Raw File Transfer Strategy

Why Come?

Why Come?

"If you had the SSDs in separate readers connected to separate busses you could get almost twice that speed"

As we say around here, "why come?" If the OWC TBolt reader is TBolt2, and is connected to a Tbolt2 bus on my Mac laptop, why would moving to a second card reader on my second TBolt2 port make it faster? The read/write speed of the discs are the limiting factor, not the 20GS TBolt2 connection, no?
 
Last edited:
Before you buy anything, I recommend testing your current setup and finding the bottlenecks. Try running a disk speed benchmark on each drive to make sure they each are properly utilizing the bus correctly as sometimes poorly designed controllers or malfunctioning buses can drastically cut your speed. If your drives independently benchmark at a higher speed than you get when copying to each other you may be dealing with a bottleneck on the bus.

Frankly, though, it's hard to really identify much of a benefit in uncompressed raw that justifies the 3+ times higher data loads, especially since the uncompressed raw originates from data that was compressed 3.6x to fit over SDI. And then you have to edit it. It's not an inexpensive proposition to have SSDs or big RAIDs for every step of the process.
 
nyvz said:
Frankly, though, it's hard to really identify much of a benefit in uncompressed raw that justifies the 3+ times higher data loads, especially since the uncompressed raw originates from data that was compressed 3.6x to fit over SDI. And then you have to edit it. It's not an inexpensive proposition to have SSDs or big RAIDs for every step of the process.
Click to expand...

Yes, I agree that for the majority of shooting Sony FS700 Raw to ProRes 422 results in pretty damn good images that grade fairly well under controlled lighting situations. I do see an advantage in certain wide range lighting situations, such as out door sunlit shoots with rapidly changing light levels and wide dynamic range requiring secondary correction in the grade. IMHO those kind of tough lighting situations can be better dealt with using RAW footage that can be more extensively corrected in post.
 
ric8138 said:
"If you had the SSDs in separate readers connected to separate busses you could get almost twice that speed"

As we say around here, "why come?" If the OWC TBolt reader is TBolt2, and is connected to a Tbolt2 bus on my Mac laptop, why would moving to a second card reader on my second TBolt2 port make it faster? The read/write speed of the discs are the limiting factor, not the 20GS TBolt2 connection, no?
Click to expand...

SATA controller in your OWC Thunderbolt 2 Dock appears to be a single SATA controller with port multiplication shared between the two docking ports instead of two separate SATA controllers. My guess is that's your bottleneck. So your total accumulative bandwidth for both read and write is around 500 with that dock. At that price point and with TB2 in the other end I think it's kind of silly of OWC to use a single SATA controller. Just my 2 cents.
 
Parrot said:
At that price point and with TB2 in the other end I think it's kind of silly of OWC to use a single SATA controller. Just my 2 cents.
Click to expand...
Agreed, and that is disappointing, the OWC dock seemed like a perfect on set solution. I don't see any TBolt - SATA card readers on the market. Would a single dedicated USB3 - SATA connector do the trick?
 
ric8138 said:
Agreed, and that is disappointing, the OWC dock seemed like a perfect on set solution. I don't see any TBolt - SATA card readers on the market. Would a single dedicated USB3 - SATA connector do the trick?
Click to expand...

Yes, one SSD on a $20 USB 3 to SATA dongle (that can take advantage of the full USB 3 speed!!! Don't get the slower ones!!!!) and the other SSD in your OWC dock should help significantly. You might need a Y cable for more bus power with the Samsung 850 EVOs.

Or you could try something like this:
https://www.akitio.com/adapters/thunder-sata-go
 
If you use the model adapter we currently sell then there is no need for additional power. Different models have varying power efficiency, which is why we have gone through a number of them until we found this one which we find highly efficient.
 
Hey Mitch, thanks for replying but,,, which one? And a search on B&H for CD-SSD-USB3 as it's described on your website returns nothing. The picture on the CD website doesn't link to anything. Kinda frustrating.
 
Last edited:
Mitch, thanks.
All, I have an Atmos USB3 card reader that I tried that in both ports on my MacBook Pro as a reader, and the Samsung in the TBolt port as the writer and only got 220 MBS.

From System report:
USB 3.0 Bus:


Capacity: 256.06 GB (256,060,514,304 bytes)
Removable Media: Yes
BSD Name: disk3
Host Controller Driver: AppleUSBXHCILPTH
Partition Map Type: MBR (Master Boot Record)
PCI Device ID: 0x8c31
PCI Revision ID: 0x0005
PCI Vendor ID: 0x8086
Volumes:
ODYSSEY4___:
Capacity: 256.06 GB (256,058,435,584 bytes)
Available: 13.86 GB (13,860,487,168 bytes)
Writable: Yes
File System: Journaled HFS+
BSD Name: disk3s1
Mount Point: /Volumes/ODYSSEY4___
Content: Apple_HFS
Volume UUID: 92A38ADB-C583-3CFD-BB7C-C2945EA8BDDD


USB to ATA/ATAPI Bridge:


Product ID: 0x0539
Vendor ID: 0x152d (JMicron Technology Corp.)
Version: 13.02
Serial Number: 012345678933
Speed: Up to 5 Gb/sec
Manufacturer: JMicron
Location ID: 0x14600000 / 11
Current Available (mA): 1800
Current Required (mA): 2
Extra Operating Current (mA): 0
Capacity: 256.06 GB (256,060,514,304 bytes)
Removable Media: Yes
Detachable Drive: Yes
BSD Name: disk3
Partition Map Type: MBR (Master Boot Record)
Volumes:
ODYSSEY4___:
Capacity: 256.06 GB (256,058,435,584 bytes)
Available: 13.86 GB (13,860,487,168 bytes)
Writable: Yes
File System: Journaled HFS+
BSD Name: disk3s1
Mount Point: /Volumes/ODYSSEY4___
Content: Apple_HFS
Volume UUID: 92A38ADB-C583-3CFD-BB7C-C2945EA8BDDD


Apple Internal Keyboard / Trackpad:


Product ID: 0x0262
Vendor ID: 0x05ac (Apple Inc.)
Version: 2.25
Speed: Up to 12 Mb/sec
Manufacturer: Apple Inc.
Location ID: 0x14400000 / 3
Current Available (mA): 1000
Current Required (mA): 40
Extra Operating Current (mA): 0
Built-In: Yes


BRCM20702 Hub:


Product ID: 0x4500
Vendor ID: 0x0a5c (Broadcom Corp.)
Version: 1.00
Speed: Up to 12 Mb/sec
Manufacturer: Apple Inc.
Location ID: 0x14300000 / 2
Current Available (mA): 1000
Current Required (mA): 94
Extra Operating Current (mA): 0
Built-In: Yes


Bluetooth USB Host Controller:


Product ID: 0x8289
Vendor ID: 0x05ac (Apple Inc.)
Version: 1.18
Speed: Up to 12 Mb/sec
Manufacturer: Apple Inc.
Location ID: 0x14330000 / 7
Current Available (mA): 1000
Current Required (mA): 0
Extra Operating Current (mA): 0
Built-In: Yes

 
Apparently there is some difference between the Atmos card reader and the CD approved, Startech USB3 to SATA cable. I plugged in the StarTech that arrived today and immediately got 500MBS average in a transfer to the Samsung EVO 850 in the OWC Tbolt dock. That appears to be about the best I can expect without going to an SSD RAID 0 setup. It works for me.
 
ric8138 said:
Apparently there is some difference between the Atmos card reader and the CD approved, Startech USB3 to SATA cable. I plugged in the StarTech that arrived today and immediately got 500MBS average in a transfer to the Samsung EVO 850 in the OWC Tbolt dock. That appears to be about the best I can expect without going to an SSD RAID 0 setup. It works for me.
Click to expand...
That sounds about right. You're not likely to get any faster than that, but it's a pretty huge improvement over what you were getting previously. There's a reason we went with that adapter. These things may look like simple cables but there's actually a circuit inside and some have greater efficiency than others.
 
nyvz said:
Before you buy anything, I recommend testing your current setup and finding the bottlenecks. Try running a disk speed benchmark on each drive to make sure they each are properly utilizing the bus correctly as sometimes poorly designed controllers or malfunctioning buses can drastically cut your speed. If your drives independently benchmark at a higher speed than you get when copying to each other you may be dealing with a bottleneck on the bus.

Frankly, though, it's hard to really identify much of a benefit in uncompressed raw that justifies the 3+ times higher data loads, especially since the uncompressed raw originates from data that was compressed 3.6x to fit over SDI. And then you have to edit it. It's not an inexpensive proposition to have SSDs or big RAIDs for every step of the process.
Click to expand...

Have you found better performance for Green Screen work in RAW over the ProRes422? Or is the Raw still too compressed for it to make a difference?
 
Raw on the Convergent Design 7Q is totally uncompressed, which is why I'm searching for a faster transfer strategy. 4K Raw onto a 256G card = 10-12 minutes of record time. I have not attempted green screen with RAW yet, but judging by the impressive flexibility I would expect it to do extremely well as long as it was well exposed.
I HAVE used the ProRes footage from the 7Q in some green/blue screen work and it is impressive, head and shoulders above the internal AVCHD. Again, you must keep the screen well up (above 30 on the WFM) to keep the noise out, but when I do, the keys are nice and clean with very smooth, well defined edges, especially the 4K
 
You must log in or register to reply here.
Back
Top