What HD should I look at for Audio Editing?

[M.J.S.]

Looking for a new hard drive dedicated exclusively to Cakewalk Sonar's audio
folder (where all the audio swapping will go down during playback).

It will be the 3rd HD on the system, huge in size (500gb minimum) and
probably SATA (I've 2 SATA slots remaining).

What do you recommend I get? Raptor? Barracuda? Something else? Fastest seek
time should probably be #1 priority, right?

Again, this will be almost ONLY for real-time audio file reading/swapping.

 

[Rod Speed]

Looking for a new hard drive dedicated exclusively to Cakewalk Sonar's audio folder (where all the audio swapping will
go down during playback).

It will be the 3rd HD on the system, huge in size (500gb minimum) and probably SATA (I've 2 SATA slots remaining).

What do you recommend I get?
Samsung.

Raptor? Barracuda? Something else?
Samsung.

Fastest seek time should probably be #1 priority, right?

Nope, there is bugger all in it between the available
drives and it isnt important for your use anyway.

 

[M.I.5¾]

Looking for a new hard drive dedicated exclusively to Cakewalk Sonar's
audio folder (where all the audio swapping will go down during playback).

It will be the 3rd HD on the system, huge in size (500gb minimum) and
probably SATA (I've 2 SATA slots remaining).

What do you recommend I get? Raptor? Barracuda? Something else? Fastest
seek time should probably be #1 priority, right?

Again, this will be almost ONLY for real-time audio file reading/swapping.

Audio editing is hardly going to give the slowest of drives a work out.
Just buy the fastest drive you can afford.

 

[M.J.S.]

Audio editing is hardly going to give the slowest of drives a work out.
Just buy the fastest drive you can afford.

Not editing per se, but when you've got 50+ 24-bit wav files playing at the
same time in a mix, there's a whole lot of disk movement going on. In fact,
the disk activity often causes dropouts before the CPU usage does.

 

[Folkert Rienstra]

M.J.S. wrote in news:[email protected]

Not editing per se, but when you've got 50+ 24-bit wav files playing at the
same time in a mix,

Isn't that what editing is about?

there's a whole lot of disk movement going on. In fact,
the disk activity often causes dropouts before the CPU usage does.

And now you know the quality of this group's participants.
Nice troll, by the way.

 

[Arno Wagner]

Not editing per se, but when you've got 50+ 24-bit wav files playing at the
same time in a mix, there's a whole lot of disk movement going on. In fact,
the disk activity often causes dropouts before the CPU usage does.

Hmm. Raptors will not really help a lot there. Their access time
is barely half as high as that of a modern 7200 rpm disk.
It is only a gradual improvement. I think you need something
significantly faster.

Lets look at the volumes we are talking here (I assume 48kHz
sampling....):

50 * 3 * 48kB = 7.2MB/sec. That is 500MB/min.

Hmm. Maybe get a 16GB or 32GB FLASH drive? They do not
have the seek-issue, since theur access times are 50-100
times lower than that of a notmal disk. Even one or several
8GB USB flash drives may solve your problem.

Also, if your software supports read-ahead, maybe get more
memory?

The basic problem is of course, that the software makers screwed
up. If you buffer sensibly, reading 50+ files in parelell
is not an issue for these speeds.

Arno

 

[CJT]

Looking for a new hard drive dedicated exclusively to Cakewalk Sonar's audio
folder (where all the audio swapping will go down during playback).

It will be the 3rd HD on the system, huge in size (500gb minimum)

I wouldn't call that huge.

and

probably SATA (I've 2 SATA slots remaining).

What do you recommend I get? Raptor? Barracuda? Something else? Fastest seek
time should probably be #1 priority, right?

Again, this will be almost ONLY for real-time audio file reading/swapping.

That's not an especially taxing (or even necessary) task. Max out your
RAM and don't worry about the disk.

 

[CJT]

Not editing per se, but when you've got 50+ 24-bit wav files playing at the
same time in a mix, there's a whole lot of disk movement going on. In fact,
the disk activity often causes dropouts before the CPU usage does.

If you're serious about speed, go with multiple spindles. Putting it
all on one drive is what causes movement.

 

[CJT]

Hmm. Raptors will not really help a lot there. Their access time
is barely half as high as that of a modern 7200 rpm disk.
It is only a gradual improvement. I think you need something
significantly faster.

Lets look at the volumes we are talking here (I assume 48kHz
sampling....):

50 * 3 * 48kB = 7.2MB/sec. That is 500MB/min.

Hmm. Maybe get a 16GB or 32GB FLASH drive? They do not
have the seek-issue, since theur access times are 50-100
times lower than that of a notmal disk. Even one or several
8GB USB flash drives may solve your problem.

Also, if your software supports read-ahead, maybe get more
memory?

The basic problem is of course, that the software makers screwed
up. If you buffer sensibly, reading 50+ files in parelell
is not an issue for these speeds.

Arno

Frankly, I think trying to mix down 50 channels in one go is a recipe
for mud anyway.

 

[Arno Wagner]

Arno Wagner wrote:

Frankly, I think trying to mix down 50 channels in one go is a recipe
for mud anyway.

Hmm. Maybe put them into a combined stream first (i.e. a 50 track stream)
and then mess around with mixing?

Arno

 

[timeOday]

The basic problem is of course, that the software makers screwed
up. If you buffer sensibly, reading 50+ files in parelell
is not an issue for these speeds.

Arno

Exactly. If it's really only 10 MB/sec total, the software could cache,
say, 10s of each track at a time, thus seeking only once every 0.2s.
Assuming 10ms seek time, the drive would only be seeking about 5% of the
time, so it should hit very near its peak sustained transfer rate.

 

[Arno Wagner]

Exactly. If it's really only 10 MB/sec total, the software could cache,
say, 10s of each track at a time, thus seeking only once every 0.2s.
Assuming 10ms seek time, the drive would only be seeking about 5% of the
time, so it should hit very near its peak sustained transfer rate.

Exactly.

Arno

 

[mjs]

M.J.S. wrote in



Isn't that what editing is about?

Uh.. no. Editing a single WAV clip via Cool Edit is considered audio
editing, and it will not give your drive the slightest workout.

And now you know the quality of this group's participants.

I'm certainly getting to know the quality of its responders. ;-)

Nice troll, by the way.

Yeah, I could tell you were right away; but I admire your humility in
admitting it.

 

[mjs]

If you're serious about speed, go with multiple spindles. Putting it
all on one drive is what causes movement.

You mean spread out the audio data across multiple drives? I'm not even sure
the audio software would know what to do with this, given there's only one
target folder for the audio files specified.

I'm trying to keep it simple... ie, trying to decide which of the Raptor or
Barracuda or insert-name-here I should get. But I get the sense that the
brand names are all pretty much equivalent to one another.

 

[mjs]

I wouldn't call that huge.

And I would. Is it really a subject worth getting into a debate over?

That's not an especially taxing (or even necessary) task. Max out your
RAM and don't worry about the disk.

The RAM is another bowl of fish altogether. Conflicting reports about just
how much of it can be seen by WinXP x86, and what point there is adding a
3rd or 4th stick in there. Ugh. :-S

 

[mjs]

Nope, there is bugger all in it between the available
drives and it isnt important for your use anyway.

You mean like this one?
http://www.ncix.com/products/index.php?sku=18576&vpn=HD300LD&manufacture=Samsung

 

[Rod Speed]

You mean like this one?
http://www.ncix.com/products/index.php?sku=18576&vpn=HD300LD&manufacture=Samsung

Yep, or a bigger one.

 

[Folkert Rienstra]

mjs wrote in news:[email protected]

Uh.. no. Editing a single WAV clip via Cool Edit

is considered audio editing,

If you must insist.

and it will not give your drive the slightest workout.

That's not what I meant. No one keeps his music in 50+ tracks un-
less he is currently editing the piece or storing it for later editing.

I'm certainly getting to know the quality of its responders. ;-)

Yup, they all denied your experience.
Your current drive must be a real dinosaur.
Maybe it still has a steppermotor actuator?

Yeah, I could tell you were right away; but I admire your humility in
admitting it.

Thus confirms my suspicion, thanks.

 

[Folkert Rienstra]

timeOday wrote in news:[email protected]

Exactly. If

Exactly if? What's that, an 'absolutely maybe'?
Now apply that (question) to Cakewalk's Sonar.

it's really only 10 MB/sec total,

Copy lots of small files and 'only' suddenly becomes very relative.

the software could cache, say, 10s of each track at a time,

2MB = (a max of) 500 4kB clusters, not necessarily all end to end.
500 possible IOs that the OS may well re-schedule/break-apart.

thus seeking only once every 0.2s.

Assuming that all individual tracks are contiguous.
Now apply that question to Cakewalk's Sonar.

Assuming 10ms seek time,

Which is not the same as access time. Add 4ms for 7200rpm.

the drive would only be seeking about 5% of the time,

So more.

so it should hit very near its peak sustained transfer rate.

Probably still, yeah, assuming that tracks themselves are contiguous.

Question,
are tracks physicaly edited or are edits written as control information.
If the first they have to be written back as well.