Windows or OSX?

@dhj - Isn’t that the truth. My mantra - many choices and the freedom to do so. GP really got it right in this area - it works as well on Mac as is does on Windows and vise versa. Doesn’t get any better than this IMO.

Looks interesting - Is that a one cable solution, i.e. just a single USB C to USB C

I use two thin cables an HDMI, because it works natively without any video driver which is nice when something goes wrong, and a USB C (in fact a USB A to USB C cable) for the touchscreen and the power supply.

1 Like

My experience.
You can easily tune a tower or desktop windows computer to produce real time audio. If you know what to do, including setting BIOS to disable energy management (making the CPU work always at maximum throughput). It’s not easy if you are not an expert but you can learn fast.
You CANNOT do it on new laptop windows computer, and this happens for ultra book compact devices which are so nice.
I think NUC are the best choice for audio, because all hardware devices are checked by Intel. I tested two or them and they were just OK.
I had disasters with two recent windows notebooks so frustrating that I switched to MacBookPro.
I strongly suggest to avoid windows if you need a notebook to produce audio.
I don’t like Apple for many things, but an M1 MacBookAir or Pro are perfect machines to play live.

2 Likes

My former small form factor PC was a NUC, but unfortunately I had difficulties to find a new model compatible with more than 32Gb and with a CPU having a base frequency close to 3GHz or more. That’s why I went to GEEKOM. Their form factor is close to the NUC form factor, but it is their own design, which I have to say, works pretty well.

This is the conclusion that I am coming to. You have limited hardware choices with Apple, but they work pretty well. The optimizations required in x86 notebooks make them less reliable for real-time audio.

I don’t think this is true. My HP ZBook can be booted into BIOS and there are settings you can change there and it’s more than capable of running GP and everythign in the optimization guide can be achieved with that laptop.

Just my experience.

With windows, a lot comes down to the individual manufacturer, what BIOS they put on the machine and how much cruft they install.

2 Likes

Also for me, the bios of my 2020 laptop is fully configurable for maximum performance, but as said above, I chose this model because it has this feature.

1 Like

Does anyone have a 2-in-1 or convertible laptop for recommendation? I hate that Mac laptops don’t have a tablet option or touchscreen, but still leaning towards one because it’s a sure bet for low latency processing,

That’s interesting. Is the performance good? I’m currently using a Lenovo M920q upgraded to M902x spec (Core i9 9900, 2Tb SSD, 32Gb ram) and may look for a newer alternative at some point in the future.

With regard to my former NUC PC, yes it is very good. And I now have 64Gb of RAM which changes from the 16Gb I had before.

3 Likes

Regarding DPC measurements. Is there a number or rule of thumb about was is acceptable? I am looking at a Surface Pro 8 with a rating of 489. That’s on the lower end, but not the absolute lowest (although it seems to be the lowest of the tablet/convertibles).

I tend to the opinion that these measurements won’t tell you much about the ability for live-audio/playing in realtime… because this ability will much depend on the audio interface, driver and settings you’re about to use, and they don’t take this in account at all when they talk about “latency” when playing a video from Youtube… that’s a whole diffrent world (which has not much to do with the requirements for a live musician).
Of course it won’t do any harm if you decide to buy a device with a general low DPC-Latency, but this won’t guarantee a glitch-less audio-performance if your audio interface, the driver and/or the settings are crappy…
just what i think.

2 Likes

Right, I guess assuming it’s a decent interface/drivers, just wanted to know if those numbers really mean anything. Obviously if there are crappy drivers and/or the machine is not configured properly that would swamp any DPC baseline. I can’t buy every machine available just to see if it has a good bios.

I have an Antelope Audio device, but would consider anything with low latency – like an RME Babyface.

The link you provide is not relevant for judging the performance of a processor for audio applications.
They use a balanced power plan when it should be a highest performance one, and the test is done by opening browser tabs and manipulating videos in youtube, which is not really relevant to a performance test of audio applications like GP.

If you are going to use an intel processor, choose a machine that has a processor geared towards the type of performance you are looking for (power saving, computational performance, etc).

Understanding Intel processor names

2 Likes

Let’s say that DPC latency performance is really necessary but not enough. You need good audio board and drivers too.
But in my experience a notebook with bad DPC latency will never be able to play a decent audio

You need to watch a video I shared previously.

You are right partially that the way they do their tests are not fully representative of they way we use our pcs/laptops but still, is a good benchmark about how optimized is an equipment for real-time audio. For example, if a laptop tested has been performing well with low dpc/isr, with balanced settings, i would expect to have similar or better performance with a high performance mode.

The other side of the story is the interface and its drivers, but as mentioned before, it doesn’t matter to have the fastest pc and the fastest drivers if the ecosystem has not good responsiveness

1 Like

Just sharing my results:

I just set up an M1 Macbook Pro with an Antelope Audio Synergy Core 4 (thunderbolt) and was able to get < 1ms latency using RTL.

Running GP and Line 6 Helix Native I can run easily < 4ms latency (in fact I haven’t pushed it to fail because 4ms is more than fast enough for now)

1 Like

Can you tell me what buffer size are you using?

48K / 128 samples.

It also works at 48K / 64 Samples

1 Like