by Sergey Samarin
11/04/2003 | 09:21 PM
Our first review of existing integrated audio solutions had a great feedback (I guess my colleagues who review mainboards should pay more attention to their audio subsystems :). I would like to express my appreciation to all readers who took pains of emailing me their responses and suggestions. Some topics I am going to discuss in this article have been included only thanks to your feedback.
And the topics of our today’s discussion are:
Every brilliant idea must be simple in the first hand! They must have posted this slogan on the wall of the Microsoft lab for the engineers to realize that their problem had a much simpler solution than they used to think. The task set for them could be worded the following way: find a way for the personal computer to identify connected analog audio devices.
In fact, a device connected across a digital interface, like USB or FireWire, can easily offer the information about itself and its characteristics to the OS. So, it is all long certain with digital peripherals, while analog devices (headphones, speakers and microphones) are usually directly attached to the audio codec and have no opportunity (or rather capacity) of sending their characteristics across the analog interface.
A majority of PC codecs are implemented on the mainboard and comply with the AC’97 specification. In other words, all AC’97 codecs have practically the same set of characteristics and work under the same specification. The interface of such a codec implies immediate operation with the input and output audio connectors. Thus, without an auxiliary detection scheme, it is a mystery to the OS what audio device is attached to the PC and if it is attached at all. Such a detection mechanism appeared in the new edition (2.3) of the AC’97 specification.
If you want to read this specification by yourself as well as to learn the differences between the 2.2 and 2.3 editions, you can follow this link.As I have mentioned above, today it is impossible for the OS to automatically configure and control an analog device. In the nearest future we will get closer to this: the software identification of the type of the connected audio peripheral will be performed with the help of the new rules from the AC’97 specification along with the new hardware, which complies with these rules. Microsoft test lab carried out a number of experiments to find ways for the most optimal classification and grouping together of various analog audio devices (microphones, headphones, speaker systems), depending on their distinctive features. If such a distinctive quality were found, the OS could use this information to configure the audio ports depending on the type of the attached device. This way not only the audio port function will be determined (namely, whether this “input” will actually be used as an “output”, or vice versa), but also the amplification coefficient could be adjusted and the echo-muffle function might be turned on and off. When experimenting, the Microsoft engineering team focused on measuring the impedance of the electrical loop of audio peripherals. This was the distinctive feature necessary for building an ID-class table and offering requirements for the manufacturers involved.
So, let me remind you of a few basic things from the physics course. The impedance is one of the most important characteristics that describe electrical loops, components and materials those components are made of. The impedance (Z) is usually defined as the full resistance of the device or electrical loop to the current. The full resistance to the sine-wave current is expressed by the ratio of the effective voltage (U) to the operating current (I) in this circuit:
where r and x are their active and reactive resistances.
Thus, the Microsoft team tested a long list of various audio peripherals to compile the results into a classification table. For each category (microphones, headphones, speaker systems), they tested 10 devices to reveal their impedance characteristic. The peripherals were tested in a frequency range from 40 to 20,000Hz. The devices that have their own volume control circuits were tested at their maximum volume to avoid any interference with the impedance measurements.
Obviously, the impedance character will be very distinctive for each audio device (you can see the difference in the diagrams below). That is, it is possible to define an impedance range for each class of audio peripherals. You can see diagrams of some audio devices as tested by Microsoft:
The figures showing the distribution of min and max impedance values for each device were presented as a diagram in the official Microsoft bulletin:
Basing on the acquired data, Microsoft worded its recommendations to the manufacturers of audio peripherals. The design of the device should take into account the following impedance criteria (it is evidently done for the WHQL certification):
The official documentation including the test methodology, data analysis rules as well as plans for classification and systematization of analog audio peripherals can be found at the Microsoft website. For those our readers who are using dial-up I would like to provide a direct link to the document, so that they did not get strayed in the maze of the Microsoft techno library: Analog Audio Classification Using Device Impedance Characteristics (the size of the self-extracting archive is 1.4MB).
The technology (or the auxiliary circuit in the audio codec) that allows the OS to configure audio ports was named Jack Sensing. It is now a part of the AC’97 rev.2.3 specification and is recommended to the audio codec makers. We will see its practical implementation in a while.
The world-renowned semiconductor company, Analog Devices Inc., has long been producing AC’97 codecs for mainboards and add-on sound cards. They work in close contact with such giants as Intel and Microsoft as well as with other big industry names. As soon as the 2.3 edition of the AC’97 specification came out, Analog Devices introduced a compliant codec, supported by fully functional software. So, we can already enjoy that Jack Sensing technology. But let’s be methodical and examine the codec thoroughly first.
The AD1985 codec was specially developed by Analog Devices for Intel’s ICH4 and ICH5 South Bridges. So, I venture a supposition that its lifetime will match that of mainboards with these South Bridges.
AD1985 codec from Analog Device with Jack Sensing support
The codec is an amalgamation of advanced technologies both in hardware and software departments. Analog Devices itself designs the codec, while the codec’s software comes under the SoundMAX trademark. The trademark is property of Analog Devices and was registered in the Integrated Audio Group within the framework of the strategic alliance of Intel and ADI. First, we will examine the hardware stuffing of the codec and then estimate the efficiency of the software.
The hardware part of the AD1985 codec features:
I would like to point out the new auxiliary circuit called Analog Mixing Control Jack Sense. This circuit allows the operating system to identify the attached analog audio device. If you are interested in getting more detailed info on the AD1985 codec, you can find it at the Analog Devices website in a PDF-file.
ASUS was one of the first manufacturers to use the AD1985 codec in its mainboards. Why? ASUS just wanted to roll out a mainboard series featuring the integrated “Artificial Intelligence” technology. The codec from Analog Devices is capable of auto-detecting the connected audio devices, which is an intelligent function, of course. This gave ASUS the reason for touting the audio subsystem of its new mainboards as Artificial Intelligence Audio (AI-Audio). Every manufacturer is prone to puff up the advantages of its products (if any). The ASUS website claims the mainboards of the AI series use the world’s first intelligent audio codec. Do not let the words mislead you into thinking it’s the only one available. Besides the AD1985, there are six-channel codecs from Realtek (ALC655 and ALC658), which support Jack Sensing, too. We are not going to have them in this review, though.
The ASUS P4P800 mainboard I took for this review as a “carrier” of the Analog Devices codec is based on the Intel 865PE chipset. Its integrated audio subsystem is based on Intel’s ICH5R South Bridge plus the AD1985 AC’97 codec. Interestingly, the main page of the SoundMAX site contains a link to this mainboard. So, we do have an official implementation of Jack Sensing here.
The connectors of ASUS P4P800 mainboard
The CD coming with the mainboard, along with various drivers, contains the SoundMAX 4 XL software, necessary for the integrated audio to work. The user-friendly and thoroughly revised interface of SoundMAX is appealing to both a gamer and a musician/music lover. This software features:
AudioESP supports Jack Sensing for the latest codec from Analog Devices, AD1985, only. This feature helps to avoid many problems that may arise when audio peripherals are attached to the PC in the wrong way and informs the user of possible conflicts. The following screenshots show you how it works:
Watch the color codes!
If everything is connected the right way, the Wizard tell you his Ok!
AudioWizard from the new SoundMAX 4 XL helps the user to connect and set up the audio peripherals. The program interface offers you a virtual control center and a toolbar for setting up and monitoring the audio subsystem of your PC.
AudioWizard will guide you through the setup process
The SoundMAX 4 XL driver for Microsoft Windows operating system continues its support of the integrated DLS2 MIDI synthesizer with XG music banks from Yamaha DLS. It also features a noise reduction technology called PureAudio and supports volume sound algorithms, Sensaura 5.1 Virtual Theater, as well as every main gaming API (including A3D, EAX, I3DL2 and Microsoft DirectX).
The original control panel is divided into three sections: playback, recording and setup.
The SoundMAX control panel
The Preferences panel, however, offers you even more settings. It contains four tabs: General, Listening Environment, MIDI Music Synthesizer and Microphone. Let’s see how flexible and far-reaching those settings are.
SoundMAX Preferences panel
The General page allows you to choose the playback and recording devices, which will be displayed in the tool panel. This is where you activate the AudioESP function, which can be set up to detect new audio devices on system startup, show a message when the device is disconnected and enable automatic speaker configuration adjustment. Besides that, you can choose between AC3 and PCM digital formats for the optical output, select the way the control panel is displayed and decide on its color (black or white).
You configure your speaker system in the Listening Environment page. As you can see in the screenshot, two items become active on choosing stereo speakers. The Close Together function optimizes the sound for the small speakers of a portable computer as well as for display-integrated ones (or for speakers that stand right in front of the display). The Far Apart function optimizes the sound for distant speakers or for speakers of a standalone music-box. The function called Sensaura Virtual Theater Surround allows you to send multi-channel music content onto two speakers. Unfortunately, Virtual Theater Surround cannot support headphones without a chargeable upgrade (I really wonder if anyone buys Sensaura libraries?). There’s another Sensaura library present – Virtual Ear (in its standard version, too). To purchase the enhanced version, the user is invited to visit the Sensaura website.
The full versions of the Sensaura libraries are offered as a paid upgrade
I think the Sensaura libraries should be made shareware, so that the manufacturers could pay for us. I guess it might be quite possible.
Besides the above-described functions, you can choose the EAX model for the reproduced composition. The EAX libraries are automatically activated in games, upon the game’s request.
This test helps you to control the surround sound
This page allows you to choose the type of the sound library for MIDI-files playback.
The SoundMAX synthesizer can play MIDI-files using DLS-banks. The bank is loaded with a special utility aka DLS Loader.
The developers of the SoundMAX software agreed to place the advertisement of ANDREA, the manufacturer of microphones and headphones, into the control panel. That’s why we can select a special headset, Superbeam Microphone, in the Microphone tab, next to the standard microphone and headphones. Analog Devices also did this for the developers of voice-recognition applications: they will be able to work with special tools, developed for this purpose. ANDREA brought to SoundMAX its own noise reduction technologies, which work for both ANDREA’s solutions and standard microphones.
I think that the SoundMAX 4 XL software is an excellent product. The developer updates the driver regularly, which is an indication of good work done by the tech support team.
Next, we are going to discuss the VIA VT1616 codec that works in pair with the VT1720 audio-controller. Unlike the AD1985, the AC’97 codec from VIA complies with the previous version of the specification (2.2), and cannot boast Jack Sensing support.
Nearly a year ago, VIA Technologies announced their improved audio-controller Envy24PT (VT1720) for mainboard integration (the Envy24HT, or VT1724, chip is for PCI audio cards; we discussed it in our previous reviews). The new chip brings to mainboards the support for eight-channel sound as well as the “audiophile” operational mode, 192kHz/24bit. The “parent” generation of the chips, from which the new VIA’s chips had originated, worked in entry-level professional sound cards .So, we can hope for that the driver of this audio-controller will support professional functions, soon. Unfortunately, the driver for the Envy24PT as it is today cannot withstand the blame both for the design of its control panel and for the implementation of the functions inherent in the audio-controller. We will review the software a bit later.
VIA Envy24PT is the “audiophile” heart of any mainboard
The audio subsystem of the Albatron PX865 PE Pro II mainboard I took for the tests consists of the ICH5 South Bridge, VIA VT1720 audio-controller and VIA VT1616 audio codec. For the sake of comparison, I also took an Albatron PX865 PE Pro mainboard, which is nearly identical to the first one. The digital part of its audio subsystem is only based on the ICH5 South Bridge, while the analog one – on the popular six-channel audio codec aka Realtek ALC650. The control panel for the codec, naturally, comes from Realtek (by the way, we already discussed the ALC650 codec as an auxiliary unit for the nForce APU in our previous Review).
Intel 82801EB I/O Controller Hub 5 (aka ICH5 South Bridge)
is the driving force for the integrated audio
on the Albatron mainboard
The differences between the two mainboards from Albatron can be seen in their connector rear panels:
Albatron PX865 PE Pro II connectors
Albatron PX865 PE Pro connectors
Albatron PX865 PE Pro II comes with a bracket carrying digital and analog connectors thus providing enhanced integrated audio capabilities.
The bracket has a CS8415A chip, which serves as a digital multiplexer, distributing the information from the serial channel across the connectors.
The difference between the Envy24HT and Envy24PT is negligible, with minor nuances like their connection to the PCI bus. Moreover, the VIA Technologies website offers scanty info on these audio-controllers, and I won’t quote the same things over again. We should better take a look at the software.
After installing the drivers from the CD enclosed with the Albatron PX865PE Pro II, I saw no trace of an exclusive control panel; there was just the standard Volume Control. So, I had to wend my way to VIA’s tech support site and download the universal “family” driver plus the Audio Deck control panel. I was surely surprised to find no such suite in Albatron’s CD, but the driver may have appeared after the mainboard. VIA software developers were writing the Audio Deck hastily and offhandedly: some captions are just abbreviations and there are only Chinese and English versions available.
When I was working on this review, version 1.30c (dated August 25, 2003) was available for download. You can visit the site for updated versions.
Now, let’s get acquainted with the Audio Deck panel:
The control panel is divided into six pages: Playback, Record, SPDIF, Speaker Config, Advanced Control and Information. The panel window is quite large (730x492) and will not scale down (it would be larger than the Desktop in 640x480 resolution!).
The Playback page contains volume controls for the components of the speaker system. You can mute any channel or adjust channel volumes in pairs (the “St.Gang” option). Usually, this option is called Link. Moreover, it is bad style to use abbreviations in control panels.
Note also that there is an indicator of the signal level, next to each slider. However, this function remained unimplemented, to my regret.
The Record page is in charge of amplifying the selected audio source (the list of sources is to the right of the sliders). Here you also enable certain advanced features: Remove Recording Noise and Enhance Mic Recording Sensitivity. The first is for suppressing the noise during recording and the second is for switching between the sensitivity modes of the microphone.
There are only two items in this page. Enable Digital Output sends the sound to the digital connectors. Enable Hi Sampling Rates (96kHz) does what it promises – turns on the 96kHz sample-rate mode.
Your speaker system is set up in this page. Note also that you can quickly select your speaker system from any page by clicking the appropriate icon above. When 5.1 mode (6 channel) is on, you gain access to DualMax functions for hardware signal down-mixing “Downmix LFE/Center DAC to Front channel” and “Downmix Surround to Front channel” (the DualMax function is implemented directly in the VT1616 codec). By placing your mouse pointer over a component of your speaker system, you should hear a test signal from this satellite.
The name of this page is rather confusing, since there are no advanced functions here. We have only two options:
From this page, you also enable the Sensaura 3D Positional Audio mode as well as reproduction of stereo across a multi-channel speaker system (“Channel Extension (2Ch to Multi-Ch)”). After activating/deactivating the Sensaura mode, you have to reboot your system, which is rather inconvenient.
You can read the driver version and names of the library files here.
As you see, this version of the Audio Deck is rather an offhand product. The software developers didn’t have enough time to do everything right. I hope the next version will include all necessary corrections.
I already described this codec in the first part of the article called Contemporary Integrated Sound Solutions: Storm is Coming. It was used in the audio subsystem of an ASUS A7N8X mainboard on the nForce2 MCP-T chipset. Of course, we had a control panel from NVIDIA there rather than the one from Realtek. The Albatron PX865 Pro, on the contrary, employs the ALC650 codec as an independent component of the audio subsystem along with the ICH5 South Bridge.
There is an announcement for people who have Realtek ALC650 sound in their mainboards. The manufacturer?fs site offers for download an update from 9.02.2003 (version A3.49). The update weighs 6668KB and includes the driver and the control panel. I strongly recommend downloading it!
The following table lists the features of the six-channel AC?f97 codecs from Realtek:
Number of Channels
Option of Crystal Elimination
Double Rate Audio
Digital Hardware Volume Control
Selectable Mic Boost
Stereo Mic Supporting
Universal Audio Jack (UAJ)
Windows 95/98/NT/ME/2000/XP, Linux, OS/2
HRTF 3D Positional Audio
Advanced Demo Program
The hardware features of the Realtek ALC650 have already been discussed in our previous review. So, we will now have a look at the control panel developed by the manufacturer. The panel has eight pages:
This page hands you the control over sound environment with an option of choosing defaults and fine-tuning them.
The Properties Editor allows you to tweak
the parameters and save them under a unique name
for future use
You can also enable Voice Cancellation from this tab, which mutes voice frequencies in the music composition. Good for karaoke, you know. Then, you can use the automatic adjustment of the gain level (Auto Gain Control) to prevent overloads and distortions when listening to music files.
The second page conceals a nine-band graphical equalizer with a number of presets for different music genres. The user is free to create and save his own presets.
You select the configuration of your speaker system here. Many modern mainboards don’t come with a bracket with the additional audio connectors. That’s why you can use the three main connectors to plug in the rear pair of satellites, the central channel and the subwoofer. In the standard layout, the connectors are intended for the microphone, stereo system and line input. The screenshot shows you that when the six-channel system is selected, the blue-colored plug is used for rear speakers, and the red-colored one – for the central channel or subwoofer. Even so, the user can click the down-pressed button to return the connectors to their regular operational mode.
This page helps you to test each component of your speaker system. The positioning of a multi-channel system around the user is show on a picture.
You can enable the digital audio-input monitoring (if it is implemented on your mainboard).
There are four operational modes for the digital audio-output: No output, Output digital only, Output digital and analog, and S/PDIF-In to S/PDIF-Out pass through mode.
The audio-visual demonstration of the environmental audio effects looks like an innovation on the developer’s side. This page explains you what three-positional sound is.
The last page tells you about the version of the installed driver, DirectX, audio-controller and the codec itself (this panel is the same for every six-channel codec from Realtek). The language selection menu is a little below. The language support embraces all European languages.
So, the software from Realtek is a well-made product, offering every single function implemented in the codec circuitry. It is good that Realtek upholds its name of a solid manufacturer by updating its drivers regularly and polishing off the software. I think the next codec to be reviewed on our site will be the one from Realtek. For example, the AC’97 2.3-compliant Realtek ALC850. I wonder how Jack Sensing technology is implemented there.
So, we have three mainboards to test today. The testbed was configured as follows:
Below you will see the diagrams with the results in 3DMark03 (Sound Tests) and Comanche 4 Demo. Regrettably, Creative’s audio cards only support 3DMark03 “60 Sounds” test mode, while the drivers for the integrated audio subsystems do not. In fact, even “No Sounds” and “24 Sounds” modes are only available after installation of the latest drivers (for the integrated sound from VIA – after enabling the Sensaura 3DPA mode).
So, let’s see the effect of the enabled sound on the overall system performance. For more illustrative results (or rather for normalization of average fps values), we used an add-on Creative Audigy2 card with every mainboard.
The last diagram shows the results of the mainboards compared against each other. We see that the audio solutions integrated into the Albatron PX865PE Pro and ASUS P4P800 equal each other (in percents). The PX865PE Pro II provides somewhat better results. The game tests showed that the audio-controller (as well as the VT1616 codec, which features a hardware processing algorithm) can take some part of the CPU workload.
The gain of 5% may seem pretty small. Well, it is. Anyway, these data tell us that the hardware processing is being performed, although not so efficiently as when an independent audio-processor is used.
I carried out our tests in the SpectraLAB suite (version 4.32.17) adhering to the methodology explained in my previous reviews. This test will help us to compare the analog parts of the audio subsystems. It is the Analog Devices AD1985 codec in the ASUS P4P800 mainboard, and Realtek ALC650 and VIA VT1616 in Albatron’s PX865PE Pro and PX865PE Pro II, respectively.
Spectrograms for Analog Devices AD1985
44,100Hz/16bit, 1kHz tone signal
Determining IMD at 44,100Hz/16bit
48,000Hz/16bit, 1kHz tone signal
Determining IMD at 48,000Hz/16bit
The AD1985 codec did well throughout the tests, showing a kind of stability across all the modes. This is perfectly seen in the diagrams. Note its low intermodulation distortion coefficient and a good signal/noise ratio. Overall, such characteristics are quite adequate for an integrated audio subsystem.
Spectrograms for VIA VT1616
44,100Hz/16bit, 1kHz tone signal
Determining IMD at 44,100Hz/16bit
48,000Hz/16bit, 1kHz tone signal
Determining IMD at 48,000Hz/16bit
There are a few things to note about the VT1616. This low SNR and high distortion may be the results of an improper wiring layout of the components of the audio subsystem on the mainboard. The manufacturer didn’t find a better solution than place the digital part (the VT1720 audio-controller) on one end of the PCB, and the analog one (the VT1616 codec) on the other. The codec itself is quite far away from the audio connectors, too. That’s why the crosstalk affected the quality of sound; this is quite clearly shown on every spectrogram. Note the parasitic oscillation (up to -80dB) at the 50Hz mark. This is nothing else but the effect of the crosstalk between the audio-signal paths and the power circuits. “Thanks” to them, the sound quality characteristics were lowered.
Spectrograms for Realtek ALC650
44,100Hz/16bit, 1kHz tone signal
Determining IMD at 44,100Hz/16bit
48,000Hz/16bit, 1kHz tone signal
Determining IMD at 48,000Hz/16bit
Unlike the VT1616, our “most popular codec” (that’s how I referred to it in the previous review), Realtek ALC650, showed acceptable results, nearly up to its own specification.
Let me repeat it once again that the results we have got here are the results of the three specific mainboards. By the way, I advise you to compare the results we have got for the ALC650 and VT1616 codecs with the ones we had in our previous review called Contemporary Integrated Sound Solutions: Storm is Coming.
I guess you don’t doubt anymore the progress in the integrated audio field. Integrated audio is constantly developing and acquiring richer functionality. The new, and already available, trends are eight-channel integrated audio for mass mainboards, intelligent identification of the type of the connected peripheral audio device. I also have no doubts that the test labs of respected manufacturers are already polishing off some new technologies to be unveiled soon. As soon as the new technologies are out, we will surely take a closer look at them.
As always, you feedback via email or in the Comments thread below is very welcome.