB/html/makingmeasurements.html and use the delta T information to adjust delay in the miniDSP device console. Edit: Sorry I forgot I also had to align subs different because acoustic timing reference doesn't really work for subwooofers and speakers.

I think the discussion is relevant and someone needs to educate me here: doesn't Dirac ART (which is included on the Tide16) take care of the dips/nulls in the curves? I've seen so many contrasting before/after waterfall graphs by ART users that no one can say "it doesn't work". What am I missing here?

 

[thcdru2k]

I think the discussion is relevant and someone needs to educate me here: doesn't Dirac ART (which is included on the Tide16) take care of the dips/nulls in the curves? I've seen so many contrasting before/after waterfall graphs by ART users that no one can say "it doesn't work". What am I missing here?

These are not my graphs I pulled this from google i just typed dirac art spectrogram . Dirac art has the best PET line but all three kind of suffer in the bass region. I'm not one hundred percent sure but I believe its because trying to time align subs to mains with room correction is not easy.

 

[Lsc]

I remember a time when MiniDSP made affordable stuff. I still love my MiniDSP 2x4HD, but if they'd update it to take multichannel PCM, added 2-4 RCA outputs (I have no need for XLR), variable loudness and a web interface it would be even better. But instead they make this monstrosity that at least I will never ever buy.
Ah well, great for those that need and can afford this!

If the Tide16 performs as well as expected then the price seems very reasonable considering it will be a preamp processor for HT up to 16 channels (I only need 5.3) and with ART may improve 2 channel sound over a dedicated 2 channel system (with no room correction) as well. This last part is what I’m really interested in because my 2 channel (actually 2.2) sound is dialed in pretty well.

We will learn more when this is out in the wild and more information is released.

Also, since I’m mainly looking for a preamp processor. $3500 is better than Marantz’s offerings where ART is extra cost.

 

[thcdru2k]

I went back through my old mdats. I remember I had an interesting sweep because on Flex HTx you can route a signal to all speakers and subs at same time. This was a frequency sweep of my entire system. 5 speakers and 2 subs. This was an old calibration so there were some errors and is not my current system state. This was probably 3-4 manual calibration sessions ago but even than I was able to get a very controlled PET line even in lower frequencies and even with an extreme situation of all speakers and subs firing at same time. This is why I think Tide 16 can be a game changer; mainly because I want to get rid of my AVR at some point. I only keep it right now for Atmos / DTS:X decoding / 4k 120hz and easier for my wife to use honestly. All bass management is handled by Flex HTx

 

[Keith_W]

I think the discussion is relevant and someone needs to educate me here: doesn't Dirac ART (which is included on the Tide16) take care of the dips/nulls in the curves? I've seen so many contrasting before/after waterfall graphs by ART users that no one can say "it doesn't work". What am I missing here?

Dirac ART DOES work. It seems to shorten the decay and flatten the frequency response, but at the expense of a non-ideal step response. I suspect that ART is playing support speakers before the main speakers to help fill in the dips, but until I gain access to an ART system and take measurements of my own, I am only speculating. How well Dirac ART would work of course depends on your system.

 

[Lsc]

I think the discussion is relevant and someone needs to educate me here: doesn't Dirac ART (which is included on the Tide16) take care of the dips/nulls in the curves? I've seen so many contrasting before/after waterfall graphs by ART users that no one can say "it doesn't work". What am I missing here?

Yes I agree.

I think the basic Dirac Live is trying to do this at an individual speaker level including subs (subs are just another speaker).

Next DLBC improves this by adding a proper bass control where a crossover can be defined.

Then ART takes this further by impacting individual speaker responses with “helper” speakers to do exactly what your point is…be able to improve the speaker response to a level that’s not possible by trying eq the speaker on its own.

Of course the effectiveness of Dirac can be questioned but that’s the intent. But based on what I read so far, since miniDSP’s target audience are folks who like to do their own adjustments that all of that will be available in the Tide16 to give users choices. I think ART will be all I need but it’s good to know that other tools will be available.

Do the other miniDSP have flexible crossovers in the digital domain that essentially replicates what I’m doing with my analog JL CR-1 active crossover? I guess I can start reading up on this - I have zero experience with miniDSP except for the UMIK-1 .

EDIT: ok I read up on some miniDSP products. Now I understand why the Tide16 is such a departure. I didn’t realize how inexpensive the other products were.

 

[Ultrasonic]

Do the other miniDSP have flexible crossovers in the digital domain that essentially replicates what I’m doing with my analog JL CR-1 active crossover? I guess I can start reading up on this - I have zero experience with miniDSP except for the UMIK-1 .

Any miniDSP is way more powerful in terms of signal processing capability than a JL CR-1. Most important for subwoofer integration is the ability to apply time-delays to the main speakers.

 

[EWL5]

EDIT: ok I read up on some miniDSP products. Now I understand why the Tide16 is such a departure. I didn’t realize how inexpensive the other products were.

It's definitely a big gamble for miniDSP. Only time will tell if the Tide16 will sell or will the hardcore miniDSP users be turned off by how much control Dirac needs vs what the user wants to do. Although this is still up in the air, Dirac traditionally locks out quite a bit of settings to the end user.

 

[thcdru2k]

Dirac ART DOES work. It seems to shorten the decay and flatten the frequency response, but at the expense of a non-ideal step response. I suspect that ART is playing support speakers before the main speakers to help fill in the dips, but until I gain access to an ART system and take measurements of my own, I am only speculating. How well Dirac ART would work of course depends on your system.

Not saying frequency response or step response is better... but I think in general a clean frequency grabs people's attention faster. It's the quickest fastest way people compare systems. Its the first tab of REW. But what is truly better? I'm not sure myself. I only have theories and experience with my own system. Listening is also so subject to bias as well. I just try to look through the REW handbook and absorb as much as I can because they definitely know more than I would ever know.

 

[Keith_W]

Not saying frequency response or step response is better... but I think in general a clean frequency grabs people's attention faster. It's the quickest fastest way people compare systems. Its the first tab of REW. But what is truly better? I'm not sure myself.

Lol, i'm not sure myself either! To answer the question - "is the non-ideal appearance of this measurement audible?" requires an understanding of psychoacoustics. Or if you don't know psychoacoustics, you need a lot of experience looking at measurements and correlating it with listening. I know some psychoacoustics, but I wouldn't say my knowledge is in-depth or authoritative.

Coming back to that non-ideal step response with ART - is it audible? Well, my current understanding suggests that the pre-masking threshold is about 20ms, although this varies with frequency and SPL. I don't know what the thresholds are, but the "pre-ring" component of ART is much greater than the 20ms commonly cited. So I think it should be audible. Yet a lot of ART people don't hear it. So either all those ART people have collective bad hearing, or my understanding is inadequate. I think the latter is more likely.

With measurements, you can go for an ideal response. An ideal response is guaranteed to be free of audible flaws. But you could also go for a psychoacoustically inaudible response. This buys you more freedom and more flexibility, but you really need to know what the thresholds are. Both the FR and step can have deviations from non-ideal that are inaudible. But what those limits are is another question entirely. It's just like high bitrate MP3's. You KNOW that an MP3 is not a perfect reproduction of the original. It is measurably inferior. But is it audible? Usually the answer is no.

 

[Oddball]

It's definitely a big gamble for miniDSP. Only time will tell if the Tide16 will sell or will the hardcore miniDSP users be turned off by how much control Dirac needs vs what the user wants to do. Although this is still up in the air, Dirac traditionally locks out quite a bit of settings to the end user.

Sorry, not following this thread as grew too fast and not much time right now. Are you saying that there will be no option to completely turn off Dirac and use just DSP if one would want so? True that Dirac locks up a lots of settings, but HTP-1 and Storm still have PEQ before Dirac and tone controls after Dirac, so some level of pre and after processing should be possible.

@Keith_W - from some graphs and settings by users posted around forums I noticed that pre-ringing is generally lower when lower support levels are engaged. Lots of people still use -18dB which is about 75% of max support levels. Not sure when will have time to play with that though. Some members also found lesser number of cross support groups also contribute to better alignment and clarity. Or just might be system and room specific. I get a fair amount of pre-ringing on some graphs, and more than 20ms in most of them, but still the best overall sound experience I ever had in my system and difficult room.

 

[thcdru2k]

Lol, i'm not sure myself either! To answer the question - "is the non-ideal appearance of this measurement audible?" requires an understanding of psychoacoustics. Or if you don't know psychoacoustics, you need a lot of experience looking at measurements and correlating it with listening. I know some psychoacoustics, but I wouldn't say my knowledge is in-depth or authoritative.

Coming back to that non-ideal step response with ART - is it audible? Well, my current understanding suggests that the pre-masking threshold is about 20ms, although this varies with frequency and SPL. I don't know what the thresholds are, but the "pre-ring" component of ART is much greater than the 20ms commonly cited. So I think it should be audible. Yet a lot of ART people don't hear it. So either all those ART people have collective bad hearing, or my understanding is inadequate. I think the latter is more likely.

With measurements, you can go for an ideal response. An ideal response is guaranteed to be free of audible flaws. But you could also go for a psychoacoustically inaudible response. This buys you more freedom and more flexibility, but you really need to know what the thresholds are. Both the FR and step can have deviations from non-ideal that are inaudible. But what those limits are is another question entirely. It's just like high bitrate MP3's. You KNOW that an MP3 is not a perfect reproduction of the original. It is measurably inferior. But is it audible? Usually the answer is no.

Yeah obviously Dirac ART is audibly excellent. There is no way it is not. It's not possible for hundreds of people to have some kind of consumer bias; at least I would hope so. I've only experience Dirac Live in my system (via Flex) and if you blind folded me and compared to my manual calibration I do not know if I could honestly pick one over the other. When I purchased a second subwoofer ; I said oh it must sound better. When I listen to more expensive speakers, I'm already thinking in my head this is going to sound great. Even just little thoughts like that perhaps alters my experience. I agree measurements are nice but I ran into that same issue where I had measurements from older calibrations that were much nicer but I think they did not sound as good as my system now but than again I cannot confirm I am not biased because I spent more money on this or I spent more money on that. Or even just the hey I spent 6 hours getting a smooth spectrogram it must sound better bias. So I can't even really trust myself... I have my own bias because I love calibration. However there have been plenty of times I thought the system sounded great and my wife says hey it's booming too much. When I go back to my old mdats to take a look I'll see that oh maybe the FR in the bass region is a little bit too much and I didn't level match properly.

 

[EWL5]

Sorry, not following this thread as grew too fast and not much time right now. Are you saying that there will be no option to completely turn off Dirac and use just DSP if one would want so? True that Dirac locks up a lots of settings, but HTP-1 and Storm still have PEQ before Dirac and tone controls after Dirac, so some level of pre and after processing should be possible.

That's why I said it's "up in the air" as no one will know until next month when hands are on devices.

You would think there would be fewer Dirac settings locked out due to the miniDSP heritage of DIY but they also may just be as restrictive as the D&M implementation since mainstream users need a lot of hand holding and more likely to "break" something than not!