Video Capture / Recording
20110304 By T. Hernandez
Hi to everybody,
Most of us know the benefits of Solid State Drives vs. traditional spinning ones, they have great features, they have the last in storage tech., they are fast, but most people don't know the disadvantages found on a lot of these drives. Yet for video.

First let me explain a litle bit about SSD, then as for video purposes, then show some tests, explaining what's the degradation & how affect video.

SSD vs. traditional spin hard disk drives

SSD Definition
   "A solid-state drive (SSD) is a data storage device that uses solid-state memory to store persistent data."
Its like a Virtual RAM drive/disk (not as much fast but fast), where on data keeps on. Information is storage permanently, not lose data when shut down the computer.

As it uses memory, this means like -in theory-, SSD use fast storage technology, well, latest yes, but fast generally not on both sides of the world.

Read VS Write
They are the sides of the storage world

Write versus read

As for Video recording purposes
Just yesterday read an aricle by Kenneth Hees saying SSDs are far better than mechanical disk, and he, as most other articles say, can be rated as True or Good for general purposes. Read rates from common SSD are awesome, compared to traditional drives. You just can boot OS and load e.g. say 30 applications, in less than 1/3 than on hdd.
I said for general purposes cause general SSDs on today's market are not well suited for profesional video recording purposes, let me explain.

Video capture uses the drive as in sequential write stress.
When you connect your camera to the computer to grab, capture video or transfer files, a sequential write is happening.

If you tipically read the same data then you need to lpook at the read rates, but if you're doing, recording, editing as a job you need to be carefull with write speeds and reliability of storage.

4 usages classify the drives: Random Read, Sequential Read, Random Write, and Sequential write.

Writing sequentially is writting continuous without intermediate seeks, where on random writes means a pattern of seek-write-seek-write ..

While your card is capturing video from external device, it records sequentially the captured stream/data directly to disk, meaning a sequential write is doing. Loading applications uses random read, playing videos uses sequential reads.

Profesional video recording at high data rates (whereon data rate is MB/s or Mbps and where on more high data rate means more quality) requires the features that are not the flagship of todays SSD, the SEQUENTIALS.
Video purposes usage sequential procedures

Where SSDs flagships are,
Acces time: this have no effect on sequential write (not random writes)
Read Speeds (both random & sequential): Capturing means write not read.

Some write the following, are true, and you need to understand
"The distinction is very important in traditional disk-based systems, where each disk seek will take around 10ms. Sequentially writing data to that same disk takes about 30ms per MB. So if you sequentially write 100MB of data to a disk, it will take around 3 seconds. But if you do 100 random writes of 1MB each, that will take a total of 4 seconds (3 seconds for the actual writing, and 10ms*100 == 1 second for all the seeking)."
Benchmark tools
Following is the list of today's common soft. bench tools for measuring transfer & data rates from HDDs
  • ATTO
  • Crystal Disk Mark
  • HDTune
  • As SSD

Lot of people doen't be affected by his unknowledgement cause they're using the system for common tasks, but professionals or extensive disk usage users note than manufacturers data rates (MB/s) not seems to be applied on his drives.

Like an example, you can't trust on the bench from BlackMagic for SSD drives, they give you not true sequential ratings.

We test some drives, and find that rating from B.M tools and ATTO, was very different than the rating provide by Crystal or AS SSD.

Benchmark tools as for video purposes
Believe in this toolsAs for video purposes, manufactures tends to provide data rates (ranking-scores) from the hightest availables(ATTO).
Remember to not believe on tools like the B.M. one that provide info about how many frame per second your drive/s are allowed to capture/storage depending on quality (resolution, color sampling, ..). As an example in our last test, we use some SSDs drives that this drive tool said allow for more frames per second than the really ones the drives was able to record. In the other hand, BM tools was right for the Velociraptor ones.

What i want to said is that if your think to buy SSD as for video capture purposes, you need to get indications from sequential benchs, forgetting soft. like ATTO that is the one used in today's SSDs manufacturer indications, otherwise you can dropp some or a lot of frames on your captures.

Why you think Corsair, OCZ, Patriot and even Intel show the same amaizing performance numbers ? cause they're based only on ATTO bench records.

Valid tools for video sequential tests are:

  • AS SSD

This week we tested some drives, fully tested as for video purposes, i will show you the sequential results.
the drives that we test this week
We tested: OCZ Vertex 2 25nm 60GB, Corsair F60, Corsair F120, Intel G25-M 160GB G2 and a Velociraptor 300GB.

How we make the tests
Computer: Intel i7 920, Gigabyte EX58-UD5, 6GB ram.
We test drive by drive, as a spare drive (not containing the OS, only for data), we install disk on a X68 AHCI mobo set BIOS on AHCI mode, then we formatt using Windows 7 disk management, the we do the following test.
  • 1st Ride bench tools
  • Record around 20-30gigs of uncompressed video
  • 2nd Ride again bench tools
  • Record uncompressed videos up to 90% of capacity
  • 3rd Ride again bench tools
  • Full Delete files (to get 100% free space)
  • Run TRIM command
  • Let the system in idle mode for 10 or more hours
  • On Intel run the optimizer
  • 4th Ride again bench tools

Records on both 1st and 2nd attempt were the same numbers, but on the 3 the degradation comes really, and in the last bench we only get onway with intel. This is what we get:
Model 1 Seq. write test 3 test (on degrade) 4 test after sanitize
OCZ VERTEX 2 60GB 25NM 35 MB/s
CORSAIR F120GB 145 MB/s 70 MB/s 92 MB/s
INTEL X25-M 160GB G2 110 MB/s 90 MB/s 110 MB/s
VELOCIRAPTOR 110 MB/s 100 MB/s 110 MB/s

Once we see the seq. numbers on OCZ 25nm, we return back the drive, horrible !
Second drive F60, we also return back cause 85MB/s is the same as traditional drives, an no one is able to record uncompressed video with.

As you may notice, all drives except intel & velociraptor don't recover its full performance after filling drive with lot of gigs.

Any case, you need to look at test 3 and 4, and so on at this time, no one is really suitable for uncompressed video. Don't worry there're solutions!

CONTINUE to 2nd page