Sorry, I'm not a fan of these guaranteed "minimum sustained write speeds"
I also need "consistency" between identical products
Regardless of brand name, I consistently find Toshiba Flash products either rating their read speeds at double what I find in actual use (PNY/Mushkin} or Heavy Throttling (Mushkin Enhanced Ventura Ultra) or inconsistency of performance between identical thumb drives (Sandisk Extreme)
The only Thumb drive using Toshiba Flash that I have ever seen consistency of performance on is the Corsair GTX but that might be due to Trim and garbage collection
The new SanDisk Extreme Pro 256 GB Flash Drive with USB 3.1 "Might" be as consistent as the Corsair but I'm not spending $80 more for a thumbdrive with 30MB/sec slower reads just to find out
Surprisingly, I have yet to kill a single thumb drive using Toshiba Flash but I'm still trying!
Is this Company doing ANYTHING about consistency of performance in their products instead of looking just at minimums which appear to change over time in certain brands like Sandisk?
It may be that the 3 (Fixed Disk) Sandisk Extreme Thumb drives I purchased had different firmware revisions (or maybe not)
What was troubling was that they all started life at the same performance levels (+/- a few %) and slowly degraded over time to 3 distinct performance levels 1 is still quite fast, 1 is so-so, and 1 is very slow
The only way to temporarily return them to anything close to new performance levels is to wipe them "completely" with Killdisk
This is because they do not have Trim or garbage collection and erased data remains on the drive when using them for Windows2Go which progressively gets worse over time
Erased data MUST be wiped to regain performance, yet each of the 3 thumb drives "consistently" degrade to 3 distinct levels of performance (bad/better/best)
"...this is for SD cards, not thumb drives." ------------------------------------------------------ Regardless, it still sucks buying an SD card rated at 20MB/sec write speed and getting frame drops on a camcorder while recording at 1.5MB/sec (PNY)
The problem is the Flash Memory Used, not the type of memory card!
Isn't SSD performance usually dictated by the controller? What is different between the flash memory used in SD cards and SSDs that make the memory chips themselves the limiting factor in the case of SD card performance consistency?
"Isn't SSD performance usually dictated by the controller?" ------------------------------------------------------------------------------ The controller, the firmware and the type of flash The Sandisk thumb drive might have the same type of flash but the controller and firmware can distribute the read/write speeds differently
The Sandisk reads 30MB/sec slower than the Corsair but the Corsair writes 30MB/sec slower than the Sandisk
Someone else will need to tell you the difference between Flash for SD cards and Flash for SSDs because I'm too lazy to Google it and need to get back to my video games....
Sandisk extreme thumb drives do NOT support trim or garbage collection while corsair does. That means in the used state sandisk will drop more than corsair
Sandisk extreme thumb drives do NOT support trim or garbage collection while corsair does. That means in the used state sandisk will drop more than corsair
Even 4K videos compressed with H.264 can make do with around 100 Mbps (~12 MBps) bitrate. HEVC should be able to deliver more efficiency. Quadruple that for 8K videos.
90 MBps is more than enough (translates to 720 Mbps) for compressed video. If you are recording RAW, you need CFast 2.0 or a proper SATA SSD (or, even NVMe PCIe drives)
Professional video recorders typically use Apple ProRes or DNxHD (DNxHR for above 1080p) rather than h.264 or HEVC. They're both intra-frame-only codecs, meaning every frame is compressed independently, sort of like MJPEG, so they have massive bitrates.
ProRes 422 HQ is 176 Mbps for 1080p24, so 4K24 would be 704 Mbps, or 88MB/s.
So, yeah, 90 MB/s is just barely enough for 4K, assuming you don't want to use more than ~24 FPS, and assuming you don't want to use ProRes 4444 or 4444 XQ. But at that point, you'll probably want to record to an SSD rather than SD.
It really depends on the bit rate and recording format. GoPros use around 60Mbit/s for their 4K recording while a Canon C300 recording 4k RAW can use over 200Mbyte/s.
Thanks. With that as context, and assuming Go Pro's 4k represents the approximate floor for mainstream 4k recording; its bitrate lines up with the bottom end of the tail of the 4k bar while leaving plenty of headroom for higher quality options above it. Recording RAW 4k blows the top; but recording RAW video's a pro feature while SDcards have always been aimed at the cheaper mainstream consumer market.
While it's true that most pro gear will record directly to SSDs (typically in at least a custom enclosure, if not an entirely custom SSD), lower-end pro recorders (like in the $500-1000 range) sometimes record to SD, as do some ENG cameras.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
16 Comments
Back to Article
Bullwinkle J Moose - Monday, January 23, 2017 - link
Sorry, I'm not a fan of these guaranteed "minimum sustained write speeds"I also need "consistency" between identical products
Regardless of brand name, I consistently find Toshiba Flash products either rating their read speeds at double what I find in actual use (PNY/Mushkin}
or
Heavy Throttling (Mushkin Enhanced Ventura Ultra)
or
inconsistency of performance between identical thumb drives (Sandisk Extreme)
The only Thumb drive using Toshiba Flash that I have ever seen consistency of performance on is the Corsair GTX but that might be due to Trim and garbage collection
The new SanDisk Extreme Pro 256 GB Flash Drive with USB 3.1 "Might" be as consistent as the Corsair but I'm not spending $80 more for a thumbdrive with 30MB/sec slower reads just to find out
Surprisingly, I have yet to kill a single thumb drive using Toshiba Flash but I'm still trying!
Is this Company doing ANYTHING about consistency of performance in their products instead of looking just at minimums which appear to change over time in certain brands like Sandisk?
Bullwinkle J Moose - Monday, January 23, 2017 - link
NOTE:It may be that the 3 (Fixed Disk) Sandisk Extreme Thumb drives I purchased had different firmware revisions (or maybe not)
What was troubling was that they all started life at the same performance levels (+/- a few %) and slowly degraded over time to 3 distinct performance levels
1 is still quite fast, 1 is so-so, and 1 is very slow
The only way to temporarily return them to anything close to new performance levels is to wipe them "completely" with Killdisk
This is because they do not have Trim or garbage collection and erased data remains on the drive when using them for Windows2Go which progressively gets worse over time
Erased data MUST be wiped to regain performance, yet each of the 3 thumb drives "consistently" degrade to 3 distinct levels of performance (bad/better/best)
A5 - Monday, January 23, 2017 - link
...this is for SD cards, not thumb drives.Bullwinkle J Moose - Monday, January 23, 2017 - link
"...this is for SD cards, not thumb drives."------------------------------------------------------
Regardless, it still sucks buying an SD card rated at 20MB/sec write speed and getting frame drops on a camcorder while recording at 1.5MB/sec (PNY)
The problem is the Flash Memory Used, not the type of memory card!
Same problem regardless.....
londedoganet - Monday, January 23, 2017 - link
Isn't SSD performance usually dictated by the controller? What is different between the flash memory used in SD cards and SSDs that make the memory chips themselves the limiting factor in the case of SD card performance consistency?Bullwinkle J Moose - Monday, January 23, 2017 - link
"Isn't SSD performance usually dictated by the controller?"------------------------------------------------------------------------------
The controller, the firmware and the type of flash
The Sandisk thumb drive might have the same type of flash but the controller and firmware can distribute the read/write speeds differently
The Sandisk reads 30MB/sec slower than the Corsair but the Corsair writes 30MB/sec slower than the Sandisk
Someone else will need to tell you the difference between Flash for SD cards and Flash for SSDs because I'm too lazy to Google it and need to get back to my video games....
drajitshnew - Tuesday, January 24, 2017 - link
Sandisk extreme thumb drives do NOT support trim or garbage collection while corsair does. That means in the used state sandisk will drop more than corsairdrajitshnew - Tuesday, January 24, 2017 - link
Sandisk extreme thumb drives do NOT support trim or garbage collection while corsair does. That means in the used state sandisk will drop more than corsairtamalero - Monday, January 23, 2017 - link
90MB/s for 8k video? doesn't 4k need more than 120MB/s?DanNeely - Monday, January 23, 2017 - link
Depends what your frame rate and compression level is...ganeshts - Monday, January 23, 2017 - link
Even 4K videos compressed with H.264 can make do with around 100 Mbps (~12 MBps) bitrate. HEVC should be able to deliver more efficiency. Quadruple that for 8K videos.90 MBps is more than enough (translates to 720 Mbps) for compressed video. If you are recording RAW, you need CFast 2.0 or a proper SATA SSD (or, even NVMe PCIe drives)
Guspaz - Tuesday, January 24, 2017 - link
Professional video recorders typically use Apple ProRes or DNxHD (DNxHR for above 1080p) rather than h.264 or HEVC. They're both intra-frame-only codecs, meaning every frame is compressed independently, sort of like MJPEG, so they have massive bitrates.ProRes 422 HQ is 176 Mbps for 1080p24, so 4K24 would be 704 Mbps, or 88MB/s.
So, yeah, 90 MB/s is just barely enough for 4K, assuming you don't want to use more than ~24 FPS, and assuming you don't want to use ProRes 4444 or 4444 XQ. But at that point, you'll probably want to record to an SSD rather than SD.
lioncat55 - Monday, January 23, 2017 - link
It really depends on the bit rate and recording format. GoPros use around 60Mbit/s for their 4K recording while a Canon C300 recording 4k RAW can use over 200Mbyte/s.DanNeely - Monday, January 23, 2017 - link
Thanks. With that as context, and assuming Go Pro's 4k represents the approximate floor for mainstream 4k recording; its bitrate lines up with the bottom end of the tail of the 4k bar while leaving plenty of headroom for higher quality options above it. Recording RAW 4k blows the top; but recording RAW video's a pro feature while SDcards have always been aimed at the cheaper mainstream consumer market.Guspaz - Tuesday, January 24, 2017 - link
While it's true that most pro gear will record directly to SSDs (typically in at least a custom enclosure, if not an entirely custom SSD), lower-end pro recorders (like in the $500-1000 range) sometimes record to SD, as do some ENG cameras.SharpEars - Tuesday, January 24, 2017 - link
Whoa, a whole 90 MB/s, whoop-d-&#$^&-do!