AVS Forum banner
Status
Not open for further replies.
1 - 8 of 8 Posts

·
Registered
Joined
·
137 Posts
Discussion Starter · #1 ·
hello


I've been reading several articles on the nForce's APU (Audio Processing Unit) where it appears that the 420-D chipset does hardware encode 6 discrete channels to Dolby Digital. Since it has a native SPDIF out, my guess is that the encoded result is sent through the SPDIF.


I think that means we can have Dolby ProLogic-II decoded by PowerDVD-XP to 6 discrete analog channels and then, thanks to the 420-D hardware DD encoder, sent in DD 5.1 format to the receiver.


Do you think that will be the case or am I totally mistaken here ?
 

·
Registered
Joined
·
4,310 Posts
Hi PortaPro:


___I am not really good at explaining this nor if this is what you wanted but I will give it a shot.


___DPL or DPL II are multiple outputs created from just two front (R & L) channels alone and mixed for a simulated surround. DD or DTS sound tracks are individually and discretely recorded and/or created channels and are provided off the disc in their digital format as the director of the DVD/LD wants you to hear it. These are two completely different sound setups. Using DPL or DPL II created sound channels and sending it to a DD 5.1 encoding scheme in a digital format, which is what the nForce APU (Audio processing unit) or MPC as Nvidia calls it, isn’t going to work like you want nor is it probably coded properly for the APU to perform the sound processing in the first place.


___What the nForce APU is supposed to do in a game, is create the 5.1 discrete sounds as if you were interacting inside the game you are playing. The game must also be coded for the process. Imagine you are a player and you hear an enemy/player moving up from behind and you spin to engage the enemy/player with guns a blazing. You first hear the sound of the intruder behind you in a 5.1 format, you than would hear the sweep of sound as you spun and finally the sound of the intruder in front of you as your head to head firing away. This is dynamic in that the sound will change as you change position in the room (spin) as you see fit. The same can be said for the sound source (player/bot) in the room as he/she changes position (turns and runs with guns blazing toward you) or the room itself changes as you walk/run at your pace or direction through it (sound in a narrow hallway sounds different than in a large auditorium if you were to walk from one to the other). Again, this is dynamic as it should be when you interact with the surroundings in the game.


___When you watch a movie, there is nothing dynamic about it in relation to where you are viewing it from other than the action on the screen itself. You do not change position (interact with) the movie but only the display is changing and the sound recorded/played back as the director sees fit to match the scene being displayed. Neither you nor your wife will hear the scenes much differently (sitting in the back or far side does sound different than sitting in the middle of course) and this is how the sound is locked down so that no matter who watches the film, they will hear basically the same sound as you do no matter where they watch the film from because its placed on every disc exactly the same. There is no processing other than decoding the 5.1 DD/DTS signal inside of your DD/DTS capable receiver or from the SW DVD players to a 5.1 sound card via analog output to a 6 channel + amp.


___All the above was posted to show that a movies sound is locked down on the disc as it is supposed to be to match the action on the disc and the nForce’s APU should not change any of the attributes of those sound channels. It should simply pass-thru to your DD/DTS decoder what the director wanted you to hear in the first place. In a game, the nForce’s APU is supposed to calculate what the various sounds will sound like from your point of view albeit your position, direction of movement, the room around you, and the various sound sources as they move as well. This of course changes continuously.


___As far as gaming is concerned, the problem with a dynamic game sound is that there is processing going on to calculate how the sound should be output and with those calculations comes a small amount of latency. As you spin for example, the APU is calculating just a small fraction of a second behind the action what/how the sounds is supposed to be presented to you as it matches the scene on your display of choice. This small fraction of a second is sometimes noticeable in that the APU gets it wrong or is to far behind the scene and you notice. You can limit the sound algorithms complexity or input even fewer variables which reduces the life-like sound you would expect. The life-like sound processing that the nForce APU is currently capable of is ~ 1 to 2 frames behind the scenes being rendered from what I have read. This is almost not noticeable in most cases but is supposedly noticeable in some situations.


___Anyways, as far as the nForce processing a 5.1 output from an already pre-processed true 2-channel to 5.1 DPL II signal, it is likely to give you a mess if I understand your question correctly. Also, I do not have a grasp on the technology nearly as well as I should so hopefully some of the audio processing astute members can teach us both a thing or two ;)


___Good Luck


___Wayne R. Gerdes

___Hunt Club Farms Landscaping Ltd.

___ [email protected]
 

·
Registered
Joined
·
1,864 Posts
I am unfortunately no expert, but here are my thoughts anyway ;)


I have already posted a similar topic to this, and my suggestion was that you could in theory (and talk about a real mess!!) play the DTS track of a DVD and software decode it, then send it out in DD encoded format with the nForce.


Here is what I think might be involved. It depends on what the nForce will accept as input internally. It most definitely supports D3D sound in which games use an xyz coordinate system to define the exact intended position of the sound source. If this is the only thing that the nForce supports, then we have a real mess on our hands doing what we want to do.


However, if the nForce supports discrete control of each speaker, then we might have something worth using. What I mean is the ability to play "Front-Left-Speaker.wav" out the front left speaker without giving an xyz coordinate, but instead specifying that speaker. In this case, I don't see why both PortaPro's and my suggestion wouldn't work. Maybe there would be a significant delay, but in PortaPro's case with Audio CD's, who cares?


I guess we will have to wait and see. The nforce should be out in a week or two.


Mike
 

·
Registered
Joined
·
137 Posts
Discussion Starter · #4 ·
Thanks for the details xcel.


One thing that I don't understand though is why the DD 5.1 output would be that bad. In my understanding, a Dolby Digital soundtrack is composed of 6 discrete channels, the mains channels (left&Right) suffers little compression and a bit more for the other channels.


Considering, the DD compression does not degrade the original signal too much, then the DPL-II should go through without too much alteration. A 1-2 frame delay shouln't be a problem as you stated.


I consider the DPL-II decoding (from a DPL or stereo soundtrack) as beeing the Director's wish even though it wasn't intended like that, I trust the author of DPL-II :)


My goal here is just to be able to play DPL/Stereo DVDs in DPL-II mode without needing a DPL-II preamp and keeping the SPDIF single cable connection since my HTPC is 15 meters away from the preamp.


Another "crazy" dream is that, since Power DVD uses DirectShow filters, maybe we will be able to create a graph (with graphedit) to feed the DPL-II decoder with an external input source. My laserdisk collection and my cable-TV box are waiting for it :D
 

·
Registered
Joined
·
402 Posts
Quote:
Originally posted by xcel
processing that the nForce APU is currently capable of is ~ 1 to 2 frames behind the scenes being rendered from what I have read. This is almost not noticeable in most cases but is supposedly noticeable in some situations.

Everything I've read has said 70ms, wouldn't that be closer to 1/2 frame behind at 60Hz?
 

·
Registered
Joined
·
4,310 Posts
Hi Darren:


___I picked up the sound being 1 – 2 frames behind the video over at ACE’s HW in the nForce Audio Processing Unit section of the MSI K7N420 Pro review .
Quote:
According to Microsoft's data, this codec has still a latency of +/- 70 ms on the xBox. xBox games will most likely be played on a Television at 25-30 fps (or 60 fps interlaced), so each frame would be displayed about every 33 ms. The Dolby Digital sound would arrive about 1-2 frames too late, which is acceptable. So, will we see PC games with Dolby digital 5.1 too?
___Having read maybe 5 MSI nForce 420-D reviews so far, this is the first one discussing the APU in any detail and they did not even have a 5.1 receiver for some DD/DTS pass-thru via HTPC type testing. This is indeed very unfortunate for us HTPC’ers, given the ability to use the AC-3/DTS stream via S/PDIF bracket or not …


___The prices that are appearing are in the $180.00 + range on Price Watch which seems a bit high to me considering the non-OCability of the chipset/board with a separate AGP card installed. OC’ing may not be in any ones best interest of course but the amount a board can OC with a given known CPU that has made it to higher frequencies in other boards gives me an idea as to how stable the nForce board may be over the short and long term when truly stressed at the defaults. None of the MSI nForce 420-D’s I have read about have any kind of great OC’ing capability with a GeForce3 installed for example :(


___Good Luck


___Wayne R. Gerdes

___Hunt Club Farms Landscaping Ltd.

___ [email protected]
 

·
Registered
Joined
·
514 Posts
Ok, I've just scrapped a longish post where I basically agreed with xcel and thorr. I was postulating that the API would be D3D only.


But I did a quick search and found this white paper .


The paper infers that *all* sounds (2D and 3D) go into the 'Global Processor' and that the encoding is taken from the mix here.


What that doesn't tell us is whether individual channels in the GP are accessible via a wave API or not. If the GP presents itself to windows as being equivalent to a 6 channel soundcard then what PortaPro suggests may indeed be possible. But it may be that the nForce appears as only a stereo soundcard to windows with all stereo sources being mixed to front left/right channels, and that the only way to access the other channels is through D3D.


At this point it is difficult to say for sure. The white paper doesn't say that PortaPro's suggestion won't work (which was a surprise to me), but neither does it have enough detail to say that it will.


I guess we'll have to wait and see ;)
 
1 - 8 of 8 Posts
Status
Not open for further replies.
Top