Operation Manual

6. Hardware Buffer on/off

Note: For WaveRT drivers (Vista), this box is labeled “

Allow Pull Mode (WaveRT)

” instead!

Enables the hardware buffer for the highlighted device. This only works for so called “WavePCI”

miniports, as other types of WDM drivers do not usually allow direct access to the hardware buffer.

Adjustment for best hardware buffer performance involves the “ASIO Buffer Size” slider and the “Buffer

Offset” slider (see below). Hardware buffering works best for rather small ASIO buffer sizes. Try

something between 128 and 256 samples as a starter!

The biggest advantage of using the hardware buffer is that this method uses a lot less CPU. In addition, it

may be possible to decrease latencies even further.

In multi-device-setups, it is possible to mix Hardware-buffered devices with devices that are not. This,

however, is not particularly recommended!

If hardware buffering is not supported by a particular audio device, there will be an additional latency of a

couple hundred milliseconds, which is clearly audible.

“Allow Pull Mode (WaveRT)”

There are two basic access methods for a WaveRT device, “pull-mode” (also called “event-mode”) and

“push-mode” (also called “polling mode”). “push-mode” is thought to be supported by all WaveRT

drivers and (Vista) OS platforms, whereas “pull-mode” wasn't even officially documented by Microsoft

until almost one year after Vista went RTM. As a consequence, “pull-mode” is still broken in a number

of instances.

If this box is left unchecked, ASIO4ALL will never use “pull-mode”, otherwise it will use it whenever

possible. If it works, “pull-mode” provides tighter timing and is generally thought to be technically better

than “push-mode”.

Note: The old ASIO4ALL 2.8 behavior for this option was as if

checked

! If you check this box and there

are no problems, keep it checked for better performance!

7. Kernel Buffers/Buffer Offset

If hardware buffering is disabled, this control lets you add up to two more buffers to be queued for audio

output. Each additional buffer increases the output latency of the device by the time it takes to play one

buffer. Therefore, the initial setting of “2” should only be changed on less powerful machines, where

reasonably small ASIO buffer sizes cannot be achieved with the default setting.

If hardware buffering is enabled, this control determines the amount of clearance (in ms) between where

ASIO4ALL will insert data into/read data from the hardware buffer, and the position where ASIO4ALL

currently thinks the hardware read/write position is. Sound complicated already? You haven't even seen

the code that calculates this...

As a general rule: Higher settings increase latencies and stability, lower settings have the adverse effect.

You should, however, be able to achieve a setting that is very close to zero (“4ms” would still be

considered “very close to zero”, while “10ms”, the default, would indicate that there is room for

improvement.)

With Envy24-based PCI-sound cards, there may be an option in your sound card control panel that reads

“DMA Buffer Transfer Latency” (Seen with Terratec products) or similar. You should set this to the

lowest possible value, e.g. “1ms” for best results.