I’ve been using the same 2 towers for about 10 years and recently bought a new laptop and have been migrating my work to the laptop. It’s been a long time since I’ve configured a system for Linux audio. I’m wondering… the laptop has an ssd and and I’ve been running the projects in my home directory. Would I be better off making a dedicated partition on the ssd or possibly use a usb hd?
Short answer is probably not. You could benefit from external USB SSD, but honestly, most SSDs these days should be more than enough to handle both your computer and the recording.
1 Like
I’d agree, for pure audio recording, or for (most?) software-based synths.
However, larger sample-based VSTi’s (think orchestral programming) are super IO-hungry: putting them, or at least their sample audio files, on a separate SSD from the O/S is an easy (and cheap) bit of insurance (and lets you move them between machines easily, another benefit).
2 Likes
Hi,
I always have all my data on a second ‘DATA’ partition at the very least and when possible on a completely separate drive(s), If you want to switch Distros or things go irreparably south then you can reinstall be up and running literally in minutes. I keep the OS partition and the home folder just big enough for some breathing room and to use as a downloads holding area but any Audio recordings, Video, Graphics and Office Docs projects are never on the same partition as the OS. As has been said I don’t think i/o speed with an SSD could even possibly be a factor, really an older 7200rpm SATA HDD would be plenty fast to have the Program and Data running simultaneously and SSD is significantly faster.
If anyone cares here are my suggestions from the AV Linux User Manual (Page 15):
1 Like
Since I am not on my phone typing now, just to clarify a thing or two:
A Spinning Disk Drive I WOULD still recommend using a separate drive myself. The reason is that when it accesses different parts of the drive (Say too write audio in one section while paging in and out of memory for another or reading samples in another) the performance drops exponentially. It is because it takes times for the heads to shift back and forth to the different areas of the platter[s] that it might need to interact with. So putting audio data on it’s own drive in those cases can make a huge difference.
That being said, Glen’s reasoning that it allows you to change distros etc. is worthwhile especially in terms of partitions. It isn’t about speed in that case, but that really just depends on how likely that is to be you that switches distros or does other operations that might need similar. A partition seperates out the data and protects it to a certain extent in those cases.
While I am not going to live or die on this hill and it certainly isn’t going to hurt, I can’t remember the last time I needed to do that even for synths, but granted I am not an orchestral composer etc. While I have used larger synths I don’t often have to deal with 20-30 going at the same time for instance, enough for live performance (Not me playing but I have set them up) and a few for some basic programming (That time it might be me:).
I also do have to toss in a caveat to something I said:
BEWARE cheap SSDs. While they are much better than before, older/cheap SSDs often had problems with performance as they filled up (Or were written enough to) if they didn’t support TRIM commands, etc. They often cheated a bit by not actually clearing sectors when deleted so that when you had to go back to overwrite those sectors, it had to wait for them to be cleared before it could write to them, tanking performance bad.
That being said I haven’t heard of this being a problem in some time, but is why i never go cheap on SSDs personally.
Seablade
1 Like
While I am not going to live or die on this hill and it certainly isn’t going to hurt, I can’t remember the last time I needed to do that even for synths, but granted I am not an orchestral composer etc. While I have used larger synths I don’t often have to deal with 20-30 going at the same time for instance, enough for live performance (Not me playing but I have set them up) and a few for some basic programming (That time it might be me:).
FWIW, I’m specifically talking about virtual instruments which load many samples from disk (as in an orchestral library). Synths which generate sounds via programmatic code / parameters are literally the other end of the spectrum.
My advice follows similar patterns on other operating systems: if your instrument depends on loading sounds from disk, then put the samples it needs on the fastest possible hard drive.
1 Like
I understand, I was referring to the same. In my response synth was a bad word choice on my part.
Yea I haven’t used many orchestral libraries on Linux, so what I am referring to was more general advice based on my experience particualrly in Mac OS in this case.
Seablade
1 Like
This topic was automatically closed 91 days after the last reply. New replies are no longer allowed.