[Hueyduck]

Hi everyone,

I'm planning to build a new PC soon and there are several points that
are unclear to me.
For the moment, I use a system where the sample libraries are stored in
2 IDE HDD connected on the RAID chipset of my motherboard (KG7Raid). I
connect the 2 HDD only when I want to make music.

For my next PC, I would like to use SSD drives.
What I imagine is an external "box" that would contain several SSD
units (starting with one since the SSD are so expensive nowadays). The
box would be connected to the PC via e-sata.
I would switch the box on only when I start my computer to make music.

What do you think?

I don't know much about the SATA system. But I'm learning.

I would love to hear about people who tried SSD to stream samples in
direct from disk mode (Kontakt, Gigastudio, etc...).
On paper, 0.1ms access time is quite tempting, right?

Thanks for your answers.


Huey


PS: I thought this NG was the more appropriate to post into. Please
correct me if there another NG that would be more suitable.

[Laurence Payne[_2_]]

On Fri, 08 Aug 2008 11:21:39 +0200, Hueyduck
wrote:

I'm planning to build a new PC soon and there are several points that
are unclear to me.
For the moment, I use a system where the sample libraries are stored in
2 IDE HDD connected on the RAID chipset of my motherboard (KG7Raid). I
connect the 2 HDD only when I want to make music.

For my next PC, I would like to use SSD drives.
What I imagine is an external "box" that would contain several SSD
units (starting with one since the SSD are so expensive nowadays). The
box would be connected to the PC via e-sata.
I would switch the box on only when I start my computer to make music.

If you've got ample RAM, disk performance may not be that much of an
issue - less "streaming" may be going on than you imagine.

[Hueyduck]

Laurence Payne a écrit :
On Fri, 08 Aug 2008 11:21:39 +0200, Hueyduck


If you've got ample RAM, disk performance may not be that much of an
issue - less "streaming" may be going on than you imagine.

Hi Laurence,

Actually, the one thing I'm not ready to do is switching to Vista & 64
bit system.
I will stick to WinXP 32bit for now since it works quite well.
With this, I'm limiting myself to 3.5Go of RAM.

And I do compose with libraries that need to be streamed.
I have to agree with you that many samples don't need to be streamed at
all. But some do :-) Large orchestral programs really does, in my opinion.

And SSD is realy getting closer to RAM, as far as access time is
concerned. Transfer rate is far lower than RAM, but I think transfer
rate is not a major issue to stream samples. Access time is.

btw: nice to read you on this NG too ;-)

Huey

[Nick Brown]

Hueyduck wrote:
Laurence Payne a écrit :
On Fri, 08 Aug 2008 11:21:39 +0200, Hueyduck


If you've got ample RAM, disk performance may not be that much of an
issue - less "streaming" may be going on than you imagine.

Hi Laurence,

Actually, the one thing I'm not ready to do is switching to Vista & 64
bit system.
I will stick to WinXP 32bit for now since it works quite well.
With this, I'm limiting myself to 3.5Go of RAM.

And I do compose with libraries that need to be streamed.
I have to agree with you that many samples don't need to be streamed at
all. But some do :-) Large orchestral programs really does, in my opinion.

And SSD is realy getting closer to RAM, as far as access time is
concerned. Transfer rate is far lower than RAM, but I think transfer
rate is not a major issue to stream samples. Access time is.

btw: nice to read you on this NG too ;-)

Huey

Huey,

First off, I haven't tried streaming samples (in the musical instrument
sampler sense) from an SSD.

Disk drive access time is (on average) in the region of 10ms, and
depending upon what other disk activity there is, (which could just be
streaming other samples - i.e. polyphony) could be much worse. At first
glance, that would appear to make disks unsuitable for streaming samples
at all. The trick (that IIRC Gigasampler used first) was to cache the
start of each sample in RAM, so it could start playing the sample as
soon as the MIDI note-on message arrived, and the disk would catch up
and deliver the rest of the sample before the cache was exhausted.

So it would seem to me that with that technique, which I thnk all the
major software samplers use now, access time is no longer a major issue.

Whether transfer rates are an issue for SSDs, I don't know. There's
probably quite a bit of variation from one SSD to the next, far more
than is the case between different models of hard drive. If you do end
up going down the SSD route, I'd be interested to hear how you get on.

Cheers,
Nick

[Hueyduck]

Nick Brown a écrit :
If you do end
up going down the SSD route, I'd be interested to hear how you get on.

Cheers,
Nick

I will post my comments the moment I try it, be sure of that. I'm so
curious for month now...

Huey

[Hueyduck]

Soundhaspriority a écrit :
"Hueyduck" wrote in message

For my next PC, I would like to use SSD drives.
What I imagine is an external "box" that would contain several SSD units
(starting with one since the SSD are so expensive nowadays). The box would
be connected to the PC via e-sata.
I would switch the box on only when I start my computer to make music.

Why? There is no advantage to making it external. ESATA connectors have a
tendency toward intermittent faults. The SSD lifetime is determined by write
cycles, not power-on-hours.


I wasn't so much concerned about the SSD lifetime as I was by "good
sense", if I may say. I don't like the thought of switching ON
something I don't need.

With my current PC, I just turn on 2 HDD in the raid array and start
working.
You could think that I could do the same with SSD, butsince they are so
small (256Go seems to be the maximum a normal human being could afford),
I was thinking about buying more SSDs as their prices go down.

I'm taking good note of your ramrk about eSata relyability, though.

What do you think?

I don't know much about the SATA system. But I'm learning.

There is nothing to learn. It's electrically equivalent to an internal
drive. All you need a

1. Adapter plates to bring out the motherboard connector:
http://www.newegg.com/Product/Produc...82E16812119021

2. An external enclosure.

This is where I haven't figured out how I'm gonna so this.
From what I read, SATA works like this: one connector for one cable for
one drive.
If I have 2 SATA ports brought out of the motherboard, I would only be
able to connect 2 drives. But I would like to be able to connect more
than that.
I understand that the transfer rate will be a limiting factor, but as I
doubt I could reach this limit while streaming samples from SSDrives.

(...)
Well, as I was searching the net to illustrate my thoughts in this post,
I found this:

http://www.addonics.com/products/hos...ad5sahpm-e.asp

This is exactly what I wanted, or am I missing something there?
One big bow able to contain 5 drives, the box onlmy needing one SATA port.
Amazing. Well, the product seems to be discontinued, but I will go onb
searching.

When I read this page:
http://www.sata-io.org/portmultiplier.asp
I learn that , in order to be really interesting for me, I must have a
Motherboard with a SATA controller that supports FIS (frame information
structure) based switching.
This would theorically a polyphony of roughly 1500 voices (with SSD
drives, of course).


Stick the SSD inside the enclosure. KISS

I will

Huey

[John Phillips]

Hueyduck wrote:
Hi everyone,

I'm planning to build a new PC soon and there are several points that
are unclear to me.
For the moment, I use a system where the sample libraries are stored in
2 IDE HDD connected on the RAID chipset of my motherboard (KG7Raid). I
connect the 2 HDD only when I want to make music.

For my next PC, I would like to use SSD drives.
What I imagine is an external "box" that would contain several SSD
units (starting with one since the SSD are so expensive nowadays). The
box would be connected to the PC via e-sata.
I would switch the box on only when I start my computer to make music.

What do you think?

I don't know much about the SATA system. But I'm learning.

I would love to hear about people who tried SSD to stream samples in
direct from disk mode (Kontakt, Gigastudio, etc...).
On paper, 0.1ms access time is quite tempting, right?

Thanks for your answers.


Huey


PS: I thought this NG was the more appropriate to post into. Please
correct me if there another NG that would be more suitable.



It is just too expensive for audio or video because you create a lot of
large files. Apple and MS OSs are not optimized for SSD yet (writes
mainly), I think that Apple may be first with new drivers. I believe
that it will be ready in the next year or two but it will still have the
"wear out" problem possibly with the Windows swap file, I do not know if
Apple has something similar. I think that you may be able to put the
swap file on a different drive if I remember correctly. With RAID
drives or the latest fast rotating drives, the bottleneck is probably
with how many plugins you can run at once so I would spend that money on
CPU and RAM.

Just my thoughts,

John Phillips
** Posted from http://www.teranews.com **

[Hueyduck]

John Phillips a écrit :

Thank you for answering,

It is just too expensive for audio or video because you create a lot of
large files. Apple and MS OSs are not optimized for SSD yet (writes
mainly), I think that Apple may be first with new drivers.

Like I said, I intend to use SSD to store (once and for all) and read
(many many times) sample libraries for direct from disk samplers. There
would never be data written apart from the first time I install the
libraries.
The bottleneck of dfd samplers being mainly access time, SSD seemed to
be really interesting.
I would not use the SSD to access and write instrument take, bounces or
session files. Never.

Huey