Latency! with Focusr. Scarl. 2i4 2. gen on old Notebook 1.8GHz,2GB RAM

Thanks for the reply!

That sounds like the bitter truth....

Still... Anybody ever used the Native Instruments Traktor Audio MK2 with Pianoteq? My beloved ancient Sony Vaios last chance....

I might end up and just get a Macbook ...

For the latency part and drivers... Windows sucks....

Or install Linux ?

What model of CPU do you have ?  ( e.g. i3-6100U, a designation like that ).

The headline clock speed is not a good guide to the available computing power.  Very similar designations can have very different performance, and different processors with the same clock speed can have quite different performance figures.

And regarding Apple kit,, my standard warning is that you are paying a considerable premium for just the name.  The same hardware in a PC laptop is typically a lot cheaper.

Hi TJ76,

I tried to estimate the latency of my Focusrite 2i2. I played and recorded a click-track at the same time with this USB-Audiointerface. The time difference between send and receive divided by two gives a vague idea of the output latency.

I set the recording-buffer in audacity to 5 ms, seems to be the minimum to get no drop-outs. As shown in the picture, the click returns 21 ms later (round-trip-latency). Divided by 2 the output-latency could be ~10 ms.

Considering the 5 ms recording-buffer the round-trip could be 21 ms - 5 ms = 16 ms.

Divided by two 8 ms could be a reasonable output-latency of the 2i2.

I used a slow Acer netbook with Celeron N3050 1.60GHz, 2GB RAM, Debian/Sid Linux, audacity 2.1.2, 24bit/48kHz,battery-powered, no performance tuning, Debian kernel 4.9.2-2.

Cheers

Just for your general information regarding your CPUs :

* Pianoteq seems to use 2 thread and having more than two cores makes no difference to Pianoteq (althought they can be useful for other apps running simultaneously).  There's a benchmarking thread somewhere in the forum where someone did exhaustive testing to find this out, but I can't find it right now.

* As a rough guide to performance I use the Single Threaded Passmark ( https://www.cpubenchmark.net/singleThread.html ) as a basis when two or more cores are available.

* Your N3050 has a single threaded passmark of 467

* Your i5-2467M has a single threaded passmark of 942 ( nearly double, despite running at the same clock speed as the N3050 ! )

* The N3050 has two cores, and the i5-2467M has 2 physical cores with 2 logical cores per physical core.  The physical cores are all that matters for Pianoteq, but overall the logical cores add about 20% more performance to the i5 over the N3050 and the normal multi-threaded passmark gives the N3050 about 900 and the i5-2467M about 2300 ( notice how the logical core - hyperthreading - raises the overal i5 performance by more than the factor of 2 the physical core alone would ).

So you would expect that not only would the i5-2467M performs much better on Pianoteq alone, but it would have enough horsepower to do more background work, which may help latency (a bit).

As a footnote I run Pianoteq on both a laptop (similar to your but an N2840) and a desktop (with a Zeon E3-1240) under Linux Mint 17 and 18 with no problems.  I haven't noticed any latency issues but I note that when I read about them they're usually involving USB audio devices and that seems to be a common issue with these things.  If I was adding an audio device like that I'd almost certainly try an RT Kernel or an audio specific distribution and get a PCI based card (like an RME) but I've no direct experience of these myself.