StorageReview's strange results with Barracudas & Samsun

Lone Ranger · Post by **Lone Ranger** » Wed Sep 22, 2004 5:14 am

Hi guys, maybe I have been out of circulation for too long but recently I came across a review of the Samsungs at StorageReview done earlier this year. See http://www.storagereview.com/articles/2 ... 14N_1.html Go have a look.

They tested a Samsung 160GB IDE drive (SP1614N) against a Barracuda PATA and a few other drives. They said the comparison between Samsung & Barracuda was not a good one on grounds of performance due to interface/BIOS. But what I think is still valid are the heat and noise tests. These are at http://www.storagereview.com/articles/2 ... 14N_6.html and they show the Barracuda dBA quieter than the Samsung (as were a Hitachi Deskstar and a Maxtor). That was odd. But they did show the Samsung as the coolest.

Well, that's just a quirk of their tests you might say. But wait. Let's turn to their rather smart comparison chart at http://www.storagereview.com/comparison.html then choose "idle noise" from the dropdown button and then press "sort".

Here we see that all the quietest drives are Barracudas and then, only after a bunch of Maxtor’s, is a 5400rpm Samsung (SV1604N) as 14th quietest at 41.2 dBA. The SP1614N above is 23rd quietest. Weird. A 120GB series V Barracuda comes out as quietest. They should come round to my PC and listen as I have both these drives in it and the Barracuda is definitely not the quieter drive.

Well, before we have any more of this let's go back to that dropdown button at the top of the StorageReview page and select the last entry "Net Drive Temperature" and press "Sort". Yes, mate. That means you. Go and have a look. I'll wait. .... Now, here Samsung makes a better showing. But if we ignore the older Samsungs which are listed then that 160GB model comes out at only 8th quietest. The 120GB series V PATA Barracuda is one degree hotter. Oh my! Har har! Now you really should come round to check my PC.

I have been out of circulation for a bit and maybe this has been discussed here (although I couldn't find it in the forums). But joking apart, what is the matter with StorageReview? They seem way out in their results. Practical hands-on experience which we have here in this forum tells a vastly different story.

What is the matter with these people at StorageReview?

What do you think?

--- rgds Lone Ranger

Dobby · Post by **Dobby** » Wed Sep 22, 2004 8:28 am

Those differences obtained are really small at given distance and it would be nice to know also error margins of the measurements.

Also in a situation like this, nature of the noise could count more than actual decibel readings.
I would say there's nothing wrong with these people, they just have different priorities and noise measurement just hint level of the noise. In some sense they also have a correct approach, because in average system modern IDE drives IDE are pretty much equal in idle, unless the drive whines.

Post by **Rusty075** » Wed Sep 22, 2004 8:58 am

The real reason for the odd results from SR's acoustic testing is their sloppy methodology:

18mm from the drive is way too close. Their readings are contaminated by all sorts of nearfield effects. Of course, they have to be that close to get any reading with the testing hardware they use. Their SPL is only rated down to 30dBa. If they tested from a range that would avoid nearfield effects, (like 1m) their meter wouldn't register anything. Testing drives that have SPL's so close the lower limit of their meter, combined with ambient noise that is probably in the 20-25dBa range, will result in meaningless measurements. They have also arbitrarily selected a point on the side of the drive to point their mic at, apparantly with no thought given to the idea that different parts of different drives may produce different noises.

Combine measurements taken from too close, with equipment that's imprecise, and you get a gobblety-gook of miseading information. They were better off when they just subjectively ranked the drives.

Lone Ranger · Post by **Lone Ranger** » Sun Sep 26, 2004 2:10 pm

Rusty075 wrote:Combine measurements taken from too close, with equipment that's imprecise, and you get a gobblety-gook of miseading information. They were better off when they just subjectively ranked the drives.

I think you are dead right. Subjective measurements woul dhave been better than StoreageReview's over-precise laboratory-like readings which give their figures an air of false authority.

I just can't agree with their results at all.

Spod · Post by **Spod** » Tue Sep 28, 2004 7:48 am

They do have some validity - clearly, loud drives (like many SCSI models) register as louder in SR's tests. I agree that they don't really measure acoustics to SPCR's standards, but it does at least give a consistently measured reference - just not a very useful one when it comes to measuring small differences between drives. And it's debatable* whether idle or seek noise is more significant anyway.

*But I'm not going to debate it here.