Arlo|Smart Home Security|Wireless HD Security Cameras

Pro 5S 2K video is unreadable (large pixelation)

Reply
Discussion stats
  • 9 Replies
  • 868 Views
  • 0 Likes
  • 3 In Conversation
RSFWolfe
Aspirant
Aspirant

I have eight 5S 2K cameras.  Four of them occasionally fail to show video and instead display large “pixelated” color blotches covering the entire screen (see screen shot).  Sometimes this corrects itself, sometimes using the app to turn off and on the camera works, but most of the time I need to remove the battery and reboot the camera.  I have strong Wi-Fi signals at all the cameras.  My Wi-Fi speed is over 200Mbps (up and down).  I am using Ubiquity Wi-Fi switches and APs.  Any clues?

 

IMG_0147.PNG

9 REPLIES 9
GeneralMeow
Apprentice
Apprentice

welcome to the club.  I have the same problem.

 

When I reported this, I was told the usual "your wifi is too weak"... or "your internet bandwidth is too low", etc...  All of weren't true.  I have FTTH (Fiber to the Home) with gigabit.  I have a Cisco AP mesh.

 

I have to reboot my Pro5 cameras every week.  All the rest of my Pros to Pro4s and doorbells are fine.

 

There are numerous complaints about the Pro5s.

 

RSFWolfe
Aspirant
Aspirant

Well that’s a bummer. Thanks for you candid response. 

StephenB
Guru Guru
Guru

@RSFWolfe wrote:

I have eight 5S 2K cameras.  Four of them occasionally fail to show video and instead display large “pixelated” color blotches covering the entire screen (see screen shot).  

 


First, this is happening because there are errors in the encoded video stream.  While this can be the camera hardware failing, it is more commonly caused by packet loss on the connection path between the camera and the Arlo cloud.

 

Start by swapping one of the cameras with the problem with one that doesn't.  Then see if the problem moves with the camera (or not).

 

You can also measure the wifi speed at the camera locations using your phone.  Use the Ookla speedtest app, and turn off mobile data during the test.  Look for a consistent wifi speed that is above 2 mbps, and also see if the app is reporting any packet loss.

GeneralMeow
Apprentice
Apprentice

While technically it is correct that pixelation is caused by poor bandwidth and/or lost packets, and thus often used by level 1 technical support scripts, I think the point was missed where he said "...using the app to turn off and on the camera works, but most of the time I need to remove the battery and reboot the camera.".  Exactly the same problem I have, and I got the same response... even tho I beat the dead horse carcass that all my previous cams worked in the same spot just fine.  If the connection was bad all the time, a reboot or power cycle wouldn't help at all.  Not even including the already answered fact mentioned "...I have strong Wi-Fi signals at all the cameras. My Wi-Fi speed is over 200Mbps (up and down)".  Yes, we know, a strong signal doesn't mean packets aren't being lost.

 

At some point the stream will refresh the frame if using error correction.  The previous cams seemed to recover just fine after a second or two of lost packets.  The Pro5s almost never recover and will continue to have pixelated streams and recordings until reboot.  That sounds of software/hardware problems, not a weak WiFi signal, as it would recover at some point IF it was working fine all along.  It is becoming infuriating that the same excuse is being used in a support section with multiple threads and especially a single year long thread stating obvious issues with a specific product.

 

 

StephenB
Guru Guru
Guru

@GeneralMeow wrote:

While technically it is correct that pixelation is caused by poor bandwidth and/or lost packets, and thus often used by level 1 technical support scripts, I think the point was missed where he said ...


FWIW, my AVD1001 often shows a lot of pixelation. I don't suspect the wifi signal, since a nearby Ultra camera is not affected.  So I was not meaning to say the cause was invariably network loss.  I started by saying it could be faulty compression hardware.  That includes both actual hardware failing, and bugs that might be resolved with a camera restart. 

 

But loss is often (and in my experience) the most common cause, so it is something that needs to be ruled out.  Especially if it is happening on 4 out of 8 identical cameras.  That strongly hints that the location of the problem cameras is relevant.

 


@GeneralMeow wrote:

"...I have strong Wi-Fi signals at all the cameras. My Wi-Fi speed is over 200Mbps (up and down)".  Yes, we know, a strong signal doesn't mean packets aren't being lost.

 


A lot of folks who report this are telling us their ISP bandwidth measured from inside their home.  The bandwidth available at the camera location is often much less.  That is particularly the case with stucco construction (which is built on a metal mesh) or siding material that includes metal flashing.  Large appliances, chimneys, and plumbing stacks in the walls can also block signal to some locations.  So I always recommend testing signal quality at the camera location.  If that rules out the wifi connection then we can move on to other possibilities.

 

Interference is another factor - a user recently spent a lot of time and energy before finally sorting out that a faulty GPS tracker on their car was interfering with the wifi connection to the camera.

 


@GeneralMeow wrote:

 

At some point the stream will refresh the frame if using error correction.  

 


AFAIK, Arlo is not using forward erasure correction to protect the video stream.  It would be useful if they did (though it would raise the bandwidth requirements and could affect power use).

 

They do periodically send I-frames (about every two seconds), so the problem will clear fairly quickly if the next I-frame is successfully received.  However, since the I-frames are a lot bigger than the p-frames, the odds of loss affecting multiple I-frames in a row are unfortunately pretty good.

 


@GeneralMeow wrote:

It is becoming infuriating that the same excuse is being used in a support section with multiple threads and especially a single year long thread stating obvious issues with a specific product.

 


Well, it often is the cause, so it is not an excuse.  Several folks posting here who took the time to measure the wifi speed at the camera location have found that the performance was a lot worse than they were assuming.  

 

In any event, you will see the same troubleshooting steps over and over here, since they are needed to try to pin down what problem that the particular poster has.  Although those steps are well-known to people who frequent the forum, they are often not known by the people posting the problem.

RSFWolfe
Aspirant
Aspirant

Thanks for all the helpful responses.  I will try them and see if any of them helps.

GeneralMeow
Apprentice
Apprentice

@StephenB  I don't disagree with anything you said.

 

My point is coming from the documented issues with respective product.

 

You are correct, I dont know if Arlo is using error correction.  I am assuming based on the difference from the old camera pixelation vs the Pro5 attributes.  I deal with high-end network gear with bandwidth over 40G where FEC is important and I inserted that in the streaming protocol and how it recovers after a few seconds.  The older cameras seem solid.

 

StephenB
Guru Guru
Guru

@GeneralMeow wrote:

I deal with high-end network gear with bandwidth over 40G where FEC is important and I inserted that in the streaming protocol and how it recovers after a few seconds.  

 


FWIW, I've implemented FEC protection for video streams (obviously not at Arlo).  While I can't see what is in their video packets due to the encryption, I have looked in detail at their recordings.  They definitely are sending I frames at about 2 second intervals. Unlike P frames, these code the full scene from scratch (not building on previously sent frames), so if they are successfully received they will clear up the video.

 

If Arlo added repair packets to their stream, they could reduce the frequency of I-frames.  That would likely give better results than what they are doing now.  The downside is that generating the repair packets in software would add some latency (which is already too high), and would increase the computational load (and power) on the camera SoC.

 


@GeneralMeow wrote:

The older cameras seem solid.


They need less bandwidth, since they aren't sending 2K.  The older ones are also single-band.  5 Ghz doesn't reach as far, and in some locations the Pro 5s camera might be frequently switching bands.  That would result in connectivity issues, and increase power drain.  FWIW, I wish there were a way to disable one of the bands in the dual-band models.

 

But it is quite possible that some (or even all) Pro 5s cameras have weaker WiFi than other Arlo cameras.  I do see some pixelation on my Pro 5s - not enough to interfere with how I use it, but a bit more than I see with an Ultra that isn't too far away.  At some point I'll probably swap the cameras to see if the problem moves or not.

GeneralMeow
Apprentice
Apprentice

@StephenB wrote:

  FWIW, I wish there were a way to disable one of the bands in the dual-band models.

This I agree with.

 

I did try at one point to change the name of the 5G band to see how the camera would resolve itself to 2.4G, but it stayed offline for 10+ mins.  That was a while ago and prob with a buggy code rev, so I dont know if it recovers better now.