Tuesday, August 29, 2006
I also know that some pro-audio external soundcards like M-Audio's Delta 66 and Delta 1010 or Echo Audio's Layla and Gina use a proprietary connection to a dedicated PCI adapter card. Both of these manufacturers are now using either USB or FireWire on all of their newer external soundcard products. We will not be using a proprietary PCI adapter card solution.
USB is the most common method used to connect peripherals to a computer. USB 1.1 was originally meant as a replacement for legacy ports, and a user-friendly, low-cost way to connect peripherals such as keyboards, mice, and printers to a PC. USB was never designed to handle intense multimedia data loads. USB 1.1 is limited to a data rate of only 12Mbits/sec, which is enough bandwidth for CD audio (1.4Mbits/sec). However, it might be difficult for USB 1.1 with it's associated overhead to handle multiple streams of CD audio or DVD-Audio (with 6 channels of 24 bit audio at a sample rate of 96KHz) which requires up to 8.6Mbits/sec. USB 2.0, which is backwards compatible with USB 1.1, provides the fastest data rate currently available at 480Mbits/sec. However, USB uses a host/client (or master/slave) architecture. The PC takes on the role as the host, which requires some overhead to handle all of the arbitration functions and dictate commands to the clients, thus reducing the overall data rates. Data rates are reduced more if addition communication between other clients is required. The maximum length of a USB cable is 5 meters, greater lengths require USB hubs.
Many of the manufacturers of pro-audio soundcards offer an external solution using FireWire (IEEE 1394). FireWire is based on a peer-to-peer technology where each "intelligent" peripheral communicates with each other to provide sustained data rates. Unlike USB, no additional PC overhead (system memory or CPU) is needed to sustain the data rates. FireWire includes support for memory-mapped devices, which allows high-level protocols to run without forcing numerous interrupts and buffer copy operations on host CPUs. The most common FireWire in use today, IEEE-1394a, provides data rates up to 400Mbits/sec with cable lengths limited to 4.5 meters, but up to 16 cables can be daisy chained using an active repeater, external hub or internal hubs included in many FireWire devices. The maximum length possible with any configuration of IEEE-1394a is 72 meters. A newer FireWire standard, IEEE-1394b (sometimes called FireWire-800) provides a data rate of 800Mbits/sec (but promises to support speeds of 1600Mbits/sec. to 3200Mbits/sec. in the future) and cable lengths of 100 meters when using optical fiber or CAT5 cabling. However, if you read the fine print, these longer lengths only support data rates of 100Mbits/sec when using CAT5 cabling. A future version of FireWire, p1394c is suppose to provide data rates of 800Mbits/sec for cable lengths of 100 meters.
Theoretically USB 2.0 should be faster than FireWire IEEE 1394a. Each FireWire device has to negotiate for bus access and the FireWire bus must wait until a given signal has been sent to all devices on the bus. The more devices on the bus the lower the performance. While USB is only limited by the host-client branch, not the whole network. USB's host-client technology also allows the host to allocate more bandwidth to higher priority devices. Even though the USB 2.0 spec of 480Mbits/sec is higher than FireWire's (IEEE 1394a) 400Mbits/sec, the increased CPU and host overhead due to the host/client technology of USB 2.0 reduces its sustained throughput to rates lower than those of IEEE 1394a. USB comes standard on over 90% of computers shipped today (probably more than 90%, now). Unfortunately, FireWire does not come standard with most new PCs. However, I've noticed that it is included on many of the motherboards that are popular with people who build their own HTPCs. Anyhow, both FireWire and USB 2.0 can handle the audio throughput needs of an external soundcard.
Looking into the future, it appears that the performance of USB isn't really going to change much. The only references to a future version of USB talk about wireless USB (WUSB). Wireless USB sounds interesting. You'll get USB 2.0 performance within a range of 3 meters. Between 3 and 10 meters, WUSB can operate at 110Mbit/sec. The future versions of FireWire are even more interesting. I'm not sure exactly what the bandwidth requirements will be for future audio formats like HD-DVD and Blu-Ray (probably, not much more than DVD-Audio's 8.6Mbit/sec, but you never know). And who knows, maybe there will be something better after these two formats kill each other off. Anyhow, the speed improvements of future versions of FireWire might be very useful. What's even more exciting is the fact that the cable lengths will let us support features like whole-house multi-zone audio much easier. With S800BaseT maybe we will be able to have an HTPC in the family room with our soundcard/amp, and others in the kitchen or dining area and bedrooms. All of the PCs are connected via FireWire to either multichannel soundcard/amps or stereo soundcard/amps. They are also connected to each other over the S800BaseT network. Because of the network connection, any of the PCs will be able to recognise all of the other soundcard/amps on the network. That is, your PC will actually have every soundcard on the network in its list of audio devices. With the appropriate software, like J.Rivers Media Center, they'll be able to set up multiple zones. From any one of the PCs you'll be able to pipe music all over the house. For parties, the same tune can play in every zone. However, you will still be able to control each soundcard from the local PC. Hopefully the software will be able to manage when the soundcard is under local control or is available across the network. Anyhow, the possibilities are very exciting since you should be able to connect all FireWire devices up to distances of 100 meters and still maintain S800 (800Mbits/sec) speeds.
After reading several articles or papers describing Audio over Ethernet, I'm not sure if it will ever come into the mainstream. First, there are a lot of proprietary solutions and without a standard, it's unlikely to move beyond its niche. The technology is really geared to professional applications like auditoriums or large recording studios. Second, if S800BaseT takes off, it looks like there will be no need for audio over Ethernet. Anyhow, it is another technology that we will keep an eye on.
I'm sure some people reading this post are wondering why I haven't included HDMI in this discussion. It's because I don't think HDMI (High Definition Multimedia Interface) will ever replace USB or FireWire as a way to connect to external soundcards. Instead, it will probably replace S/PDIF because it is also basically just a one way connection. It is a much higher bandwidth digital connection. Instead of the S/PDIF limits of 2 channel PCM or multichannel Dolby Digital or DTS compressed audio, HDMI can handle 8 channels of uncompressed audio (in addition to all of the HD video). We are considering using an HDMI input instead of/or in addition to S/PDIF as our digital input option. We are already beginning to see HDMI connectors on graphics cards, high resolution displays and A/V receivers. In the next few months HDMI connectors will be included on new motherboards. Another big part of HDMI is content protection and HDCP, but I don't really want to cover that in this blog entry. Maybe some other time.
Tuesday, August 22, 2006
Monday, August 21, 2006
When asked, if after you plug your PC into an A/V receiver and perform the room correction calibration on the PC and then disconnect the receiver from the PC, 'will the receiver retain the calibration settings?', Amir replied, "Unfortunatly not. All the processing is done inside the PC and only works if you use the PC as the source, driving your receiver. Take away the PC and there is nothing there anymore. For your receiver to have similar capabilities, it would have to have a ton more hardware in it." Anil then favorably compared Vista's room correction capabilities to a high end TacT processor by saying, "The TacT processors do what you say but the one I have cost me a cool $10K and it does less in some respects than Vista!"
For a quick explanation of what DRC is and how it can benefit you and your system, check out this DRC guide.
The bass management feature is very flexible and doesn't have some of the limitations you find with the bass management on most A/V receivers. For example, with Vista's bass management, you can have any combination of large or small speakers, with or without a subwoofer. The crossover point can be set to whatever your loudspeakers need. When sending a full range signal to any of your small speakers, the portion of the signal below the crossover point will be routed to the subwoofer, or in the case of a system without a subwoofer, the lower frequencies can be routed to the large speakers.
Most A/V receivers won't let you send a full range signal to the front speakers if the subwoofer is engaged. They force you to filter out the low frequencies from the fronts and send everything below the fixed crossover point to the subwoofer. This works okay with a sub/sat system, but isn't very appealing for anyone who has large or full-range loudspeakers in the front. There are also some A/V receivers and processors that will not send a signal to the subwoofer when it is set to stereo mode. Also, with most of the A/V receivers and processors, the subwoofer output jack is low-pass filtered so any signal above the frequency setting (usually in the 100Hz-150Hz range) is blocked. Most powered subwoofers also have a built-in low pass filter, and when the two filters combine you are increasing the steepness of the filter. Too steep of a filter slope makes for worse sound. Anyhow, it sounds like Vista's bass management feature will let you avoid these types of limitations.
Vista now has a new audio mixer that according to Amir works much better than Windows XP's Kmixer. Kmixer resampled all audio to 48KHz, unless the sample rate was already 48KHz. Audio originally from a CD source, which uses a sample rate of 44.1KHz, was always resampled to 48KHz and the method used by Kmixer degraded the sound quality. With the new mixer, maybe the algorithms for resampling are much better so the sound quality remains high. The new mixer also has a user mode audio engine where a user can set the default sample rate and includes an exclusive mode which provides access directly to the soundcard. With the user mode, the mixer will not apply resampling at all. However, its not clear if the mixer will just leave the source material alone so that if you are playing ripped CDs it will use 44.1KHz and when playing from a DVD, it will automatically switch to the native 48KHz.
Be sure to read Amir's posts in the thread linked above to learn about all of the other great audio processing features. Also, take a look at this post in the Windows Vista Team Blog about the new audio features in more detail. And here's an interview with Microsoft PM Hakon Strande who also talks about high definition audio in Windows Vista.
These features will only work if you send either a 2-channel PCM stream over S/PDIF to your A/V receiver, or analog signals directly to your receiver/amps. So for anything more than 2-channel stereo, including 2.1 (stereo plus subwoofer), you will need to use analog outputs. This might be a problem because most people prefer using their A/V receiver's processor and DACs especially when they perform better than the DACs included with inexpensive soundcards or the built-in audio chips on the motherboard. They also feel anytime an audio signal is present inside a computer, the signal will pick up noise generated by other internal components like the computer's power supply or hard drives, thus decreasing the sound quality even further.
However, using one of our future products, you'll be able to take full advantage of Vista's advanced audio capabilities and enjoy excellent sound quality because we use the same components found in high end dedicated DACs and the best pro soundcards. Since these DACs are located in our box, you don't have to worry about any possible noise generated inside the computer. In addition, our amp modules are very efficient, have lots of power, generate less heat and sound better than most amps inside your typical A/V receiver.
Friday, August 18, 2006
So, I apologize for posting something that isn't directly related to our technology or development. I'll try to keep these posts to a minimum and will probably delete them if I think they no longer serve their purpose.
BTW, if anyone knows of other useful techniques to help generate more (some) traffic, please let us know. Hmmm... maybe we should change the name of the site to "AmplioPod" and see if we get a cease-and-desist letter from Apple. Maybe that will draw some attention. Probably not the kind we want.
One of the recommendations in Blogger's help tells us to sumbit (I think they really mean "submit", but sumbits might be cool) your address to blog search sites and directories. There are a few sites they recommend like Technorati (Technorati Profile), Daypop, Blogdex and Popdex. Eventually we'll try them all. For the submissions to work, I have to insert these links into my post. Done. Now we'll see if that works.
Wednesday, August 16, 2006
Since we have fairly limited funds, it was important to make the prototype with components that wouldn't take a lot of time or cost a fortune to build. So the conceptual prototype is made up of some "off-the-shelf" components. We used a commercial external FireWire based soundcard typically used for professional audio engineering applications and class-D amplifier modules that are available for DIY audio amplifier builders. Since this is a 6 channel system which requires 5 channels of amplification (the 6th channel for the subwoofer or LFE channel does not require amplification because the subwoofers typically are purchased with their own amp), and we want our system to be a modular or flexible system, so we included 3 400W amp modules for the front left, right and center speakers and 2 180W amp modules for the rear left and right speakers. Since the 400W and 180W modules require different operating voltages, we also had to include separate power sections for the front and rear speakers. So we had to have 2 custom toroidal transformers built for the prototype because we couldn't find any stock toroids to meet our needs. Next, we had to add circuitry for converting the AC voltage into the DC voltage needed to power the amp modules and circuitry to limit high in-rush currents during start up. To complete the prototype, we had to build a chassis that could hold everything together in a nice neat package and dissipate any heat generated by the components without generating any noise.
So I now have a box that connects to my HTPC using a FireWire cable and to my Precision Monitor 10 loudspeakers using speaker wire.
Instead of trying to explain the advantages of using a PC to play music, I can now demonstrate that it sounds really good. So during the demonstration they get to hear a really good sounding system and learn a little about the advantages of using a computer for managing your collection and playing music.
It really does sound good. I don't consider myself an "audiophile," but I do appreciate good sound. I've owned a variety of fairly high end esoteric systems in the past, including a pair of Beveridge Model 2 SWs, a modified HQD system, B&W Matrix 801s with an assortment of electronics from Mark Levinson and Audio Research. Not the highest of the high end, but still pretty good sounding stuff. Actually, I had to get rid of the HQD system because it just was too big for my space. My system included a 27" Hartley driver in this huge, heavy subwoofer enclosure that was something like 4'-6" wide x 4'-6" deep x 3'-6" high. The double stacked Quads looked like a pair of big solar panels. It just took up too much space and was far from meeting good GAF or WAF (girlfriend/wife acceptance factor). I really downscaled after the HQD system. Through a series of trades and purchases, I ended up with an upgraded Dynaco ST-70 amp, an Electro Research EK-1 Phono Cartridge/Preamplifier System, Linn Sondek LP-12 turntable and the Precise Monitor 10 loudspeakers. The speakers were probably the weak link in the system (and the amp didn't have a lot of power). However, they were about the same size as the B&W 801s and seemed to have a little more clarity. The tube amp at 35W/channel provided enough power to drive the Monitor 10s, which have a rated efficiency of 88db, to pretty loud listening levels. Anyhow, I was pretty surprised to hear how much better the Monitor 10s sound now that they are hooked up to the class-D amp modules. They are really amazing sounding speakers. I hate to sell them, but could use the cash.
I haven't had as many demonstrations/auditions as I would like, but I've had the opportunity to play the system to a few investors or potential investors, along with a few of my friends and family. Even my wife is impressed and it takes a lot to impress her when it comes to hifi. My Dad, who is a fan of big band music, swing, etc. was very impressed. I played some Frank Sinatra recordings from the 60s and he was amazed. Then I put on a recent Michael Buble' recording and he nearly fell off his seat. Obviously the modern recording took advantage of the latest and greatest studio recording equipment and would naturally sound better than something recorded in 1961.
Now we are working on the "working" prototype. This will be much closer to the actual commercial product. The pro soundcard used in the conceptual prototype has a lot of functionality specifically for recording. Our soundcard component won't need a lot of the recording components, but will need functionality specific to home theater like bass management and digital room correction features. There are also a few options for enhancing the amp modules and power supplies which we will want to include in the working prototype. In future postings, I'll share some more of these details.
Tuesday, August 15, 2006
Our company, Amplio Audio, is currently developing a product that combines a computer's soundcard with high quality class-D amplifier modules. Our first "conceptual" prototype is basically a 5.1 surround sound (6 channel) product. In a future post, I will provide photographs of the completed conceptual prototype. The prototype (I haven't given it a name -- any suggestions?) has performed very well. It is currently hooked up to our HTPC (home theater personal computer). The HTPC is also connected to our network. I currently use J.River's Media Center 11 to play music from my library of losslessly encoded WMA tracks, which are stored on a media server on the network. To listen to the system, I have a pair of Precise Monitor 10 full range 3-way loudspeakers connected to the prototype. We also have a full Era 5.1 speaker system that we plan to use for future testing (so, the Monitor 10s are for sale!!).
The prototype product currently includes the following components:
- the chassis
- power supply - toriodal transformers and power supply PCBs
- amplifier modules
- FireWire interface
- digital/analog section (DACs, ADC, etc.)
I will provide more information about each of these sections in future posts.