Comparison of DS, KS, Asio, and wasapi output in win7)

Transferred from: bytes

Comparison of DS, KS, Asio, and wasapi output in win7

Next we will introduce the history of wasapi and the development history of Microsoft's voice processing:

In Windows before WINXP, General DMA was used to process sound. DMA = Direct Memory Access is the IO mechanism of all the storage devices, such as harddisk, LAN Card, and drives, most of my computers in the 386 s used Pio to process HDD Data Access. That is, all data written by HDD must be directly controlled by the cpu so that Ram and HDD can transmit data to each other, however, this will occupy a lot of CPU loading, and the higher the HDD volume, the worse the situation.
So in the Intel Pentium MMX era, drive controllers began to switch to DMA, enabling the chip to easily control or direct HDD/drives to Access RAM directly, without the need to process all transmitted content and materials through the CPU, the CPU can completely eliminate the burden on HDD data, even if the former IDE/SCSI to the present SATA uses high-performance DMA, however, the DMA approach is to wait for interaction. All the information is not only based on RAM as the buffer Transfer Station, but also the base on Ram clock timer to control the transmission time. Although it saves computer performance, compared with audio Io, this is not stable, slow, and large latency, and it is difficult to restore the original sound, because sound is an energy that requires time, and sound can only be made up of time, in addition, PCM signals require that each sample be delivered on average in real time to maintain the original sound.
What DMA audio uses is what kmixer is generally called.

However, Microsoft claims that Vista/7 has begun to abandon kmixer and audio Io dependent on DMA, and developed what they call wavert (wave realtime) and their wasapi, mmcss and so on use wavert as the core, while wavert has its own mixer, but as long as you start the exclusive touch style, you can bypass this mixer, and mute drops all otherProgramThe voice, mmcss enables you to improve the priority of audio input and output Io to the highest time pulse. What Microsoft wants to do is to use real-time clock timer to control the audio stream without DMA, communicate directly with uaa audio devices hareware, and even let the hareware clock of the sound card or audio interface directly control audio data. This is intended to be very similar to ASIO.
Even in the wasapi sharing mode, no SRC is available. Instead, you can freely set the sampling rate, bit size, and channels for the shared audio on the console, therefore, the original 44100Hz signal can be retained and will not be retained by Src.
Now all Intel motherboard or Intel chips have hpet (High Precision Event timer), which can make video and audio processing more real-time and precise processing of high sampling rates, low bus latency, the number of data stream events that can be responded to per second has increased significantly, but I do not know whether AMD has
Http://www1.uwants.com/viewthread.php? Tid = 8157140 & page = 1 # pid116989187

About wasapi by Microsoft
The http://msdn.microsoft.com/en-us/library/dd371455 (vs.85). aspx

Next, we will introduce how to get the perfect sound effect from win7 + foobar without the sound burst and output it from the original sound source:

Note: From dearhoney sound studio

I have read some domestic and foreign discussionsArticleThen, I did it myself and confirmed that the following method can ensure that the output of 16 bit 44.1 kHz digital signal is completely original.

1. Play software please use foobar2000, download: http://www.foobar2000.org /? Page = download, and then install it.
2. In addition, download the wasapi output file on the same page. Put the file foo_out_wasapi.dll in the components folder of the foobar2000 installation directory. For example, in this example, because it is installed in the x64 environment and everything follows the system preset, it is placed in the path c: \ Program Files (x86) \ foobar2000 \ components.
3. Start foobar2000 and go to files → preferences to adjust the output page under playback. If the page goes smoothly, two sound devices starting with wasapi will be displayed. If you use the digital output/HDMI output, select the digital output. If you do not have an external decoding device, you want to use the hdav 1.3 analog output, select the Speaker: (there are three wasapi devices, one of which is not hdav 1.3, which is the HDMI output of the motherboard M3A78-EM)

 

Win7_foobar_wasapi

You will find that you can play music, and there are several features:

1. You cannot adjust the sound size from the system volume console.
2. All other system sound effects cannot be heard.
3. If your device is good enough or you have a golden ear, you will find that the sound quality decoded by your peripheral amplifier is better.

But how can we ensure the original 16-bit 44.1 kHz digital signal? I verified this: I have some wave files at hand, which are essentially DTS-encoded signals. This method is used to play the files from foobar2000 and immediately turn on the DTs signals around the expansion machine, is the best proof. Why? If it is a normal music file, that is, it will eventually be a stereo 16-bit 44.1 kHz digital signal, if there is a SRC will be changed content, but the ears are not necessarily heard; however, I am playing a data transmission-encoded wave file. If it has been processed by Src as a two-channel music, it will completely become a noise and be lost, however, the sound is not heard after such a setting, but the correct content is used to enable the surround amplifier to light up the DTs signal. As you can see, this setting is completely out of the control of SRC and Windows multiplexing, the original digital signal is successfully output.

Summary:

Only DS andWasapi, DS with y460 Dolby sound effect is good; but there will be broken sound during the pitch, the MS bass is too large, it is difficult to control; addWasapiAfter that, the output sound becomes very delicate, and the high-pitched voice is bright, but the bass is not as deep and thick as DS, but after all, the high-pitched voice cannot be clear or even the bass is good. After all, it is not an external speaker on the book, and jbj's audio is enough;

On the contrary, DS is automatically enabled when the external access headset is enabled (the Dolby sound of y460 is enabled ), it works well when playing games, but DS cannot listen to music on foobar, so it is even more common to listen on WMP. Compared with wasapi of foobar, the output effect is far .