In response to quite a lot of feedback on our original article entitled "The forgotten component - What's up with computer audio?" since its birth, it's been brought to our attention that not everything we mentioned was entirely profound in reference to the SoundStorm's Dolby encoding. We might have referenced SoundStorm as being "mighty" a little too much although we still believe it is just that but we do have a few adjustments we need to make to deliver entirely accurate content to you, our readers.
We, like many of you, learn the ropes by producing an argument or stance on a particular aspect of the IT industry (in this case in-depth technology) and through feedback we can learn and absorb true and previously uninformed facts from the inevitable others out there who are more 'in the know' within this particular area, like any area of life. Constructive feedback and making an effort to point out any flaws or miscalculated judgments on our behalf is something we welcome with arms wide open and benefit from, not only for us but also for everyone else out there who wishes to learn more about the technology in full view.
In order to be assured we are on the right track, we have since spoken with nVidia's Micah Stroud, Senior Product Manager for nVidia Audio. During our discussion with him highlighting some of our earlier content for perusal, we are now able to present everyone with a few corrections and in several cases lack of noteworthy expanded comments on our behalf.
Let's get started with the follow-up.
Following it up
The main point to the article regarding SoundStorm's uniqueness is that *IF* you were wanting to isolate your sound output to Digital coaxial or optical, the only possible sound solution on the market capable of outputting PC generated sound to the independent 5.1 speakers is when the source has been encoded into those 5.1 channels at the PC's end prior to being sent to the amplifier in which it is then deciphered and sent out to the corresponding channels. This is where Dolby's Live encoding process is given full credit, something that only nVidia's now dead SoundStorm was able to do in hardware on the fly.
It is important to note that the Dolby Digital Live encoder is a separate and final step of the SoundStorm audio hardware process. It takes a mix down set of 6 channels and encodes them in real-time (at 640kbps) into a Dolby Digital bit stream and outputs this stream over SPDIF. Dolby Digital Live does not position, clone, create, or otherwise "3D-ize" the audio stream, the nVidia APU itself is in charge of all these tasks. All of the processing, sample rate conversion and acceleration happens within the nVidia APU before the Dolby Digital Live encoding process begins. Credit should be given to the nVidia APU itself and not just the Dolby Digital Live encoder part of the silicon.
Other sound solutions outputting via optical or coaxial can also pass through an already encoded 5.1 signal (DTS or Dolby) such as DVD's. This untouched audio source is sent to the amplifier then it decodes the raw format and outputs to all 5.1 channels accordingly - like PowerDVD with "SPDIF pass-through" set in the audio preferences. Games, MP3s, and all other PC generated sounds are NOT already encoded into 5.1 therefore you'll be limited to sending a two channel signal to your amplifier on anything but SoundStorm due to lack of on the fly encoding of 5.1 Dolby Digital at the source end.
However, most people still use analogue outputs for their sound in which case this has no influence, only in that the hardware acceleration for 2D, 3D, occlusion, obstruction, EAX, DLS, DSP effects, sample rate conversion, mixing, EQ processing, up conversion (cloning), down conversion (folding) and other compute intensive tasks are still resident through the SoundStorm APU which is in charge of taking on these important tasks. Unfortunately, the DAC used in conjunction with SoundStorm for analogue output is that of your common el'cheapo variant (mostly Realtek) seen on the majority of onboard sound solutions from nVidia's motherboard partners. This detracts from what would have been far superior sound quality when utilizing the three analogue output jacks over SPDIF out. (front, rear, centre/sub) - Basically, if more expensive codec's were used, the SoundStorm would have been better.
The article was only focusing on SoundStorm's own ability to send 5.1 channels continuously in a Digital format. Quality of sound, S/N ratios, lack of ASIO support etc... We could have gone into all of that during our comparisons but that's not what the sole purpose of this article was intended for. It was not meant to be a full blown and in-depth soundcard roundup. Dolby encoding is a feature we believe many people would like to see as an option when purchasing a sound solution for their computer system but is no longer available, unless of course you want to consider the use of software based Dolby encoding solutions now on the market, several of which are Intel branded for what we believe to be only about three 915 and 925 motherboards at this stage. Also, C-Media's own 9780 based card is capable of software based on the fly Dolby Live encoding. However, software encoding is of course inferior to the dedicated hardware implementation SoundStorm makes use of, as many more CPU cycles are used in the continuous process with the impact being mostly noticed during gaming.
The benchmarking tests in our article were merely to highlight that an onboard audio solution working just as hard as the other sound solutions is not only able to rival them, but in several cases come out on top. Obviously the main factor contributing to this is the inclusion of its own hardware acceleration which is something the majority of other onboard solutions out there currently lack - since it is the cheaper design method.
It is indeed a shame we couldn't be more precise in obtaining the results of our performance tests. We put a lot of thought into what method(s) would be best used for the purposes of this article prior to conducting these tests. It was what we felt the best real world analysis of how different soundcards influence the performance of gaming only.
Getting back to highlighting a couple of things previously untouched, first and foremost, when SoundStorm converts the digital stereo signal of your source into Dolby Digital 5.1 it is in effect not only re-producing the content to be outputted on all five speakers, but also compressing the signal at the same time, as that's what Dolby Digital and DTS are - compressed raw audio signals with DTS to a lesser extent (hence it's slightly richer and clearer sound). Taking this on the fly compression into account, the sounds being reproduced at the other end (your speakers) are actually of a lossy format over simply outputting directly from your PC without the influences of encoding beforehand, although the quality of your soundcard's DACs also play a role here. Technically, this would attribute to a slight loss of sound quality (a small amount of content will be lost) at the end of the line but as to how noticeable is purely subjective. At the end of the day, digital signal or not, the sound has to be converted back to analogue somewhere along the line as that's what your actual speakers are outputting.
Second of all (and in many ways linked to the paragraph above) although we stated the quality of Dolby Digital 5.1 to be somewhat clearer and more wholesome when gaming versus that of analogue 5.1, this isn't exactly true. Although the signal being reproduced is now in crystal clear digital, it couldn't possibly be better than the original source sound as that would require a raw untouched digital signal to be sent to the decoder (amplifier) in the first place. (i.e. direct from DVD movie content). With gaming and so forth the sounds are being converted and compressed into digital a second time over (introducing even further but mild sound degradation) from what was previously already singular compressed content (also noting that most games use some form of audio compression for their sound effects such as the common MP3 standard). This logically concludes what you would believe to be a lossy result when encoding to DD, if only minimal.
In reference to the following paragraph within our original article, primarily focusing on "positional" audio:
First and foremost, the beauty of the nVidia SoundStorm APU is that it is capable of encoding Dolby Digital 5.1 on the fly via hardware acceleration and not software (CPU). This means that in any games you play and as long as you are using optical or Digital coaxial cable with your surround sound speakers (anything above 2.1 channels), the hardware APU will do the intensive job of reproducing the sound from the game to Dolby Digital 5.1 or AC-3 so you get proper positional surround sound.
Some of you have been puzzled and/or misguided as to how this is possible. In a clearer and more constructive means of wording, allow me to go back over the above in a better fashion.
SoundStorm, although capable of encoding any source into 5.1 Dolby Digital will still only output what the original source was intended to be. For instance, taking two channels of sound into account (i.e. stereo) from a game that doesn't support positional surround sound, encoding into Dolby Digital 5.1 does not introduce positional surround sound to these games. It simply couldn't as the instructions for positional surround are not embedded into the game title. At best, it could attempt to emulate positional surround (i.e. similar to Pro Logic II). Instead, what Dolby Digital encoding does for original stereo sources is merely replicate the front two speakers to your rears, in effect sounding much the same as simply not encoding to Dolby in the first place but setting your amplifier to "Stereo x 2" which also sends what's being outputted at the front speakers to the rears at any given time.
In games that support positional surround sound, however, is where Dolby Digital Live encoding is a truly useful feature since it allows the positional channel instructions to be carried through the newly Digitally encoded bit stream whereby the amplifier then knows exactly how to decipher the stream and output to the corresponding channel/speaker - all this done on the fly in real-time as you play. Providing the original game makes use of a full 5.1 setup, your centre channel and sub will be given their own independent channels as well, and in the case of the few titles out there limited to quadraphonic (4 point surround) the DD Live encoding will only output to those same four speakers.
To confirm Dolby encoding for gaming was capable of this, we fired up Doom III with Dolby Digital encoding enabled (knowing full well its unique standalone audio engine is capable of outputting a full 5.1 channels of sound) then loaded our way into the opening map. We figured that the best and quickest way of determining if the positional data was being deciphered correctly at the receivers end was to walk up to one of the in-game characters and get them to start rambling on. We then swiveled around 360c very slowly as he talked away and surely enough his voice was moving across all 5 channels independently and correctly during this time.
One other thing you must obviously realize if you want to take advantage of SoundStorm's Dolby Digital Live encoding process is that you must be outputting the digital optical/coaxial cable to a set of PC speakers or dedicated home theatre amplifier (such as our Onkyo use din the original article) capable of decoding Dolby Digital 5.1 or you will not be able to experience the benefits of Dolby Digital. This is something many people would believe to be too costly for their basic PC audio needs but it's an option. The complete PC speaker options are definitely more affordable but in terms of 5.1 sets capable of taking an SPDIF coaxial or optical input, there are not that many to choose from on the market at this stage.
Other then those fresh pointers, we feel the rest of our article is still an excellent representation of what we are all missing out on within the PC audio industry, not just in the SoundStorm / Dolby encoding area, but with audio solutions across the board. We hope this has cleared up a few things for some of you and apologize if some of our previous content was difficult to follow.
Please feel free to post your constructive views/comments in the dedicated forum thread in our Audio and Visual Forum and either Cameron or myself will respond accordingly.