[ubuntu-studio-devel] i386 EOL plans for Ubuntu Studio?
len at ovenwerks.net
Thu Jul 21 14:49:58 UTC 2016
On Tue, 19 Jul 2016, lukefromdc at hushmail.com wrote:
> One issue in audio is that huge numbers of older 32 bit machines considered
> "slower than smartphones" for huge, JS heavy websites and Internet video are
> still as good for audio work as they ever were. The very first Pentium 4's were
> able to run things like Audacity with no video rendering glitches for realtime
> spectrum display, and back on 2004 I even did audio news reports on an
> original Pentium laptop and in another location on Pentium II's. On those
> machines realtime spectrum display lagged badly but the recordings were fine.
Another thing to concider is PCI based audio interfaces... there is no
replacement of reasonable performance (performance in the audio world does
not equal speed BTW) at a reasonable price. It is quite possible with a 10
year old or more P4 to have sub ms latency performance with a PCI RME or
delta series audio IF. While USB AIs have improved the latency is not even
the same from one startup to the next, even in audio IF costing more than
the computer. So a new computer with only PCIe slots means the user is
unable to use their older but still servicable... and actually better
performing audio inteface. This has made audio users less than happy about
upgrading to new faster, but poorer performing computers. Any of the PCIe
audio interfaces I am aware of are in the range of double the cost (or
more) of a good computer (i5 or so).
I personally use a D66 PCI audio IF. There are still a lot of people using
the 1010LT as well. People who have bought the 1818vsl as a replacement
have had trouble with USB3 drivers in linux and have had to black list
USB3 drivers to get them to work. They still have the odd - upgrade system
new kernel now audio doesn't work kinds of troubles too.
People often make the faster is better for audio statement, but faster
through put does not mean better audio performance. Turning "Boost" and
hyperthreading off still gives better performance in audio for example.
Running a CPU forced to 800Mhz will give better audio performance than
running the same cpu in powersave or ondemand at 3.2Ghz.
Performance in audio means how low can the latency be with no xruns. A
single xrun is considered a failed take when recording. Glitches in the
audio on stage will quickly make one consider an alternative (Other people
will suggest a Mac).
Speaking of hardware and performance, the laptop seems to have hit a limit
where newer HW is not really any faster. Yes, I can get an i7 in a laptop
but it is only 2 cores with limited speed and for audio work HT is turned
off. How much faster is it than a 32 bit 2 core machine really?
Anyway, it seems 64bit only will be forced on us in any case it is more a
matter of how soon than anything else. Ubuntu is for "mainstream" use in
North America... and (semi)profesional audio use does not really fit.
More information about the ubuntu-studio-devel