Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Motherboard VS graphics card integrated audio (Read 4456 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Motherboard VS graphics card integrated audio

Recently I connected the monitor via DisplayPort to my AMD Radeon R7 260X, what amazed me is that using it as an audio output device I am able to set DirectSound in my DAW to 1ms latency and there are no dropouts whatsoever. With mobo integrated Realtek chip I have occasional dropouts even at 25ms latency. I normally use ASIO, but that DS efficiency made me wonder if audio implementation on the graphics card isn't superior in other aspects too, are there any tests?
 

Re: Motherboard VS graphics card integrated audio

Reply #1 –
I do not know about tests but it is possible that graphics card audio channel and driver is so simple (and maybe lower quality) that it is faster than complex DACs on sound cards/mbs.

Re: Motherboard VS graphics card integrated audio

Reply #2 –
Quote from: rutra80 on
Recently I connected the monitor via DisplayPort to my AMD Radeon R7 260X, what amazed me is that using it as an audio output device I am able to set DirectSound in my DAW to 1ms latency and there are no dropouts whatsoever. With mobo integrated Realtek chip I have occasional dropouts even at 25ms latency. I normally use ASIO, but that DS efficiency made me wonder if audio implementation on the graphics card isn't superior in other aspects too, are there any tests?

You're comparing apples and oranges.  The Graphics card audio is very simple, it  stays in the digital domain - no DACs.


The Realtek chip carries the burden of digital to analog conversion.

A good rule of thumb is to keep keep audio in the digital domain as close to the listeners ear as possible. So relying on the graphics card all-digital audio is usually the the better idea.

Re: Motherboard VS graphics card integrated audio

Reply #3 –
The latency value you see in a DAW is only based on buffer size. For example it will say 44.1k@256 samples = 5.8ms, no matter what interface you use (PCI/PCIe/USB/firewire). This value is unreliable. The reliable way of measurement is round trip latency, perform a loopback test in the whole signal chain and measure the latency. In this way the result will at least be doubled (larger than 11.6) and there will be some overhead from less than 1 to more than 4ms depends on the interface's type, result in somewhere around 12-18ms round trip latency. Usually PCIE interfaces are the best.

http://www.dawbench.com/audio-int-lowlatency2.htm
http://www.dawbench.com/audio-int-lowlatency3.htm

I am not very surprised audio interface from GPU can achieve very low latency since the card is usually plugged into the fastest x16 slot and the PCIE lanes from that slot are provided by the CPU directly instead of the motherboard's chipset (assume using Intel Core i processors).

The problem is that standalone HDMI DACs are somewhat rare and expensive, and I've never seen a DisplayPort one. A quick google search showed a "cheap" HDMI one cost $199

However, if the Realtek is glitchy even at 25ms (non round trip DAW provided value) it is really a bit crappy.

Re: Motherboard VS graphics card integrated audio

Reply #4 –
Quote
However, if the Realtek is glitchy even at 25ms (non round trip DAW provided value) it is really a bit crappy.
But, it's not caused by a "crappy" soundcard/soundchip.   It's caused by the multitasking operating system...

If the input/recording buffer doesn't get read in time, or if the buffer is too small, it overflows and you get a glitch.

If the output/playback buffer doesn't get written in time, or if the buffer is too small, it underflows and you get a glitch.  

Re: Motherboard VS graphics card integrated audio

Reply #5 –
It doesn't matter. If you see the DAW bench links I posted, all interfaces are stress tested by the same OS and DAW with the same set of plugins. If one performs well and another performs poorly in the same stress test condition then it simply means the glitchy one is incompetent, in either hardware design or software/driver side.

Re: Motherboard VS graphics card integrated audio

Reply #6 –
Quote from: DVDdoug on
Quote
However, if the Realtek is glitchy even at 25ms (non round trip DAW provided value) it is really a bit crappy.
But, it's not caused by a "crappy" soundcard/soundchip.   It's caused by the multitasking operating system...

We have plenty of of reliable real-world experience-based proof that a well-designed multitasking operating system and a well-designed audio interface can operate glitch-free.

Re: Motherboard VS graphics card integrated audio

Reply #7 –
Quote
and downloaded the correct Ffmpeg and LAME files for it.
Of course you can get glitch free audio on a PC or a Mac, but not without a buffer and the associated latency.  

Often the latency can be brought down to an acceptable level, but some people have trouble getting glitch-free recording when they reduce buffer size for low latency.

Re: Motherboard VS graphics card integrated audio

Reply #8 –
First thing to check on a Windows system with audio problems: DPC latency.
Next is to check priority of your audio application process/thread and system load.

But yeah, while increasing buffers will increase delays it does not tell you much about the total delay. So if comparing apples and oranges you should measure latency.
"I hear it when I see it."

Re: Motherboard VS graphics card integrated audio

Reply #9 –
Quote from: DVDdoug on
Quote
and downloaded the correct Ffmpeg and LAME files for it.
Of course you can get glitch free audio on a PC or a Mac, but not without a buffer and the associated latency.  

The buffers are not required because the OS is multitasking.  The buffers (usually there are at last 2) are required because the audio interface or other piece of I/O gear runs asynchronously or at its own rate as compared to any other tasks. 

How else can we manage the situation where the I/O gear runs at a different rate than any of the other things that the O/S does, such as run user programs? There is an answer, which is to halt everything else that the system is doing, and run the I/O at its own rate. This was used in early computers (e.g. up through the IBM 1401) through the mid-1960s, but is usually very inefficient because I/O gear usually runs relatively slowly, and stopping everything else and waiting for I/O  to start and then finish is time-consuming.

In the case of an audio interface, stopping everything else and waiting for the data to be read in or written out in is incompatible with having sound that is uninterrupted and free of clicks and pops.

Re: Motherboard VS graphics card integrated audio

Reply #10 –
Some popular keywords about improving latency:

power saving related:
speedstep
c-state
core parking

timer related:
hpet
useplatformclock

Be careful when changing power saving settings, some settings may simply make the CPU hotter without real improvement. The effectiveness is highly hardware/OS/software dependent.