nvidia to release SoundStorm as add-in card

[BD2003]

As far as I know, it is a custom implementation of DS3D. Could be open AL. Whatever it is, its roughly equivalent to EAX, that being the reverb part of the positional sound. Basically, games are programmed in ds3d on the xbox, and its either sent in two speaker mode to the stereo speakers, or to 5.1 surround through the spdif cable.

Dug, I hate to be the one to say this, and although Im sure its going to fall on deaf ears, in all the time Ive been on this board I have had so few examples of someone so unilaterially, unequivocally, just plain dead wrong about something.

1) DD and EAX are compliment technologies, not exclusive technologies. --- No they are not.

Wrong. EAX is a reverb layer applied on top of DS3D. EAX is reverb, occulsions, obstruction and other stuff ONLY. DS3D handles all the positioning. DS3D works with or without a center channel.


2) The Nforce isn't just taking sound from a game and sending the same signal to each speaker. That would be like circle surround or stereo surround. This is encoded on the fly!

Yes, it is encoded on the fly. But it merely lets you connect your PC to your receiver through the one spdif wire. Nothing more and nothing less. If youre watching a DVD or playing a game encoded in AC3 (to my knowledge there are none), then it does nothing and just passes the spdif stream through. In all other instances, including 3d games, mp3s etc, all you get out of the speakers are exactly what you would get if you had a standard 5.1 PC speaker setup. NOTHING MORE, AND NOTHING LESS.

Virge is absolutely positively right. You are NEVER going to get 3d positional audio out of 2d sound, the absolute closest you will ever get to doing that at the moment is dolby prologic II. But using dolby digital prevents that.

If you think youre getting 3d sound from games that dont support it youre either:
1) Mistaken about that games support and its using 3d audio regardless, which is then encoded to the 5.1 channels.
2) Really convinced that the nforce is doing something it isnt and you either a) Are experiencing a placebo effect or b) need your ears checked.
3) You are on a hallucinogenic drug.

 

[Dug]

If you think youre getting 3d sound from games that dont support it youre either:
1) Mistaken about that games support and its using 3d audio regardless, which is then encoded to the 5.1 channels.
2) Really convinced that the nforce is doing something it isnt and you either a) Are experiencing a placebo effect or b) need your ears checked.
3) You are on a hallucinogenic drug.

You are correct on all three accounts

I need to hook up my nforce computer right now to do some testing.

 

[Dug]

Well after doing a little testing and researching this is what I have found:

EAX is a propriety method of determining occlusion, reverb and echo of the audio.... ie. merely effects, not position. DirectSound3D or openAL provide multichannel interfaces.

Every game I've played must have some direct sound positioning built in. Because without selecting EAX or 3d sound, I'm getting positional sound. (Hey just like the Xbox)

No one in the nforce forums recommends using EAX if you are using the Soundstorm outputing to a DD receiver.


I still want to know how the hell the Nforce2 DD encoding can put the sounds in just the right place. But when using analog out to a 4.1 system there's nothing unless you turn on EAX or A3D.

 

[BD2003]

Originally posted by: Dug
Well after doing a little testing and researching this is what I have found:

EAX is a propriety method of determining occlusion, reverb and echo of the audio.... ie. merely effects, not position. DirectSound3D or openAL provide multichannel interfaces.

Every game I've played must have some direct sound positioning built in. Because without selecting EAX or 3d sound, I'm getting positional sound. (Hey just like the Xbox)

No one in the nforce forums recommends using EAX if you are using the Soundstorm outputing to a DD receiver.


I still want to know how the hell the Nforce2 DD encoding can put the sounds in just the right place. But when using analog out to a 4.1 system there's nothing unless you turn on EAX or A3D.

You must be still hearing things then, because theres nothing magical happening. The DD encoding has absolutely nothing, 100% zilch zero nada to do with positioning, its just a different way of connecting your speakers. And anyone who tells you that theres any good reason not to use EAX while using DD (other than that the nforce's eax sucks to begin with) is also confused. Theyre two entirely separate things, that have absolutely no effect on one another.

The only thing I can think of is that you have dolby pro logic II overriding the DD5.1 stream on your receiver, because what youre saying is happening is pretty much against the laws of logic. If youre not using 3d sound in the game, theres absolutely NO WAY that the 3d positional information can be somehow created into it and have it be accurate.

Youre getting positional sound on the xbox because the xbox game is programmed using positional sound, not because the nforce audio is doing something magical! Think about it!

 

[Dug]

No I'm not hearing things.

Like I said everygame I've tried must have DS3D in it. Why else would I assume that nforce2 is supplying sound to all speakers? This is also why I believed the Nforce2 was doing all the work.

I never believed it was magical, in the same way ds3d is not magical, or cicrcle surround, dobly pro logic II, descrete matrix, NEO6, and any DSP listening modes that add effects. Why is it so hard to believe that the Nforce 2 could do this? It's not really. Not when you look at something like Yamaha's newest receivers suppling 80+ surround effects for 2 channel stereo. Hey I may be wrong, but its not hard to believe. Especially when you read the specs.
Ripped from Nvidia's PDF


Check out these forums if you want to see why no one recommends eax with the dd encoding.
A g
nforcehq

 

[BenSkywalker]

It's actually pretty simple, it works with audio CDs. I like the beginning portion of Zombie's SuperBeast myself

Youre getting positional sound on the xbox because the xbox game is programmed using positional sound, not because the nforce audio is doing something magical!

If you listen to a train through a single speaker set directly in front of you can you tell if it is coming at 'you' or going away? Is that magic?

 

[ViRGE]

Originally posted by: BenSkywalker
It's actually pretty simple, it works with audio CDs. I like the beginning portion of Zombie's SuperBeast myself

Youre getting positional sound on the xbox because the xbox game is programmed using positional sound, not because the nforce audio is doing something magical!

If you listen to a train through a single speaker set directly in front of you can you tell if it is coming at 'you' or going away? Is that magic?

You could hear it, but that's because of the dopplar effect. I don't really get where you're going with this line of questioning though. If he's hearing multi-channel when it's not natural to the signal, then the nForce is using some sort of algorithm to artificially create it, in which case it's "faulty" sound(or at least, more faulty than something based on EAX) because there's no real position data.

 

[BD2003]

Originally posted by: Dug
No I'm not hearing things.

Like I said everygame I've tried must have DS3D in it. Why else would I assume that nforce2 is supplying sound to all speakers? This is also why I believed the Nforce2 was doing all the work.

I never believed it was magical, in the same way ds3d is not magical, or cicrcle surround, dobly pro logic II, descrete matrix, NEO6, and any DSP listening modes that add effects. Why is it so hard to believe that the Nforce 2 could do this? It's not really. Not when you look at something like Yamaha's newest receivers suppling 80+ surround effects for 2 channel stereo. Hey I may be wrong, but its not hard to believe. Especially when you read the specs.
Ripped from Nvidia's PDF


Check out these forums if you want to see why no one recommends eax with the dd encoding.
A g
nforcehq

If its not magical, than what is it, where is it documented, and how do you enable/disable it? I'm not denying the fact that the nforce COULD in theory be possible of doing such a thing, but the fact is that it ISNT. The only changes you can make to the sound on the nforce is to add some reverb or chorus if I remember correctly, but theres absolutely no stereo separation DSP modes, such as Creatives CMSS, Dolby Pro-Logic, etc. They could program them in if they want, and Im sure the nforce can handle such a codec, but at the moment, it does not exist! In 4 channel mode, the stereo image is mirrored on the back speakers. In 5 channel mode, both channels are mixed into the center.

But you know what though, this is a waste of time. I am now assured that people will truely believe what they want to believe regardless of the facts. I just hope no one buys the nforce expecting to get the effects youre describing only to find out that you have no idea what youre talking about.

 

[Dug]

but theres absolutely no stereo separation DSP modes, such as Creatives CMSS, Dolby Pro-Logic, etc.

Actually there is. Look at the control panel.

And I'll say it again- it must be that every game I have tried with it must have DS3D and I didn't realize it. I'm not saying you are wrong.

 

[BD2003]

What is the entry called in the control panel then?

 

[Dug]

Dolby Surround Encoding.
Create center channel.
Center channel 3d pan
Creat LFE channel.

Encode MP3's into DD 5.1

Nvidia Control Panel

 

[BD2003]

Originally posted by: Dug
Dolby Surround Encoding.
Create center channel.
Center channel 3d pan
Creat LFE channel.

Encode MP3's into DD 5.1

Nvidia Control Panel

Dolby Surround encoding is akin to dolby digital encoding, it lets you get rough 3d sound on a Dolby Surround (not digital) receiver.
Create center channel merely mixes the front two channels into the center. Some people like it, some people dont.
Create channel 3d pan lets you decide whether or not you want to use the center when playing a 3d audio game.
Create LFE channel lets you decide whether or not you want bass routed to the .1 LFE channel when using DD output.

None of the above do anything even close to what you are thinking of.

Like I said, Ive been using the nforce myself for years. Id know about this feature you speak of if it existed.

 

[Dug]

None of the above do anything even close to what you are thinking of.

No, it was in answer to your comment about it not having anything like CMSS or Dolby Pro Logic.

 

[BD2003]

Originally posted by: Dug

None of the above do anything even close to what you are thinking of.

No, it was in answer to your comment about it not having anything like CMSS or Dolby Pro Logic.

It doesnt! None of those do anything close to what CMSS or Pro Logic do, which is to separate a stereo source into more than two channels.

 

[BenSkywalker]

You could hear it, but that's because of the dopplar effect.

The Dopplar effect works on all sounds of course. If you take any 2D stream that uses movement around the microphone your ears are capable of hearing movement forward and backwards due to Dopplar effects on any sound.

If he's hearing multi-channel when it's not natural to the signal, then the nForce is using some sort of algorithm to artificially create it, in which case it's "faulty" sound(or at least, more faulty than something based on EAX) because there's no real position data.

There is not precomputed positional data involved, the alterations in the sound waves themselves can be discerned.

 

[Dug]

It doesnt! None of those do anything close to what CMSS or Pro Logic do, which is to separate a stereo source into more than two channels.

What the hell do you think pro logic is doing when you enable it. Just playing 2 channels out of 2 speakers.
No. It's just like Pro logic for movies, which is a 2 channel source going to the receiver.

Do you really think all those settings are for nothing?

 

[BD2003]

Originally posted by: BenSkywalker

You could hear it, but that's because of the dopplar effect.

The Dopplar effect works on all sounds of course. If you take any 2D stream that uses movement around the microphone your ears are capable of hearing movement forward and backwards due to Dopplar effects on any sound.

If he's hearing multi-channel when it's not natural to the signal, then the nForce is using some sort of algorithm to artificially create it, in which case it's "faulty" sound(or at least, more faulty than something based on EAX) because there's no real position data.

There is not precomputed positional data involved, the alterations in the sound waves themselves can be discerned.

But this isnt real sound, its game sound. The doppler effect does not exist in game sound, unless it is calculated by some algorithm in the game (rare) or....a 3d audio codec. Wow! Wave files that when not in a 3d environment, are at most panned left to right and varied in volume. There is no alteration in the sound that positional data can even be discerned from. Do you know what kind of a momentous achievement it would be in this world to create something entirely from nothing, and have it be accurate? Believe me, I would like nothing more in the realm of sound for that to be possible, and if it were, I'd throw my audigy in the trash and switch to my onboard audio. But its not going to happen.

What the hell do you think pro logic is doing when you enable it. Just playing 2 channels out of 2 speakers.
No. It's just like Pro logic for movies, which is a 2 channel source going to the receiver.

Do you really think all those settings are for nothing?

At this point, Im convinced that benskywalker is just slightly confused and has wishful thinking, but youre just trolling. I just told you EXACTLY what every one of those settings did to the letter and you are still insisting that its doing some sort of multichannel separation. I know what pro logic does and how it works, and its an incredible technology. I wish the nforce possessed it at the very least. But it doesnt even come close, the very most it can do to a two channel source is mirror it on the rear and combine it in the center.

 

[Dug]

I give up. If you really believe the Nforce2 is doing nothing so be it.
If find it hard to believe that all the rave reviews about it are all bs.
In fact I'm suprised Nvidia even has a control panel, sense according to you, it does nothing.

I'm starting to wonder if you even used a motherboard with Soundstorm on it or just the standard MCP.

 

[BD2003]

Please then, by all means link me to a page where what you are describing is explained, because so far you have done nothing of the sort.

 

[Dug]

I don't have any links-- I have sound coming out all 6 of my speakers.

Oh, and there's nothing fancy about Dolby's surround scheme.
Its an inverted encoding scheme for the surround channels. The center channel is just the left and right combined into the center +3db.

Edit- Actually I don't have sound coming out all 6 of my speakers anymore because I don't use the Nforce board as my main system anymore. Just had to bring it back into the loop for testing so I didn't think I was imagining things.
I only get sound out of 2 speakers playing games, and all 6 for movies from my onboard digital out of the IC7.

 

[PliotronX]

Originally posted by: Dug
I don't have any links-- I have sound coming out all 6 of my speakers.

Basically the DD sends the discrete 5.1 channels so it acts like the direct 5.1 analog hookup. The positioning is done in the software, not in the NF2 I don't think. BTW I have an NF7-S and while the digital capability is excellent, the rolloff introduced at 16KHz to keep the latency down (which is still noticeable though) makes it a less than stellar overall implementation. But all arguing aside, this is great news because the sound card market is in dire need of competition. Perhaps, and I'm hoping, Nvidia can fill the void that Aureal left and since they have the only other hardware support for OpenAL besides CL's lineup, this might... or actually should help OpenAL challenge EAX as the dominating "useless reverberation feature"

 

[yhelothar]

No Dolby Digital 5.1 encoding is done by hardware. This is why the MCP-T is such a powerful chip. No way the CPU is powerful enough to handle the task of constantly encoding all sounds to DD5.1.. that would probably constantly suck up a good 10% of it...

Dolby Surround is supposed to be a form of 3d sound algorithm... I use dolby surround with my headphones and I get 3d sound with them

 

[BenSkywalker]

But this isnt real sound, its game sound.

Well, since it is just imaginary to you, I guess you can imagine it is in full Dolby Surround without a soundcard at all

Wave files that when not in a 3d environment, are at most panned left to right and varied in volume.

That's nice, how many games do you own that use .wavs? Hitman uses some for a few effects, and Incoming does. Out of the ~20GBs of games that I have on my HD those are the only ones I found that did.

There is no alteration in the sound that positional data can even be discerned from.

Actually, there is numerous data as long as you aren't talking about a 1D recording without any pretext of location being factored in to the equation. Setting up the front three speakers and the sub woofer is obviously very siplistic, all you need is to analyze the data and immediately move the bass over to the subwoofer which anything can do simplisticly. For the center channel you simply need to analyze the seperation involved in the audio stream and calculate out the relative strength of each seperate voice and then calculate out what ratio output should be utilized by the various speakers(ie- an even 50/50 split is dialed down for both channels and moved to the center, 80/20 would be muted on one speaker, halved on the center and halved on the other channel etc). All of this is extremely simplistic, it can be converted on the fly without any issue from a modestly powerful solution.

So the tricky portion is handling the rear channels. First thing you would need to do is convert the data type from a fixed function to a floating point to allow for the generation of positional data. Then you would need to take each given voice and sample it multiple times to detect movement of the object and cross reference that with the interger data for the soundstream and how it relates to the channels being utilized. By sampling the same sound multiple times and comparing it to where the 0 space for the soundstream initiates from the only time I can see serious issues cropping up is if you are talking about a directly in front off/directly behind situation(where the sound remained constant/dead center). Of course, this method would introduce latency in to the equation.