Setting DAW Input and Output Buffer Sizes

[Amack]

This post contains the attachments for post #2. Please read post #2 first!

[Amack]

As shown below Ableton Live 9 Lite has a rather nice feature that facilitates the proper selection of audio device (Playback Engine) buffer size. Dropouts while listening to a generated tone indicate that the buffer size needs to be increased. There were no dropouts during operation with the configuration shown. I.e. with ASIO4ALL set for a 192 kHz sample rate and 64 sample buffer size and Ableton simulating 80% usage while restricted to using a single CPU core. Pro Tools 12.1 on the same Windows 8.1 desktop required the buffer size to be increased to 512 samples to prevent dropouts with a similar test signal while using all 8 CPU cores with no added usage simulation.



As shown below similar testing can be done with Pro Tools by creating an instrument track with a signal generator plugin/insert. The 96 kHz sample rate and 64 sample buffer shown below had an approximately 50% audio dropout rate. With a 128 sample buffer it had approximately a 5% audio dropout rate. It was dropout free with a 256 sample buffer. Details for these buffer sizes are included in the attached “ASIO4ALL Noise GeneratorC.pdf” (in post #1) along with higher resolution versions of these images.

[Amack]

To witness Pro Tools struggle communicating with the Playback Engine (as shown on the "System Usage" meters) it's not necessary to have anything on the track or have anything assigned to its input. The only thing necessary is to have the track's output assigned to a Playback Engine output. The apparent reason is that, as Bharath said at http://duc.avid.com/showpost.php?p=2...&postcount=13: "Even when idle, engine is still streaming data in and out of the device (albeit silence) and that i how both ASIO and Coreaudio systems work. CPU meter is just saying that engine is spending that percentage of its time period just pumping data in and out of the device."

[Amack]

If possible PT will use the same playback engine, sample rate, and buffer size used in its last session. So it's not even necessary to have any tracks in a session to see the impact of changing PT's "Playback Engine H/W Buffer Size" on it's "System Usage". The easiest way to do that on PCs is to use ASIO4ALL with one of the PC's audio devices.

Originally Posted by Amack

To witness Pro Tools struggle communicating with the Playback Engine (as shown on the "System Usage" meters) it's not necessary to have anything on the track or have anything assigned to its input. The only thing necessary is to have the track's output assigned to a Playback Engine output. The apparent reason is that, as Bharath said at http://duc.avid.com/showpost.php?p=2...&postcount=13: "Even when idle, engine is still streaming data in and out of the device (albeit silence) and that i how both ASIO and Coreaudio systems work. CPU meter is just saying that engine is spending that percentage of its time period just pumping data in and out of the device."

[mesaone]

You're making the mistake of assuming Pro Tools behaves the same way with all audio interfaces as it does with your audio interface. You're also making the mistake of assuming Pro Tools behaves the same way with all computers as it does with your computer. In actuality, it does not.

I shouldn't have to tell you that it's not good science to make pronouncements based on an the results of an ill-informed experiment with a sample size of 1.

[Amack]

What specifically? Can you demonstrate contradictory behavior?
I see the same behavior on three Windows 8.1 PCs with both PT 11.0.2 and 12.1.

Originally Posted by mesaone

You're making the mistake of assuming Pro Tools behaves the same way with all audio interfaces as it does with your audio interface. You're also making the mistake of assuming Pro Tools behaves the same way with all computers as it does with your computer. In actuality, it does not.

I shouldn't have to tell you that it's not good science to make pronouncements based on an the results of an ill-informed experiment with a sample size of 1.

[mesaone]

Aside from ASIO4ALL and your internal sound card (which you seem to be using as the control) how many different audio interfaces have you tested? From what I see here, that answer is "one".

[Amack]

How many would satisfy you?

Originally Posted by mesaone

Aside from ASIO4ALL and your internal sound card (which you seem to be using as the control) how many different audio interfaces have you tested? From what I see here, that answer is "one".

[mesaone]

This isn't about my satisfaction. This is about your habit of making broad assumptions and conducting ill-informed "tests."

It's like you started with a conclusion and worked backward.

[musicman691]

Originally Posted by mesaone

This isn't about my satisfaction. This is about your habit of making broad assumptions and conducting ill-informed "tests."

It's like you started with a conclusion and worked backward.

Don't let amack bait you into another one of his 'down the rabbit hole' excursions. He's a solution in search of a problem (or he has too much idle time on his hands).

As Mark Twain said:
“Never argue with a fool, onlookers may not be able to tell the difference.”
Or the more recent George Carlin profferred this advice:
“Never argue with an idiot. They will only bring you down to their level and beat you with experience.”