I have been getting slower playback from my directly plugged in SSD drive than from my NFS share. I’ve been using the Jellyfish test files to compare playback speed. NFS - almost smooth playback on the 300mbps test file, on the SSD, very stuttery.
My external SSD drive is this one, it is not powered, however, should that really mean it is in effect slower than an NFS link?
The log: https://paste.osmc.tv/vopafuyice
Quite possible. It’s easy to test it you have a powered USB hub to test with.
Interesting. I’ve ordered a hub from you guys on 14 december, hopefully it will arrive soon.
Did you measure throughput via dd?
Also why to test with unrealistic files that are not available in “real” world encodings?
How is the SSD formatted?
If it’s NTFS this would explain why you don’t get the best performance
Cheers
Sam
I test with these to benchmark the speed, simple as that. How do i test with ‘dd’?
It’s ext4 (as was suggested in this topic) plus it now has the usb hub from your store, but it doesn’t make any difference…still the exact same stuttering.
By the way, with the hub, is it normal that the power adapter plug doesn’t fully plug in (so you don’t see any metal from the plug anymore)? (photo)
Thanks for confirming disk is formatted as ext4
I see you are having problems with Jellyfish clips. Some are not possible to play back smoothly on the device. Can you confirm which version you are using?
The hub looks Ok
That should play if your drive is fast enough.
Did you try playing from the eMMC as a test?
Sam
Just checked it, it plays at about the same speed as the one from the nfs source.
Tested this on a Vero 4k and a Full HD Plasma which is limited to 100 Mbit/sec but has an external USB gigabit ethernet adapter which makes up to 320 Mbit/sec bandwidth available:
From my point of view this mkv requires a bandwidth beyond possible troughput of USB 2.0 on the Veros.
Interesting. Is there a way to test the speeds, or get some numbers? It seems weird that an nfs share on the other side of my house is way faster than a local, USB 3.0, recent model external ssd, right?
Ethernet will give you 930Mbps (real world)
USB 2.0 will give you about 280Mbps (real world).
Buffering is also handled differently depending on source
Fortunately the 4K UHD spec maxes out at 128Mbps; which means USB 2.0 is more than adequate for 4K playback
You could use iperf or smartmontools to do some benchmarking
Wait, I have a USB 3.0 SSD, and the USB 3.0 hub from your store, so that means I should have USB 3.0 speeds right?
No, all actual Veros have USB 2.0.
Oh wow, can’t believe I overlooked that. I guess I took it for granted that a new mediaplayer would have USB 3.0.
Not sure why I even bothered with buying the USB 3.0 hub Tom Doyle recommended though, as in retrospect it was kind of obvious that USB 2.0 would be, and always was the bottleneck.
Just think about that and the sense of artificial test files.
It was never the issue that I couldn’t play standard 4k files. It wasn’t about that at all.
It was about the difference in speed from a local ssd drive and an nfs share on the other side of the house that was noticeably faster which made me think something should be wrong here.