For you firewire guys out there

[rtcstudio]

I knew this already, but it was really brought home to me this past week.

Among other things at a tracking session going wrong, I noticed that my system seemed REALLY sluggish. I was only recording about 16 tracks (Drums, Bass, Electric guitar), but everytime I hit play with those tracks in record mode and QUICKPUNCH enabled it would literally take at LEAST 10 seconds for play to initiate. I got the rapping fingers for that long.

Well, I found out that my assistant had placed an order for the IBM 60 Gig 7200 RPM drives I usually use, but the order had not gotten here in time, so he went to CompUSA and got Western Digital drives off the shelf for my session.

As soon as I got the IBM's and copied all my recordings to it, I opened up one of the sessions, enabled record on the same tracks (16), enabled quickpunch and hit play. No rapping fingers and play started in about 2 seconds or less.

Goes to show that not all 7200 RPM drives are created equally. I know the WD drives would probably cut my track count down as well, and would kludge faster once I started editing. They just can't handle the load as well as the IBM's.

[Bookerv12]

Hi,

I had the exact same thing happen;
Western Digital and IBM 60 Gig.

Thanks for the post.
I thought I was crazy.

Booker

[Lee Blaske]

Hey, let's not be too hasty in condemning WD drives!

I'm not sure which particular WD drive you were using, but if you go to www.barefeats.com, you'll see that they are very similar.

Here's a good link with other related links:

http://www.barefeats.com/hard22.html

My guess is that the firmware for your Oxford 911 bridge needs to be updated with the latest definitions. The WD drives you're attempting to use might not have been included in the firmware version you're currently running.

If you're using Granite Digital enclosures, the firmware updaters are on their site.

http://www.granitedigital.com/custsup port/openingpage.htm

If you take a look at the firmware revision history, you'll notice some recent revs. that dealt specifically with WD drives.

At any rate, I hope that's the problem in both of your cases. Firewire devices seem so plug and play that some folks forget that the bridge chip firmware needs attention, especially if you're popping new drives into older cases.

Lee Blaske

[rtcstudio]

Originally posted by Lee Blaske:
Hey, let's not be too hasty in condemning WD drives!

I'm not sure which particular WD drive you were using, but if you go to www.barefeats.com, you'll see that they are very similar.

My guess is that the firmware for your Oxford 911 bridge needs to be updated with the latest definitions. The WD drives you're attempting to use might not have been included in the firmware version you're currently running.


If you take a look at the firmware revision history, you'll notice some recent revs. that dealt specifically with WD drives.

At any rate, I hope that's the problem in both of your cases. Firewire devices seem so plug and play that some folks forget that the bridge chip firmware needs attention, especially if you're popping new drives into older cases.

Lee Blaske

<font size="2" face="Verdana, Arial">I didn't want to be hasty condemning WD drives so I went to the Granite site, and I was a revision or so behind (thanks for reminding me, Lee). I had updated my drives once, but it was many months ago. So I updated all my bridgeboards to the latest revision.

I then went and re-did my test. With 16 tracks in record ready and Quickpunch enabled the Western Digital drive was still at LEAST 5 seconds behind the IBM in initiating play. The IBM initiated play in 2 seconds. The Western Digital took just over 7 seconds. Same session, different drives. The Firewire Update DID shave 3 seconds off the already poor performance of the WD drive (it used to take 10+ seconds to initiate play).

This was a NEW Western Digital hard drive, 60 gig, 7200 RPM drive (the current model CompUSA is selling). They are just not as fast as the IBM's. I've run my own benchmarks on them using Hard Disk Speed Tools and the IBM always beat the Western Digital.

I wish it wasn't so, because you can walk into CompUSA and see an entire wall filled with WD drives. I have to order the IBM's online and wait a day. But until somebody comes along with a faster drive, if you want maximum performance from firewire, you'll get the IBM Deskstar 60 Gig 7200 RPM drive.

[Lee Blaske]

Interesting.

I do know that there can be quite a bit of variance between differently sized drives in a given series. As a matter of fact, I'm not sure if the WD drives are as much of a technology sharing series as the IBM 60GXP or 120GXP series (or the black sheep 75GXP series).

I've got some WD 80 GB drives with the 8 MB cache showing up in a day or so (haven't decided if I'll use them for PT, or a second GS Windows machine I'm assembling). I've got 120 GXP 80 GB drives (not the 60 GB ones), so I'll do a comparison of those two drives in a Granite Digital case. I haven't seen much written about the WD 60 GB drives, compared to their 80 GB drives. Maybe their 60 GB drive is a dog, and they know it.

Out of curiosity, did you measure the time to start without QuickPunch being enabled? Also, what setting are you using in DAE, and what cache setting are you using in the Memory Control Panel? Also, what Mac are you using?

What's Fingaz's take on the situation?

Questions, questions, questions...

Lee Blaske

[Fokke van Saane]

A drive that is quite full and needs to be optimised takes a lot more time to get into record. You should do these HD comparision test with empty or optimised drives. You could also check the differences when you make a selection and then go into record (then PT knows how long the recording will be), or in Prefs enable a short maximum recording length. You should partition your drives to overcome this.

Fokke

[Mount Royal]

I missed the basic premise here: the Western Digital drive was a 60 GB IDE drive in a firewire enclosure, or it was running on an internal IDE 0 or 2 bus? Does it make any difference if it was running on an internal IDE bus whether it was a master or a slave?

John Caldwell

[Mark Haliday]

But until somebody comes along with a faster drive, if you want maximum performance from firewire, you'll get the IBM Deskstar 60 Gig 7200 RPM drive.

<font size="2" face="Verdana, Arial">something is unclear here : who makes the firewire enclosure for your IBM ? from what I know (I may be wrong), it's all third party (QPS granite etc...)since IBM only supplies the drives themselves (apart from small portable products)

[PiratePost]

I have a Western Digital 120 GB drive (8 meg buffer) on my internal ATA 100 bus (Dual 1 gig mirrored door version) . No problems...so far.

[rtcstudio]

Originally posted by Fokke van Saane:
You should do these HD comparision test with empty or optimised drives.
Fokke

<font size="2" face="Verdana, Arial">I humbly disagree with this statement, sorry. It's that kind of testing that makes one drive look as good as another drive ON PAPER. Testing an empty drive is not the same as a drive's performance three days into a session in the real world.

These drives had EXACTLY the same material on them. ABSOLUTE mirror images. They were brand spanking new drives, both formatted with the Apple format.

They were both 60 gig 7200 RPM drives in IDENTICAL Granite Digital HotSwap bays. They were not daisy chained to each other. They were both plugged into one of the two Firewire ports on my G4/733.

My DAE buffer setting is 2.

The WD drive is simply slower than the IBM in a ProTools recording environment. As I stated, this is not the first time I've come to this conclusion. Since Lee pointed out the Granite Digital revision issue, I was willing to take another look at the WD, and give it every opportunity to match the IBM's performance. It doesn't, and I don't think it can. The time difference initiating play is an indication of bigger problems. Once you get into heavy editing, the WD will kludge up, and your track count will go down.

I'm NOT saying the WD drive won't work at all with ProTools. I'm saying that if you want maximum performance, use the IBM.