Yamaha introduced Surround:AI with their previous-generation receivers. I had a RX-A3080 with me to test for a month – not provided by Yamaha. While I appreciated the clean beautiful sound of the receiver – Surround:AI did not leave me impressed. Read on to see why!
What is Surround:AI?
I have written at length about Yamaha’s Cinema DSP and why I like it. All of the modes – but especially the movie modes – are well-researched and well-implemented and they work really well in re-creating a movie-theatre-like experience to literally make the walls disappear.
Surround:AI uses the same CinemaDSP modes and tries to dynamically switch between them and dynamically configure them based on what is happening in the soundtrack.
What it also tries to do is focus on the dialogue by highlighting it from the mix and turn down the reverb / echo on the dialogue. Admittedly, this is a useful thing to do.
What’s the Issue?
While this may sound like a great idea in theory, the implementation left me feeling like the soundtrack was getting continuously altered by the algorithm in ways that I felt messed with the integrity of the original soundtrack, its tonality and steering especially.
This is in contrast to Yamaha’s normal DSP modes, such as Sci-Fi or even the new Enhanced DSP program developed for object-bases soundtracks. With those “simple” programs engaged, the soundtrack never lost the original steering or tone, but the mix was actually made clearer and its different parts more distinct – allowing it a lot more room to breathe.
Any Way to Fix It?
I actually feel like Yamaha may have overstepped the line a little with Surround:AI and I would advise they go back to the drawing board or at least give us some configuration options. For example, allow us to adjust the strength, speed and whether steering is affected. It may even help to force Surround:AI to work within the confines of one or two traditional DSP programs so it doesn’t seem so schizophrenic.
One incredibly useful feature however is the highlighting of the dialogue track by trying to separate it from the mix! If Yamaha was able to remove reverb and echo from the dialogue track while applying dialogue emphasis based on user settings for the traditional DSP modes, it would make those modes a lot more useful.
This is because the traditional DSP modes such as Sci Fi apply reverb and echo to the centre channel – and therefore the dialogue as well. It can make the dialogue harder to hear or muddier – especially in a non-treated room. Using AI to remove this muddiness – without messing with the balance of the rest of the soundtrack – would be incredibly useful!
For now, I would recommend purists – and those wanting to do critical listening – stay with the traditional DSP programs – my personal favourite still being Sci-Fi. The only time I feel Surround:AI might be of benefit is for late-night listening: it makes action more discrete and easier to follow and it does make it easier to focus on the dialogue. Of course that is all at the expense of soundtrack integrity which is not nearly ad critical for late-night listening and therefore late night processing modes.
Related Articles
Configuring YPAO – The Right Way
Recreating the Cinema Experience: Yamaha’s Cinema DSP
Get Disney+ to Sound Great with YPAO and Audyssey
Please elaborate on your final sentence, “…as it does make it easier to focus on the dialogue.”
Thank you, James. That does need further elaboration. I’ll update the article shortly. It will be earlier on in the article however!
G’day Roland, interesting take on yamaha AI, I was sceptical about it, you’ve reaffirmed this with me, I had a choice of 3070 or 3080 when I was shopping around, I went with 3070 and bought the last one on the shelves in Australia, because I was put off by the AI dsp, I’m so glad I did, the extra coin I saved went into my speakers, I put some effort into researching and setting it all up, you were a great help with your knowledge of yamaha, I’m loving my sound system, my favourite dsp is enhanced because most of my movies are 4k with atmos or dtsX sound tracks, it does such a great job, sci-fi is good too, I like it with some music, I bought yesterday on blu-ray disc The day the earth stood still 2008 version, played it last night, using surround decoder dsp, with a 5.1 sound track, my Yamaha rx-a3070 did such a good job, when the helicopters flying over in movie, the sound was directed to my 4 atmos speakers, I was amazed at how well it up scaled, even my rear speakers kicked in at some points through the movie, from a 5.1 sound track, it might not be as defined as a atmos or dtsX sound track, but it’s still impressive, I’ve limited my volume at +5 above reference, but most movies I’ll play them depending on how good the sound track is at around -6 to -12, music sounds great, if I crank it at reference or cap limit, you can hear it 3 to 4 houses down in my street lol, it’s a beast and I’m loving it.
That’s great to hear, Stephen! The 3070 is a cracking receiver! 🙂
It sounds like you did an excellent job in setting it up!
The Enhanced program is indeed excellent for object-based soundtracks.
I think AI has a place in receivers but it needs to be configurable and tackle the issues we normally have: dialogue and late night listening. It will also help to more accurately dissecting the soundtrack so the different elements can be more effectively steered using ONE sound-program. Not what they are doing, it’s a bit all over the place.
Let’s see what Yamaha does from here.
Yes, it will be, as a start point they didn’t get it so good, I’m sure yamaha will improve it over time like all their dsp’s.
This is my set up.
Yamaha rx-a3070 is the master, being a 3070 it let’s you assign atmos to pre out’s, that I have front and rear atmos powered by my old set up a kenwood krf-v5100d-s with it’s 6 channel in, it came in handy for my set up, I bought second hand for $120 in great condition another matching kenwood krf-v5100d-s to power my rear’s and side’s, the power match is spot on with 3070 per channel, leaving the power of 3070 to power front’s and centre. Front speakers are polk signature series s60. Centre, atmos, rear’s and side’s are all Accusound reference 8.6xd, my subwoofers are two accusound reference 8.6xd 200w class a/b amps and two emotiva basx s12 300w class d amps, they took a bit to get right, set up at different heights, configured as front/rear, one emotiva at front with accusound daisy changed to it, same for rear subwoofers, the subs aren’t over powered in my system though, yet well distributed low freqs, I can be anywhere in my house and it’s beautiful bass, in my prime listening position it feels like I’m sitting on a subwoofer lol, with movies it’s awesome, you can feel all kinds of low freq effects, more than enough chest slam too, it’s a Frankenstein Beast, but it sound’s awsome, I’ve done what was required for room treatment, I made myself, nothing over the top though, all I’m doing now is totally enjoying it all, no matter what I’m playing.
Do you have an Hungarian background Roland ? From a post I read, my father was Hungarian.
Sounds like you built a setup you’re really happy with which is the main thing. Enjoying it is what it’s all about. Sometimes we techies forget to do that! 🙂
I thought your surname sounded Hungarian. Yes, I’m from there originally. Left at 18 though. 🙂
OK, I’ve been to Hungary, Dad took me as a kid, good to meet Roland, your help is great 👍😉
Good to meet you too, Stephen. Hungary is great! 🙂
Thank you for your kind words and glad you’re finding the blog useful. I have a lot of things to get out for posterity so I don’t feel like I spent all this time learning just for my own benefit. So full steam ahead.
Cheers! 🙂
Nice article, definitely would be interested see what you or any of your readers think about the gen 2 surround:Ai that comes with the new A6A & A8A models.
From what I’ve read/seen online they have made improvements to make the effects more subtle and and blend much smoothly
Ah that’s cool, Alex. I haven’t heard it yet. Been busy with Audyssey and Dirac a bit more but I’ll get back to Yamaha in the new year. 🙂
If you hear anything more in the meantime, please let us know!
Actually moving to a A6A soon from my 3070 so will def see how it works, shame never had the 3080 to compare to for ai
Not sure why but not a lot of english sites go in-depth about this feature, most of the ones I read/vid I see are from asia only.
But they def say on the 3080 the ai works but you can tell when it shifts the effect , but in the newer ai it’s improved
Yes, I really disliked it on the 3080. It was very distracting. I’m glad to hear they have improved it.
I think the normal DSP programs work great. Shifting BETWEEN programs on the fly is a bit of a silly idea. Instead AI should variate the DSP itself (effect, size, etc) on the fly whine sticking to the same program but that’s just me. 🙂
You’ve considerably helped with my understanding about my 5 year old RX-A660 receiver. I’ve upgraded speakers three times now and have a decent 5.1.2 system used primarily for HT/TV. Starting to look at the RX-A4A & RX-A6A units because of AI, pre-amp capability, better DAC, better YPAO, improved MusicCast, and (of course) more power. I think that I’ll continue to use the 660 for a while until Yamaha can bring along some of the firmware changes that they have promised. Also, the supply shortages bother me a bit, afraid that some parts swapping might be going on around the world today.
I’m glad the blog has helped, James. I’m always happy to hear that! Thx! 🙂
You will definitely notice a difference when you upgrade but as you said it isn’t wise to do it until at least March 2022 when Yamaha promised to enable HDMI 2.1 features by so we’ll know if they can make good on that promise. By then chip shortages should also ease considerably. 🙂
I got the rx-a4a in August 2022. If the source has a surround track I’ll use straight mode, but if the source is 2 channel I will always use surround ai. I consider myself an audiophile and have absolutely enjoyed surround ai. It does a very good job at isolating dialog and sending that to the center channel. It also creates really nice surround signals, sometimes I forget that it’s a 2 channel signal.
Thank you, Luis. Surround:AI is supposed to be better on the new RX series so people have been taking to it better. I’m glad you’re enjoying it! 🙂