USB corruption for dashcam recording

[Knightshade]

So now there are 3 cameras being recorded. That would imply 50% more data per second (of course, depends on sampling rate and compression method) so perhaps the memory cards need to be 50% faster.

.... what?

There's been 3 cameras for quite a while.

They just added a 4th, which would be 25% more, not 50%.

Anyone measure/analyze the data amount per second being written on a AP2.5 system under V10?

The cameras are writing 0.5 megabytes per second per camera... so it went from 1.5 to 2.0 per second with this 4th camera.

Which is still several times slower than even cheap USB 2.0 keys can usually manage for sustained writes, let alone any newer/decent ones.

The fact folks are getting the "too slow" error even on USB devices that benchmark 10-20 times faster than that for writes is, yet even more, evidence this is a software problem, not a hardware problem.

 

[BioSehnsucht]

There may be practically zero buffering of the writes, in which case write latency is important. If they had nice big write buffers with lazy writing then it would be much more forgiving of bursty behavior.

 

[CSFTN]

.... what?

There's been 3 cameras for quite a while.

They just added a 4th, which would be 25% more, not 50%.



The cameras are writing 0.5 megabytes per second per camera... so it went from 1.5 to 2.0 per second with this 4th camera.

Which is still several times slower than even cheap USB 2.0 keys can usually manage for sustained writes, let alone any newer/decent ones.

The fact folks are getting the "too slow" error even on USB devices that benchmark 10-20 times faster than that for writes is, yet even more, evidence this is a software problem, not a hardware problem.

My error on the 3 vs 4 .... but then going from 3 to 4 is a 33% increase not 25%. But thank you for the answer. Looks like write speed isn't the issue.