[LAU] Affordable well-working USB interface with low latency at 48kHz?

[Giso Grimm]

On 23.11.20 14:55, David Kastrup wrote:
> David Kastrup <dak at gnu.org> writes:
>> Giso Grimm <gg3137 at vegri.net> writes:
>>> On 23.11.20 02:03, David Kastrup wrote:
>>>> "Jeanette C." <julien at mail.upb.de> writes:
>>>>
>>>>> Nov 22 2020, David Kastrup has written:
>>>>> ...
>>>>>> So I am lacking a reasonably solid recommendation of (possibly older but
>>>>>> reasonably quality) hardware with 1 or 2 mic inputs and +48V phantom
>>>>>> power and likely USB2.0 HiSpeed class compliant, possibly 2 headphone
>>>>>> outputs
>>>>> ...
>>>>> Is it worth investigating the Focusrite Scarlett interfaces? I have read
>>>>> here - a few weeks ago - that the 3rd generation is now supported on
>>>>> Linux. But the 2nd generation is still listed in some stores. The
>>>>> Scarlett 2i2 with two mic preamps +48V phantom power costs about 139EUR,
>>>>> the solo - with just one mic preamp - is less than 100 EUR . I think
>>>>> these smaller Scarletts used to be a favourite with RaspberryPis, as
>>>>> well.
>>>>>
>>>>> I don't know the latency, but many people speak very favourably about
>>>>> them.
>>>
>>> I recently measured latencies (analog to analog) for some sound cards,
>>> including the Focusrite Scarlett solo 3rd gen (2i2 should be the same):
>>>
>>> https://github.com/gisogrimm/ovbox/wiki/Soundcards
>>>
>>> From all the USB sound cards I tested the Focusrite Scarlett was the one
>>> with the shortest delay.
>>
>> Very good to know, thanks!  I'll also take a look at the rest of your
>> tests.
> 
> Question: is there a reason that you are going with buffer sizes of
> 3*2^n rather than 2^n?  I've not actually tried what Jamulus does in
> that situation (but will do) but it seems a bit strange to me.  Does it
> have some inherent advantage?

it is based on rumors that with the USB 1.1 audio protocol it latency is
lowest if the fragment size is a multiple of 1ms. I am quite sure that
this was a problem with USB 1.1 devices about 10 years ago.

Whatever jack supports should be fine, with the USB 2.0 devices I don't
see any differences. However, the latencies I measured behave like

 4*P + X

where with older cards I have very faint memories of achieving something
like 3*P + Y (although with a longer total delay, at least for short
period sizes).


> That being said: I was particularly looking for the Presonus Audiobox
> USB these days (since I guess Presonus cards newer than the lacklustre
> Firebox are supposed to have good preamps/gains) but with regard to
> latency, it doesn't seem like the greatest choice.

The only Presonus Audiobox USB I have at hand has a broken phantom power
supply, thus I cannot judge the preamps.


Best,

Giso