or Connect
AVS › AVS Forum › Video Components › Home Theater Computers › Audio Processing in Vista Explained
New Posts  All Forums:Forum Nav:

Audio Processing in Vista Explained - Page 9

post #241 of 971
Thread Starter 
Quote:
Originally Posted by Ozy666 View Post

The mid to high-end Denon receivers have room correction. Is the Vista correction superior to how they implement it?

http://www.twice.com/article/CA6345135.html

Ozy

I have not yet tested that in my Denon (too lazy to set it up . But since it does time domain correction, it is similar to what we do with Vista. Many others though, only have simple EQ and as such, they are not very effective.
post #242 of 971
Quote:
Originally Posted by amirm View Post

I have not yet tested that in my Denon (too lazy to set it up . But since it does time domain correction, it is similar to what we do with Vista. Many others though, only have simple EQ and as such, they are not very effective.

Can Vista take an external, multi-channel audio signal, apply the correction, and ouput the result in near-to-real time? I'm just wondering if people could use it for non-PC sources (Tivo's, HD-DVD players, etc...)

Ozy
post #243 of 971
Amir, thanks for your note and response. When Microsoft chooses to focus on solving a problem, it usually does a great job of it. Most of the consternation in these forums centers around differences in what problems need solving.

First off, using Vista effectively as a pre-amp will requires high quality DAC's in the sound card. Most sound cards sold don't, in fact the builtin motherboard audio codecs are usually quite poor when it comes to this. Even the new Intel Hidef audio standard doesn't specify high quality DAC's as a requirement for motherboard designers. Most of the people who have high quality audio systems already have recievers that have high quality DAC's in them. Hence the focus on these forums on finding a way of feeding these receivers with SPDIF in bit-perfect mode. The struggle to accomplish this in MCE is well chonicled in this and other forums for several years now. So I wish you guys would have focused on this.

Next, my Yamaha and my sister's pioneer all have digital room correction, and come with mics that support automatic room calibration and setting. You may add some functionality to what these recievers already do, but this is becoming normal in mid to high end recievers. In the vast majority of current and I will assert future MCE installations, MCE will co-exist with a reciever and other audio components. For example, cable company DVR's are common and will not replaced by MCE easily, especially given the crippled cable card support available and the cheap price the cable company provides the DVR for. The same is true for DBS boxes, and PC support for XM and Sirius also seems to have been crippled. This all means the future holds a lot of devices that will have digital audio outputs that will not go through MCE for A/V switching. This is the common case, not the case where amps are directly hooked to an MCE system.

As an engineer, I try and optimize for the common case solution, which is why I don't understand why you have focused on DRC and such and ignored bitperfect music passthrough.

I completely agree with you about recievers that correct in the digital domain. Noone I know of who cares about music quality uses one of those. Hence the reason you see things like the harmony remote control being so popular, because you have to control volume on a reciever that only has IR control available. Again, digital volume control is hardly the common case.

I am very glad to hear the internal vista mixing and conversion has been improved substantially. Still, improving something that many of us see little value in while ignoring a pressing issue is a design choice that isn't winning you a lot of kudos here. We would prefer Vista get out of the way as much as possible between our media's data stream and our receiver DACS are steadily getting getting better, as are high end sound cards for PC's. However, if I buy a reciever with high quality DACS, it is used for my cable DVR, DBS DVR, satellite radio (not that it matters much with that source), etc... If I buy a high end sound card, that doesn't help me with these other sources. Again, this is the common case today, and should be what you guys are optimizing for. Adding PC DRC etc... is a fine thing to do, but ignoring the other is the issue.

Maybe we can get a powertoy in tweakVista that could support the kind of passthrough most of us here are looking for? :-)

thanks,
Mike
post #244 of 971
Quote:
Originally Posted by MikeSM View Post

We would prefer Vista get out of the way as much as possible between our media's data stream and our receiver DACS

Amen.
Ouch


Quote:
Originally Posted by greggplummer View Post

I'm not sure I understand you guys. Are you saying the DACs in all soundcards are bad? If so, have you looked at the DACs and analog section of something like an RME Fireface or the Lynx Aurora? Even the E-MU and X-FI DACs are better than most of the ones you will find in an AVR.

Hey Gregg, we are getting there ! you exactly made *my* point. So you were first ...

To answer your first question ... YES.
I do have the Fireface800 and it sounds ... ugly. Just plain ugly. Of course this is relative to the other leage of DACs, which were just made for the best audio experience, and they cost a car ... E.g. look for the Zanden DAC (oops, tubes again). Oh, I do have an EMU as well.


Kind of fun is, that no matter a company like RME delivers a fine soundcard *they* recognize that there's more than PA stuff, so they provide the non-certified MME driver I implied earlier. This has one goal only : Bit perfect SP/DIF pass through, with the lowest latency possible. Mind you, lower than ASIO once you write a dedicated player for it, which I did.
This player alone brings you in "this other leage" (and it's bit perfect as ASIO/KS can be).

So what it comes to, is that numerous "other leage DACs" are used by people who start thinking that the PC as a transport might be better than the CD box player. And indeed, the only thing you have in mind then : please let my OS be out of the way !

Btw, if we out here refer to room correction pre/pros, we'd better start referring to the TagMcLarens et al.

Peter
post #245 of 971
Quote:
Originally Posted by amirm View Post

2. The pipeline in Vista is designed with care. Just because we say it is not bit perfect, it doesn't mean it is the same as XP. The only processing that happens when you use a single app (e.g. MCE), with volume set to 100%, is that we dither the conversion to and from internal floating point representation. Such a processing is not as harmful as you might think.

How hard would it be to add a registry setting to turn dithering on/off? That would help with the sp/dif crowd.

John
post #246 of 971
This looks to have absolutely huge potential, and especially when HDMI hits PC audio. I'm mostly interested in the response flattening.

On many high end DRC units, they have their own microphones and their own microphone amps so they know as accurately as possible how much they need to alter the frequency. Could you please explain how you have simplified this, and is it to any detriment?

Thanks
Matt
post #247 of 971
Quote:


How hard would it be to add a registry setting to turn dithering on/off? That would help with the sp/dif crowd

+1
post #248 of 971
Amir:

Most all of the Yamaha receivers have room correction built in as well (and it's fully automated).

Until HD DVD and BD software support full aspect ratio and blanking control (a la Theatertek or Zoomplayer), using HTPC for any HD media (except WMV which is supported through DirectShow, is a non starter for me and anyone else here who uses a constant height 2.35:1 screen.

Vern
post #249 of 971
Off Topic
Quote:
Originally Posted by PeterSt View Post

I do have the Fireface800 and it sounds ... ugly. Just plain ugly. Of course this is relative to the other leage of DACs, which were just made for the best audio experience, and they cost a car ... E.g. look for the Zanden DAC (oops, tubes again). Oh, I do have an EMU as well.

Have you compared the sound of your Fireface800 connected to good amplifiers to the output of a mid to high end receiver?

BTW, it seems to me, given your affection for Zanden DACs and tube equipment, that you have an audio preference for a smooth musical reproduction. As you know, that is a common characteristic of tubes. I know, I've owned several very good tube amps and preamps in the past. However, that is no longer my preference, but I respect other people's choices. So I do believe saying that the sound of the Fireface800 is ugly is pretty misleading and many people would argue that it provides a very high quality, extremely transparent sound. I've had some experience with software DSP and I believe in the future, with software, you will be able to emulate almost any sound you like. A few pro-audio software applications (my former employer's products included), do an amazing job of emulating the sound of vintage equipment such as some microphones and amps and inserting this sound as an effect into the path. These are now relatively common effects used by many audio engineers.
post #250 of 971
Quote:
Originally Posted by amirm View Post

I appreciate your point of view. But I like to make a few points:

1. I am not aware of any receivers that have built-in room correction. My Tact processor cost me $10,000. The correction in receivers is rudimentary eq which tends to aggrevate room response.

2. The pipeline in Vista is designed with care. Just because we say it is not bit perfect, it doesn't mean it is the same as XP. The only processing that happens when you use a single app (e.g. MCE), with volume set to 100%, is that we dither the conversion to and from internal floating point representation. Such a processing is not as harmful as you might think. For example, the HDCD format which encodes 20 bits in 16, does so by coding the extra resolution in the low order bits. But by randomizing it properly, it is able to actually improve the linearity of some DACs.

3. If you use any A/V receiver with volume control in digital domain, you get the same processing at Vista. You will NOT get bit perfect pass through. That is just the nature of the animal in digital domain.

4. Unless you have a very high end receiver, it is doubtful that your DACs are linear to the last bit of 16-bit samples. Most receivers are linear to 14 bits or so, when fed 16-bit signals. High-end equipment of course, can go well above this but not most systems.

Anyway, I don't mean to dispute your point. We have tough goal of satisfying everyone and sometimes that means doing 99% of what you want, rather than 100%.


These are very good points. Controlling the volume in the digital domain, as is done by most AVRs today is not as bad as many people seem to think. If the DACs SNR is very good, digital attenuation may actual sound better than adding extra analog to the signal path. The reason many audio engineers disliked digital volume control is because of the lack of dithering in some of the earlier products. Correctly dithered digital does not sound like it is losing any resolution. The key advantage of using high quality analog volume control is to improve the SNR at low levels. This is the main reason we are leaning toward adding something like the Wolfson analog VC chips to our signal path. Using other technology might be impractical in a multichannel system.

So I guess my point is Vista's audio effects may be as good and probably better than you already get with your AVR. The addition of DRC and bass management make it very attractive. AND, software can always be improved, upgraded (and with the performance of todays and future processors it will only get better), while your hardware will have to be replaced. If you paid attention to several threads about open source digital crossover and DRC projects, you will appreciate where I am coming from. And there are also some very cool commercial software products for cross overs and room correction as well.
post #251 of 971
Quote:
Originally Posted by amirm View Post

I have not yet tested that in my Denon (too lazy to set it up . But since it does time domain correction, it is similar to what we do with Vista. Many others though, only have simple EQ and as such, they are not very effective.

Can 3rd party apps, using the Exclusive mode, access the Room Correction and Bass Management features?
post #252 of 971
Thread Starter 
Quote:
Originally Posted by hellokeith View Post

Can 3rd party apps, using the Exclusive mode, access the Room Correction and Bass Management features?

You mean exclusive mode as in "don't change the bits on the way out"? If so, how would one do any of the above processing?
post #253 of 971
Quote:
Originally Posted by greggplummer View Post

BTW, it seems to me, given your affection for Zanden DACs and tube equipment, that you have an audio preference for a smooth musical reproduction. As you know, that is a common characteristic of tubes.

I just referred to the Zanden because of its price. Although I used a tubed DAC, now I don't anymore. Not because I don't like the characteristic of tubes in this part of the chain anymore, but because it "adds", in fact like you are implying.

Talking about amplifying, I think tubes (single ended) are used to allow the highest frequencies to be more accurately amplified than a solid state amp can do.
Here too, all is about a consistent whole, which btw makes me express "the other leage". For instance, where a single ended tubed amp has no feedback, all begins with speakers not needing feedback. So, in order to compare properly, you should have 112dB speakers like I have in my room.
Now tubes vs. ss has nothing to do with smooth musical reproduction (being negative) anymore ... now it just comes to representing the music "data" as accurate as possible. Note that in no way (in my system by my ears) a colouring or otherwise can be heard, when the tubes are compared with ss, but, that the highs sound less distorted.

And the funny thing is, the distortion is brought forward by my XXHighEnd player ...
Without it, there's no distortion, but also there is nothing !!! With it, there's distortion, which turns out to be an instrument once the distortion was tweaked into accuracy.
This is such a strange phenomenon, that it hardly can be explained.

Right now I use tubes for the mid-high and ss for the bass, because this gives the most accurate representation.

Now allow me to a. show myself it bit more as this person on the other planet (but we are with many overhere) and b. explain how audio can become a phenomenon that can be objectively analyzed, instead of the subjective stuff as usual :

Already a long time ago I found that room correction to the sense of reflection and standing waves is not needed at all. You "just" need to have the sound good.
For that matter, my room is 12 x 8 x 3 meters, and in fact it's a tad too large to have good music reproduction. In theory that is ...

By pure coincidence, last week I could prove it is just true that room correction isn't needed (observed by several people btw).

First it needs the reference that -no matter how I achieved that- the room is without audible reflections (at normally listening to music !) and is without audible standing waves at any point in the room.
Then it needs a big change, like changing amps and speakers, the room now being full with reflections and standing waves.
Lastly it needs the adoption of "knowing" this can be tweaked by "good sound". Good supposedly is "accurate".

In this case the speakers were provided with temporary adjustable filters, providing volume level balance between bass and mid/highs and impedance matching between all elements. And there you go ... as long as you keep on thinking all reflections etc. can be removed, and of course you know the elements in the basis are ok, then you will manage. And so I did.
Note that I'm playing well over 120dB (which is loud, I promise ) and that there's no grain of sharpness or unstandable bass waves anywhere.
Oh, during this process the speakers weren't moved, and now you can move them without damaging the sound (but with respecting the general rules of course).

Well, why this story ? just to show that I'm working in the domain of stupid accuracy, instead of having a nice smooth whatever sound to *my* ears. This is unrelated to my ears, so it would count for you and anyone just the same.

For me this implies a non oversampling solid state dac, single ended tubes to the mid/high and a solid state amp to the bass (but the latter is prone to change).
Please note that there's much more to it, like bi-amping where a main amp is acting as a preamp for the other main amp, and the dedicated speaker filtering for that.

Although it would require many pages to work it out, please respect my expression about the Fireface being "ugly", which already begins with its oversampling principle. If you'd take the time to scope the output of a os-DAC vs. a nos-DAC and observe the waves, there's no way that you will think the Fireface can be better than a nos-DAC. That -of course- the nos principle has its "negative" aspects as well, is quite another story, which all deals about how we perceive that and if we can hear that.
Of course any soundcard doesn't sound "ugly" when it's in its own leage. And of course the Fireface is waaay better compared to the mobo's sound or the EMU for that matter. Btw, I think a Lynx is better than the Fireface if it were for analogue out.

So all is in perspective, but accurate sound really is not.

Thanks.
post #254 of 971
Quote:
Originally Posted by amirm View Post

You mean exclusive mode as in "don't change the bits on the way out"? If so, how would one do any of the above processing?


EXACTLY..... this whole "bit-perfect" thing has gotten way out of proportion, to the point that most people chasing it have no idea why or if they should !!!
post #255 of 971
Quote:
Originally Posted by jimwhite View Post

EXACTLY..... this whole "bit-perfect" thing has gotten way out of proportion, to the point that most people chasing it have no idea why or if they should !!!

Well said, well spoken.

Vista is still going to give audio professionals and true dyed-in-the-wool Audiophiles the ability to purchase hardware and supporting drivers that give them bit-perfect output if they need it (and many of these folks do have valid reasons for wanting this), but for the majority of people improving the quality of audio processing within windows is what MS is and should be focused on.

My opinion is that it would be stupid for MS to waste resources on developing this... I'd rather have them put their resources into a large number of higher priority items... HD-DVD, Blu-Ray, Satellite supt in MCE, etc..
post #256 of 971
Amir,

Any chance of MS coming out with a sound card? The reason I ask is because Creative practically being a monopoly in that field has completely let support go down the drain. Their Vista drivers are non existant and we're already at RC1. Their harware makes minimal gains over long periods of time, the software blows and costs a fair amount. Some competition would be nice.
post #257 of 971
Sound card manufacturers have been on the wane since the volume of oppty has diminished due to nearly every PC shipping with sound on the MB.

I don't see a reversal in this trend occurring. The MB sound chips are slowly improving as is the quality of MB implementation.

What's really needed is to have the MS Vista logo program require a higher standard of performance in this area and then the computer makers (after complaining) would fix their implementations.

That's probably the shortest path to better quality.

jtm
post #258 of 971
Quote:
Originally Posted by hellokeith View Post

Can 3rd party apps, using the Exclusive mode, access the Room Correction and Bass Management features?

Quote:
Originally Posted by amirm View Post

You mean exclusive mode as in "don't change the bits on the way out"? If so, how would one do any of the above processing?

Exclusive, as in audio data signal which is never part of the Main or Wave mix, so that I can use different outputs on my multiple-output sound card. Audio signal uses the "Room Correction DirectX effect" (for lack of a better term), but is in a protected stream and uses a specified output, not DirectSound.
post #259 of 971
Amrim,

With the new mixing engine in vista does it make any sense to upsample at all?

I mean if you are looking to play back ripped CDs is 16b/41k the way to go, or should you consider 24b/96k?

assuming you're doing digital out to a reciever that can handle 24b/96k
post #260 of 971
Thread Starter 
Quote:
Originally Posted by sotti View Post

Amrim,

With the new mixing engine in vista does it make any sense to upsample at all?

I mean if you are looking to play back ripped CDs is 16b/41k the way to go, or should you consider 24b/96k?

assuming you're doing digital out to a reciever that can handle 24b/96k

You shoud try both to see which one sounds the best to you. There are a lot of variables here between the quality of the DACs at various sampling rate and your preference for specific sonic quality.
post #261 of 971
I think maybe this feature of Vista may be why Creative has been so slow with upgrades to the XP drivers for the X-Fi.... I smell the midnight-oil !!!

post #262 of 971
Quote:
Originally Posted by jimwhite View Post

I think maybe this feature of Vista may be why Creative has been so slow with upgrades to the XP drivers for the X-Fi.... I smell the midnight-oil !!!



Sorry to burst your bubble, but it's more likely creative is just being lazy, they've had crappy drivers with slow updates for the last 10 years, no reason to believe anything has changed.
post #263 of 971
I believe that Creative(and nVidia, and ATI) are spending 90+% of their time on new products to fully take advantage of the changes in Vista, both audio and video.

So I guess I am agreeing with both jimwhite and sotti! ;-)

They are burning the midnight oil, and they are not supporting their current customers with updated/working drivers.
post #264 of 971
I'd disagree on the video side of things a bit.

When nVidia and ATi both finally come out with "release" drivers I believe they will support older cards with at least the same functionality they provide in XP. But I'd say you are correct in that 90% of their energy is getting their dx10 cards up and running.

No back to audio
Creative probably isn't burning any oil at all.
post #265 of 971
Quote:
Originally Posted by amirm View Post

we offer the ability to separately indicate the presence or absence of the subwoofer (.1) channel, and for the user to indicate that a specific set of speaker/speakers (say side, or rear, or center) is not actually connected. When we detect that a speaker(s) is missing, we will virtualize that channel(s) so that the information in the missing speakers is still presented to the listener in an engaging fashion.

Just wondering--do you use Dolby Virtual Speaker for your speaker virtualization or your own technology?
post #266 of 971
Thread Starter 
Quote:
Originally Posted by oscarruitt View Post

Just wondering--do you use Dolby Virtual Speaker for your speaker virtualization or your own technology?

It is our own technology.
post #267 of 971
Hi Amir,

Can you give us a list of supported PCI and USB sound cards?

TIA,

Phil
post #268 of 971
Good question. I gave up on getting Vista onto my Asus P5B Deluxe-based MCE machine because the RC1 Soundmax drivers are not yet ready; Intel High Definition Audio is detected, but there is not sound out. Although the boardhas both RCA and TOSLINK outputs, only the latter is shown in the sound control panel and neither produces sound. Installing the XP drivers results instabilty.
post #269 of 971
Quote:
Originally Posted by jimwhite View Post

EXACTLY..... this whole "bit-perfect" thing has gotten way out of proportion, to the point that most people chasing it have no idea why or if they should !!!

Surely the point is that you should either get bit-perfect output and let your receiver do the processing, or let the PC do the processing (in both cases for PCM and DD/DTS).

As things stand, the PC won't do any DSP processing for DTS/DD output (which is reasonable) and it won't generate bit-perfect output for PCM.
post #270 of 971
Quote:
Originally Posted by jrpavel View Post

I gave up on getting Vista onto my Asus P5B Deluxe-based MCE machine because the RC1 Soundmax drivers are not yet ready.

Have you tried installing Vista RC1 build 5600?

This wiki lists SoundMax Intergrated Digital Audio (32 Bit) as compatible:
http://www.iexbeta.com/wiki/index.ph...st#Sound_Cards
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Home Theater Computers
AVS › AVS Forum › Video Components › Home Theater Computers › Audio Processing in Vista Explained