Rather than adding this hugeness to the stick at the top, I fleshed the info out and decided to make it it's own post. Enjoy. 
The entire audio stack in Vista is rewritten to be largely in software user mode, for stability reasons. Yet despite this, it is indeed faster and far more robust than the old audio stack, but this has many new implications.
Nowadays, consumers have essentially have three choices for audio in Vista:
1) Creative cards
2) Non-creative cards
3) Onboard sound (Primarily Realtek and Analog Devices)
Simply, what should you use? If your primary concern is:
[*]3D Gaming Audio[/LI] - Creative X-Fi or Analog Devices onboard device.
[*]Home Theater/HTPC[/LI] - A quality, Non-Creative card or onboard solution, that supports SPDIF digital out and/or 6 discrete outputs.
[*]Music Creation[/LI] - Either a creative or non-creative card.
[*]General purpose sound[/LI] - Onboard will suit you fine.
3D Audio - Before I go on, there is something that is a common source of confusion, so let me clear it up from the start:
DS3D and OpenAL are only responsible for the *positioning* of sound.
EAX, which runs on top of either of them, provides the advanced reverb and other effects. They are two completely independent things, that are typically used in conjunction with each other.
There is now only one direct hardware path for 3D audio in Vista - OpenAL (with EAX typically on top). If you?re wondering why 3d audio doesnt work in most older games ? they use DS3D, which is no longer supported.
But there is a solution - ONLY if you have a Creative (X-Fi/Audigy) card. A wrapper named "Alchemy", which can translate DS3D calls into OpenAL calls. It works seamlessly - download it at creative's site. Again, you MUST have a creative card for this to work.
So you might be thinking "Great! Now creative no longer has a monopoly on 3d sound, since OpenAL is *open*!" Think again. OpenAL is "open" in only the most basic sense - guess who had a large part in creating it? Creative still has a monopoly on 3d gaming audio. They own all the patents for anything but the most basic positioning, and EAX v3 and up is proprietary. It's a closed system, and their monopoly is complete and uncontested. It will be a cold day in hell before there is another manufacturer that can make a card that supports advanced 3d audio.
Future games will certainly use OpenAL, but on a non-creative card, they will never be able to take advantage of anything more than basic EAX 2. And in my experience, the vast majority of non-creative cards do a particularly terrible job of *any* type of EAX.
In particular, the EAX implementation on Realtek onboard sound is *completely broken*. Occlusions and Obstructions, one of the major foundations of 3D sound, simply do not work properly. Read more about it here:
http://techreport.com/discussions.x/12970
Realtek provides the *vast majority* of onboard sound chips. If you have a Realtek chip (you probably do), and you use onboard sound for gaming - do NOT enable 3D sound or EAX under any circumstances. It will sound terrible. You may have been doing this for years and not realizing how screwed up it is!
Analog Devices are the other major supplier of onboard chips - and their EAX does indeed work properly. But like all non-creative audio, they only support up to EAX 2, which is very basic by todays standards (EAX5). Using 3D sound with these cards will not sound bad, making them a realistic choice - but the X-Fi sounds better.
Improper or incomplete support of EAX with the majority of onboard sound, and the ability to use DS3D through alchemy is the reason why an X-fi is *required* for complete 3d audio support.
If only it was that simple! Almost a year from release, the X-fi STILL has major driver issues:
Frequent loud scratchy noises that can disable audio.
Issues with suspend mode causing BSODs/resets.
Improper SPDIF support in X64
Software implementation of CMSS that hits a Core2Duo at 5%+
Issues with systems with 4GB+ RAM.
Lack of support for the new HD audio extensions
So while theyre usable (I use one), and these problems are hit or miss, its not perfect, so factor that in to your decision.
Home Theater/HTPC - You may have heard of some new, very cool audio features in vista, especially for HTPC users.
http://arstechnica.com/news.ars/post/20060907-7682.html
This post on AVS forum goes into great detail, directly from the horses mouth (Microsoft), from where many of the quotes come:
http://www.avsforum.com/avs-vb...713073&highlight=vista
However, if you have an X-fi or any other PCI card, or older onboard audio, you can?t use them. They only work with onboard ?Azalia HD Audio?. You can access them in the properties tab of your speakers in the sound CP applet.
Furthermore, they are intended for use with discrete 6-channel output. This can be either a set of PC speakers, which all use 6channel input, as well as an option on most receivers. These effects cannot be applied to SPDIF digital out, since that merely sends the compressed signal to your receiver.
The most important new feature is Room Correction:
This is HUGE. Stanalone equipment to do this job costs several thousands of dollars. I've tried it, and it works great. Its similar to the auto-calibration feature of newer receivers, although it goes one step further.
One thing you must realize - you need to turn ALL equalization, delay, gain, or any other processing OFF in your receiver, otherwise it will not work properly. And you need a halfway decent mic, either carotid (the type you sing into) or omnidirectional (much more expensive). The better the mic, the better it'll work. It wont work with a crappy headset mic, and if its marginal, Vista will let you know your mic isnt up to snuff.
For those that dont understand the mumbo jumbo:
Equalized frequency response - All speakers sound different - ideally, they would emit the same volume of sound across the entire frequency spectrum. Yet this is almost NEVER the case, especially with less expensive options. But using an equalizer, even cheap speakers can be calibrated to sound much flatter, and thus much better, but it usually requires a sound level meter, a card that has a decent equalizer, test tones, and a lot of patience. This single improvement alone will make your speakers sound MUCH better.
Time delay - Sound takes time to reach your ears, and unless all your speakers are exactly the same distance from your ears, the sound will reach you at different times, and thus have an overall negative effect on the soundstage. By adding a short delay to the closest speakers, all of the sound will reach your head at the same time.
Gain - For perfect sound, all of your speakers must also be at the exact same volume relative to your listening position, and things such as their distance, the length of wire and reflections can inhibit this.
Reflection cancellation - Ideally, you would only hear exactly what the speakers output, and nothing more. But thats never the case, because sound bounces off walls, and even worse, these walls will be in different places, throwing the sound around the room, all ending up in a mess by the time it reaches your ears. By measuring the exact amount of sound reflected, opposing sound waves can cancel this effect out.
By using specific sound patterns and a microphone situated at your listening position, it can correct all of these things, and the result is incredibly impressive in practice. I definitely suggest giving it a try for your home theater PC, perhaps even your gaming PC. This is really a killer app here.
The rest isnt nearly as earth shattering, but notable to HTPC users is the bass management, loudness equalization and the speaker fill.
Most receivers handle this internally, but as with everything else here, you want this disabled and left to be handled by Vista.
Most receivers can handle this for AC3/DTS over SPDIF, but this will apply to ALL your audio, especially important since if you're using room correction, you WONT be using SPDIF. Very useful at night.
Unlike most processing, such as Dolby Pro Logic or CMSS which takes the signal and splits amongst speakers, the Vista speaker fill is very mild in comparison - adding a slight bit to the other speakers without necessarily taking away from the original stereo source.
So to sum it all up - Vista can be a pain in the ass for 3D audio unless you have an X-Fi, it can be incredible for a proper home theater setup, and for everyone else, you'll barely notice a difference at all.
The entire audio stack in Vista is rewritten to be largely in software user mode, for stability reasons. Yet despite this, it is indeed faster and far more robust than the old audio stack, but this has many new implications.
Nowadays, consumers have essentially have three choices for audio in Vista:
1) Creative cards
2) Non-creative cards
3) Onboard sound (Primarily Realtek and Analog Devices)
Simply, what should you use? If your primary concern is:
[*]3D Gaming Audio[/LI] - Creative X-Fi or Analog Devices onboard device.
[*]Home Theater/HTPC[/LI] - A quality, Non-Creative card or onboard solution, that supports SPDIF digital out and/or 6 discrete outputs.
[*]Music Creation[/LI] - Either a creative or non-creative card.
[*]General purpose sound[/LI] - Onboard will suit you fine.
3D Audio - Before I go on, there is something that is a common source of confusion, so let me clear it up from the start:
DS3D and OpenAL are only responsible for the *positioning* of sound.
EAX, which runs on top of either of them, provides the advanced reverb and other effects. They are two completely independent things, that are typically used in conjunction with each other.
There is now only one direct hardware path for 3D audio in Vista - OpenAL (with EAX typically on top). If you?re wondering why 3d audio doesnt work in most older games ? they use DS3D, which is no longer supported.
But there is a solution - ONLY if you have a Creative (X-Fi/Audigy) card. A wrapper named "Alchemy", which can translate DS3D calls into OpenAL calls. It works seamlessly - download it at creative's site. Again, you MUST have a creative card for this to work.
So you might be thinking "Great! Now creative no longer has a monopoly on 3d sound, since OpenAL is *open*!" Think again. OpenAL is "open" in only the most basic sense - guess who had a large part in creating it? Creative still has a monopoly on 3d gaming audio. They own all the patents for anything but the most basic positioning, and EAX v3 and up is proprietary. It's a closed system, and their monopoly is complete and uncontested. It will be a cold day in hell before there is another manufacturer that can make a card that supports advanced 3d audio.
Future games will certainly use OpenAL, but on a non-creative card, they will never be able to take advantage of anything more than basic EAX 2. And in my experience, the vast majority of non-creative cards do a particularly terrible job of *any* type of EAX.
In particular, the EAX implementation on Realtek onboard sound is *completely broken*. Occlusions and Obstructions, one of the major foundations of 3D sound, simply do not work properly. Read more about it here:
http://techreport.com/discussions.x/12970
Realtek provides the *vast majority* of onboard sound chips. If you have a Realtek chip (you probably do), and you use onboard sound for gaming - do NOT enable 3D sound or EAX under any circumstances. It will sound terrible. You may have been doing this for years and not realizing how screwed up it is!
Analog Devices are the other major supplier of onboard chips - and their EAX does indeed work properly. But like all non-creative audio, they only support up to EAX 2, which is very basic by todays standards (EAX5). Using 3D sound with these cards will not sound bad, making them a realistic choice - but the X-Fi sounds better.
Improper or incomplete support of EAX with the majority of onboard sound, and the ability to use DS3D through alchemy is the reason why an X-fi is *required* for complete 3d audio support.
If only it was that simple! Almost a year from release, the X-fi STILL has major driver issues:
Frequent loud scratchy noises that can disable audio.
Issues with suspend mode causing BSODs/resets.
Improper SPDIF support in X64
Software implementation of CMSS that hits a Core2Duo at 5%+
Issues with systems with 4GB+ RAM.
Lack of support for the new HD audio extensions
So while theyre usable (I use one), and these problems are hit or miss, its not perfect, so factor that in to your decision.
Home Theater/HTPC - You may have heard of some new, very cool audio features in vista, especially for HTPC users.
http://arstechnica.com/news.ars/post/20060907-7682.html
This post on AVS forum goes into great detail, directly from the horses mouth (Microsoft), from where many of the quotes come:
http://www.avsforum.com/avs-vb...713073&highlight=vista
However, if you have an X-fi or any other PCI card, or older onboard audio, you can?t use them. They only work with onboard ?Azalia HD Audio?. You can access them in the properties tab of your speakers in the sound CP applet.
Furthermore, they are intended for use with discrete 6-channel output. This can be either a set of PC speakers, which all use 6channel input, as well as an option on most receivers. These effects cannot be applied to SPDIF digital out, since that merely sends the compressed signal to your receiver.
The most important new feature is Room Correction:
Well, in Vista, we have a perceptually tuned version of room correction. Because we don't expect everyone to own an instrumentation microphone, we have designed it so that you can use a cheap cardioid or omnidirectional microphone costing a few dollars to do an excellent job. IF you have a more expensive microphone, you can calibrate your room very accurately. All of this is implemented in in-box Vista software, using less than 5% on 3GHz CPU. Our room correction will equalize frequency response, time delay, and gain between all of your main channels, as well as build a first-reflection-cancellation filter if one or more of your speakers are near a reflective surface. If you do have a high-quality microphone, the room correction system will also flatten the overall frequency response and adjust the subwoofer delay accordingly. All of our adjustments are done as to capture the most obvious problems from the point of view of human hearing, allowing us to do a very effective job with a few machine cycles.
This is HUGE. Stanalone equipment to do this job costs several thousands of dollars. I've tried it, and it works great. Its similar to the auto-calibration feature of newer receivers, although it goes one step further.
One thing you must realize - you need to turn ALL equalization, delay, gain, or any other processing OFF in your receiver, otherwise it will not work properly. And you need a halfway decent mic, either carotid (the type you sing into) or omnidirectional (much more expensive). The better the mic, the better it'll work. It wont work with a crappy headset mic, and if its marginal, Vista will let you know your mic isnt up to snuff.
For those that dont understand the mumbo jumbo:
Equalized frequency response - All speakers sound different - ideally, they would emit the same volume of sound across the entire frequency spectrum. Yet this is almost NEVER the case, especially with less expensive options. But using an equalizer, even cheap speakers can be calibrated to sound much flatter, and thus much better, but it usually requires a sound level meter, a card that has a decent equalizer, test tones, and a lot of patience. This single improvement alone will make your speakers sound MUCH better.
Time delay - Sound takes time to reach your ears, and unless all your speakers are exactly the same distance from your ears, the sound will reach you at different times, and thus have an overall negative effect on the soundstage. By adding a short delay to the closest speakers, all of the sound will reach your head at the same time.
Gain - For perfect sound, all of your speakers must also be at the exact same volume relative to your listening position, and things such as their distance, the length of wire and reflections can inhibit this.
Reflection cancellation - Ideally, you would only hear exactly what the speakers output, and nothing more. But thats never the case, because sound bounces off walls, and even worse, these walls will be in different places, throwing the sound around the room, all ending up in a mess by the time it reaches your ears. By measuring the exact amount of sound reflected, opposing sound waves can cancel this effect out.
By using specific sound patterns and a microphone situated at your listening position, it can correct all of these things, and the result is incredibly impressive in practice. I definitely suggest giving it a try for your home theater PC, perhaps even your gaming PC. This is really a killer app here.
Beyond advanced processing like room correction, we have also added fully configurable Bass Management, Loudness Equalization, Headphone Virtualization for private headphone listening, and channel control tools to fill speakers (for example 5.1) from stereo material, as well as build phantom channels when, for example, your listening setup has no center speaker. In addition, both Headphone Virtualization as well as bass boost are available for laptops, for headphone and laptop speaker use respectively.
The rest isnt nearly as earth shattering, but notable to HTPC users is the bass management, loudness equalization and the speaker fill.
For Bass Management, we have provided the tools to set up your system with all large (bass capable), Left and Right large, or no large speakers, with or without a subwoofer. For instance, if you have small speakers and a subwoofer, the bass from all channels will be routed to the subwoofer, and if you have some large speakers and no subwoofer, the subwoofer signal will be routed to the large speakers. We also handle the case where you have some large, some small speakers, and will route the bass accordingly. You can set the crossover point of your system to whatever the loudspeakers require.
Most receivers handle this internally, but as with everything else here, you want this disabled and left to be handled by Vista.
In order to explain what Loudness Equalization does, we must first explain some terminology. In the Psychoacoustic Discipline, the term "Loudness" refers to the listener's evaluation of how loud a signal is, and the terms "intensity", "sound pressure" and the like refer to the measured, external to the listener, mechanically determined level. Loudness and intensity are related, but not in a direct fashion, because the response of the outer, middle, and inner ear must be modeled in order to relate the two. Our Loudness Equalization does exactly that, and then equalizes not the intensity, but rather the loudness, of presented signals in order to avoid blasting the listener when switching, for instance, from an analog to a digital TV channel, or from an older, dynamic recording to a modern, highly compressed recording. While this does reduce the dynamic range of the sound, this is often desirable, both in a loud environment where the quiet parts of the music are hard to hear, and in a quiet environment where you want to avoid blasting others with the loud parts of the sound.
Most receivers can handle this for AC3/DTS over SPDIF, but this will apply to ALL your audio, especially important since if you're using room correction, you WONT be using SPDIF. Very useful at night.
When we detect that the source material has fewer channels than the listening setup, we provide the ability to "fill speakers" with signal that creates an experience that is more enveloping, and that has a stronger central image and wider listening area.
Unlike most processing, such as Dolby Pro Logic or CMSS which takes the signal and splits amongst speakers, the Vista speaker fill is very mild in comparison - adding a slight bit to the other speakers without necessarily taking away from the original stereo source.
So to sum it all up - Vista can be a pain in the ass for 3D audio unless you have an X-Fi, it can be incredible for a proper home theater setup, and for everyone else, you'll barely notice a difference at all.
