182,818 users have contributed to 42,261 threads and 254,931 posts.
In the past 24 hours, we have 3 new thread(s), 16 new post(s) and 44 new user(s).
I don't really know if adding hard drive will help to that but on some big templates, I have pops, clicks when several instruments play together. In this kind of projects, I use Play, Kontakt and of course VI2. So, the only thing I have to do is increase the MIR engine latency (up to 8000 in some case !)
So, I hope I could find a solution :)
Sonophile,
Speaking of harddrives.
I'd say you need to spread the load acoross multiple harddrives...
SSD's will improve many things, including how your 72 gigs of RAM, pre loads data.
Still, there's nothing wrong with good old fashioned HDD's, provided you have enough of them.
And Raid 0 is always a plus.
As for your clicks and pops.
@Mike B Studios said:
And Raid 0 is always a plus.
Not really. Most of the time, sample playback is limited by access speed, which limits simultaneous voice triggering. When combining several drives into a RAID-0, they all have to synchronize their data transfer, eliminating the benefit of using several drives.
The best performance ought to be have when striping the sample library up manually over your sample drives, 001.dat on drive 1, 002.dat on drive 2, 003.dat on drive 3 and so on. But it would be quite annoying. SImply distributing the sample content randomly over the drives should be enough.
Before going to a raid of SSD I had a raid of 6 x 32 GB Raptor 10 000 rpm with a Rocket raid card, it was day and night compared to an HD non raided
Martin did you do trial with Rocket Raid cards ?
@Mike B Studios said:
And Raid 0 is always a plus.
Not really. Most of the time, sample playback is limited by access speed, which limits simultaneous voice triggering. When combining several drives into a RAID-0, they all have to synchronize their data transfer, eliminating the benefit of using several drives.
The best performance ought to be have when striping the sample library up manually over your sample drives, 001.dat on drive 1, 002.dat on drive 2, 003.dat on drive 3 and so on. But it would be quite annoying. SImply distributing the sample content randomly over the drives should be enough.
Martin,
I believe one of my suggestions was simply spreading the sample load across multiple hard drives.
To your point, (for sample playback) how many drives in Raid 0 would you say represent too many? 4... 6?
I would think 2 drives are fine, or am I mistaken?
Before going to a raid of SSD I had a raid of 6 x 32 GB Raptor 10 000 rpm with a Rocket raid card, it was day and night compared to an HD non raided
Cyril,
Was it night and day compared to 1 HDD, or 6?
Before going to a raid of SSD I had a raid of 6 x 32 GB Raptor 10 000 rpm with a Rocket raid card, it was day and night compared to an HD non raided
Cyril,
Was it night and day compared to 1 HDD, or 6?
What it is said on professionnals forum is that you multiply by 1.5 the speed of reading by each disk added.
For six disk
Speed of the disk x 1.5 x 1.5 x 1.5 x 1.5 x 1.5
This is of course limited by the speed of the SATA channel
Splitting manually you libs across HD, this is hell ! Raid is doing it for you !
