- Jul 21, 2019
Yes I would like to set up a sufficient Atmos system that would at least accommodate 4-6 people in the sweet spot. I think you would be getting into 32 channel range just for that
For example AirPods Pro support Atmos with head tracking. What would be the downside of such an implementation?
Audyssey does not create a wider sweet spot.
Maybe 7.2.4 is enough. SOTA it isn’t though. The sky is the limit there.
SOTA is not defined by how many channels or how much a system costs. It defines the QUALITY of the system, not the number of speakers. A sound system of very high quality can be had by a 7.1.4 system (there is no such thing as .2) just as easily as a 22.1.10 system.
Fair enough. So at the very least you need multi sub and probably cardioid fronts and dsp for it work in your room at anything approaching SOTA.
At least below the transition frequency we have to have systems which can properly correct for room modes.
Those issues have barely been solved in 2 channel stereo. I mean they have been (sort of) if you can spend $30k on genelec w371 or kii bxt system. They still revolve around the paradigm of “stereo bass”. Far from holistic bass.
What do you do between where subs go omnidirectional and the transition frequency?
Afaik Dirac bass control is the only one that handles proper multi sub. Without you needing to buy a mini DSP and tweak it for hours. How do you get that? $4k + $399 and there is no hdmi 2.1 yet. I wouldn’t buy anything until that is sorted out.
Maybe you can argue these issues have been solved by the beolab90 for stereo. Now where’s the solution for multichannel? 8 x $45k and 8 beolab 90s?
Since we have so many channels there would be other creative ways to tackle the problem of room modes. What a consumer grade avr can currently do is good enough but it’s not SOTA. Limited bandwidth speakers plus one or even 2 subs isn’t going to cut it. Maybe you need 4 subs 4 mid bass plus extra midbass units in the front to keep localization accurate then you can add all your extra channels.
If you tell me current state of the art is a trinnov and just 7.4.4 with 11 mains of beolab 90 or 8351/w371 or kii + bxt and 4 good subs then I would agree.
Now you say Dolby doesn’t use dsp in their demo. Apart from the fact that entire Atmos format is based on dsp you mean they don’t use dsp room correction.
Amir is right when he says the kind of stuff they have to put in there is too cheap to store even 2 target curves. Then if you keep reading you will learn it will apply phase/delay independently to each sub but peq is the same for both. Seems half baked to me. Dirac maybe fixes all this? We won’t know for years until someone actually figures out what the black box is doing. Case in point? Multeq XT.
Right and they don’t need to. You can create a large enough room that the transition frequency is lower than 80Hz. You put an architect, plus acoustic engineer together and enough real estate why bother with dsp room correction?
In a real SOTA multichannel system the room is the most valuable part of the system. The equipment just needs to put out the required SPL once the room is built and treated and monitors soffited etc.
The entire point of DSP room correction is to bend the laws of physics enough get that performance in a domestic environment. Which is not a trivial thing to do. This is why Dolby doesn’t need DSP but you do.
Audio is a 3 dimensional thing. Creating a SOTA cinema sounding experience in your domestic environment is not easy. It’s just that people can be so overjoyed by all the whizz bang sound effects to not care that the tonality is off.
"In fact you are right. It’s just sound who cares? It’s what? A music track + a dialogue track and a bunch of whizz bang effects for entertainment."
"Now to even try to answer that we must at least be knowledgeable about what’s going on in the 3D audio world."
Just like computational photography is revolutionizing the photography world and really bending old rules about light hitting the sensor the same is happening in 3D audio.
If one can get over how what they own is 95% SOTA (I’ve personally bought a lot of junk over the years with that line of thinking) then one can open their eyes to what’s out there.
The Japanese want 22.2 for their next broadcast standard. Why? Are they stupid? The standards are flying out there. MPEG.H. 6th order ambisonics.
You may refuse to believe it but we are just now seeing the tip of the iceberg in terms of computational audio. Being happy with what you own is a good thing. But having been doing this for a while it doesn’t surprise me anymore to learn that what I used to think was the best thing since sliced bread is now objectively considered crap. Things change, progress happens, stuff gets discovered.
From your perspective we already have it. Just take Trinnov 32 and make it cheaper. Fire all the other engineers involved in 3D audio. Just make them all work on Trinnov 32 for the masses.
If you mean SOTA for today then in terms of dsp room correction it would be hard to argue against the Trinnov 32. But if you SOTA in terms of cinema experience in a domestic environment there is a lot of work being done on so many facets of the problem. A few google searches will help you realize that.
Like a lot of people on this site, I am amazed that there is no really State Of The Art home theatre solution.
Even $17,000 AV processors perform far worse than a $100 DAC.
I would love to have some input on how could we do a SOTA HT.
For a stereo system it's not too hard - with AES outputs we can do all the equalisation and crossover before the DAC.
Plenty of nice DACs with >120 dB performance.
Active speakers and we're done.
For home theatre with active speakers we end up with a lot of channels if we have 7.1 or more system x 2 or 3 way.
But a harder problem is the source material, we seem more or less stuck with an AVR for all the licenses for the latest Dolby and DTS lossless CODECs.
And the onboard DAC implementation in even the most expensive is not SOTA.
And the license rules mean no accessible PCM streams we can send to be equalised, crossed-over and then DACed in a SOTA DAC.
So far the best approach seems to be to use the excellent @Weeb Labs videos to hack an AVR to extract the I2S streams, possibly convert to AES.
Once we have I2S or AES then what?
We need volume control, usually done in the AVR after the DAC AFAIK.
So a multichannel DAC with volume control, maybe the Okto 8 once it becomes re-available.
But for multi-way active speakers, can multiple units be tied to work in unison?
What about the DSP and crossover?
Does this sound like a project anyone else would be interested in?
I’ll reply to this post in more detail later.Sort of embarrassing to admit I’ve been writing in this thread and just now really read your original post.
In terms of a theoretical discussion about SOTA i would be inclined to argue the Trinnov would fit that bill. So yes you would have to hack the I2S out of it and then directly convert to AES and use something like Genelec GLM for whole system volume control. That way you keep everything digital up to the last point. You would use the DSP of the Trinnov but GLM only for volume. I’m not too familiar with the other digital monitor systems and whether they offer full system digital volume control. So this is an option.
If you feel digital volume control has no place in a SOTA system then this problem gets complicated. However I would argue this is an opinion matter. Especially when the signal is output at a 24bit depth.
SINAD to compare DACs is important to see what kind of engineering went into them and compare them. However if you think of a room with even 55dB ambient noise and let’s say a horrible SINAD of 80dB you still have output to 135db before the system’s poor SINAD becomes audible in that room. Now it would be hard to call the system SOTA with such poor SINAD but the audibility of it is debatable. Perhaps it makes up for it in terms of being exceptional in something much more audible.
Personally I would take the SINAD hit to avail of the DSP in the Trinnov. This is borderline blasphemy in audiophile circles I suppose but I think in real world use the DSP is more important. But it clearly is an imperfect situation. Why not have higher quality D/A conversion in such an expensive unit? Would it add $1-2k to its cost? About 6% extra? People in that market would pay it for piece of mind. Clearly people are hacking these to get it and thinking of solutions that would cost more than $1k to fix this. Add the value of your time to the equation and this is an expensive fix for what arguably should have been dealt with by the designers.
There are people who feel modern a/d and d/a conversion is transparent. Many of them make the music you listen to. Who knows what their chain involves and what SINAD the A/D and D/A converters they used. I wouldn’t be surprised to see multiple unnecessary conversions in there. I hope not. I like to think apart from initial A/D conversion everything is kept purely 24/192 to the end of production. I somehow doubt this is the case.
Personally I feel SOTA sound in a domestic room has a lot of issues which need to be tackled and too many with higher actual audibility. It’s easier for me to just rule out DAC SINAD (likely you would not agree) as one of those issues and just settle for what’s commercially feasible. Because audibility is questionable.
SOTA requires a level of OCD and maybe we disagree on which aspect it should be applied to. You may believe it should be focused on the DAC SINAD. I think it should be focused on using advanced DSP techniques to create a flat response (you can add your curve later) at the listening position. In a domestic room this is a huge challenge which we haven’t really gotten creative enough about.
You can have your system certified to SINAD 120dB but room modes under 300Hz plague the output. So we both feel commercial implementations are flawed for different reasons primarily yet all reasons are still valid and likely agreed upon but with different priority levels.
I suppose the DRM restrictions in play prevent a 2 box solution to improve SINAD so that is going to be out of the question. So it would involve some box within a box construction and adds let’s say $2k to the retail cost of the solution. If you want SOTA in that aspect you have to pony up for it. But the market for that kind of chassis design probably isn’t there. An AVR has to be built cheaply to sell. Most of them cost less than $2k. I would prefer that money to be spent on better DSP software research and design and I think that’s what Trinnov did. Basically a philosophical difference which not everyone would agree with. The market seems to be dictating that money be spent on the DSP software/hardware instead of chassis design to get higher SINAD.
This is why given commercial feasibility plus DRM restrictions and current correction and measurement paradigms we are far from SOTA even in the aspects we are aware of. There are aspects we may not discover until another round of listening tests get done. This is like saying there known problems which we work on and implementing solutions for at a commercially feasible cost. Known unknowns where we recognize a gap in our knowledge and we work towards filling them until they are known enough to develop solutions for. Finally unknown unknowns which we will learn about as we keep doing listening tests and psychoacoustic research.
I guess SOTA in this 3D audio in domestic space problem will always involve some compromise somewhere even if we simply deal with known problems at this point. Consider Schroeder frequency. We’ve known about that since the mid 1940s. 80 years later we still don’t have a definitive consensus solution for under 80hz, which thankfully, we are close to solving.
Now how many years to solve 80-300Hz? Thankfully modern DSP tech evolves at an exponential rate. So it won’t be 80 more years but maybe 5 hopefully. Commercial feasibility is a huge issue and that’s why I think the stage is set for multi million $ companies to pay for this kind of development pace. Companies I respect like Trinnov, Dutch and Dutch and even Genelec can’t even dream of creating the tech found in a $199 HomePod. While, in and of itself it is not satisfactory for a high end system, that concept scaled up is where we need to go. We are just getting started IMHO.
That is not true. Innovation is not limited to those mega companies. It could actually be said that smaller companies could be more nimble and focused. That is one reason why these behemoths , regularly acquire smallish companies whose innovations their billions could not come up with. Apple had to gobble Promephonic to make Apple Music compete with Spotify, for example. In spite of all its billions, Apple is not the leader in streaming space. Spotify and Netflix are. So …Companies I respect like Trinnov, Dutch and Dutch and even Genelec can’t even dream of creating the tech found in a $199 HomePod.
I’ll reply to this post in more detail later.
One thing caught my eye:
That is not true. Innovation is not limited to those mega companies. It could actually be said that smaller companies tend to ne more nimble and focused. That is one reason why these behemoths , regularly acquire smallish companies whose innovations their billions could jot come up with. Apple had to gobble Promephonic to make Apple Music compete with Spotify, for example. In spite of all its billions, Apple is not the leader in streaming space. Spotify and Netflix are. So …
Could have but wanted to point out there’s a lot going on out there.Honestly, you could have saved the 10,000-word comments and said this, and I would agree totally.