What's new

Puzzling speeds with WiFi, DLNA and a smart TV

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

TimTheToolMan

New Around Here
Hi All. So I've been lurking around these forums for 3 weeks or so reading as I slowly set up my new quarantine zone aka home network, and I've finally run into a problem for which the search function hasn't availed me. So figured I should put my hat in the ring, see if anyone might be able to offer some wisdom.

The short version:

My Tv's inbuilt 5Ghz AC wifi for some reason seems to give far lower throughput than my other AC devices, even when connected in the same way, doing the same tasks.

While the router has shown itself capable of happily pushing ~200Mpbs, streaming DLNA media between computers on wifi alone, the TV will only manage 70Mpbs despite being on the same 5Ghz network, manually locked to AC at 80Mhz bandwidth. I would conclude the TV's card just can't process data any faster, BUT when I connect the *computer* to the router using gigabit cable, suddenly the same TV will accept 130Mpbs by wifi ( or even higher, but only if I pause playback).



The long version:

Hardware:
-New smart TV (Hisense 50R7) client. Plays ~80GB 4k HDR files from USB fine.
-Dell XPS 9550 laptop, Ubuntu 1604 client
-Dell XPS 9360 Win10 client
-Old Asus RT-AC66U (_not_ the B1, the 600Mhz original) running Merlins last official firmware (380.70). Mode set on "N/AC", with an 80Mhz band. Hardware acceleration showing as enabled. No other 5ghz nearby. Everything in the same room



What's happening is:

Withough the TV, i.e.

Laptop -> 5Ghz Wifi -> AC66U -> 5Ghz Wifi -> Laptop

I get 170-200 Mpbs thruput, measured with iperf3. Checks out on the media, too - 60gb remuxes play smooth on either laptop, streamed by DLNA from the other. If I put an Ethernet cable on one of the links, the rate boosts to 4-500Mbps, which makes sense due to no router re-transmission. All fine.

BUT,

If I connect as

Laptop -> Gigabit Ethernet via USB C adapter -> AC66U -> 5Ghz Wifi -> TV

I suddenly only get about 120 Mbps of throughput (measured using laptop connection monitor which matches Ipef3). Enough for some movies, not enough for many more. Lotsa buffering. Confusing, as I know the router can do more as above.



Even stranger, if I go

Laptop -> 5Ghz Wifi -> AC66U -> 5Ghz Wifi -> TV

I then only get about 60-70Mpbs. Again, I KNOW the router can serve over double this, even on re-transmit duty. And I know the TV can handle AT LEAST ~120Mpbs...but why only if I use a cable from Laptop -> Router?


Whats odder still, is that in either scenario if I pause the movie, the rate INCREASES by about 70Mpbs! So if the laptop is hooked up with cable to the router (but TV still by wifi) , it surges up to about 200Mpbs, but if it's all on wifi, it jumps to about 140Mpbs...but only for about 30s, then it stops entirely. Either way, I only get get maybe 30s of clean palyback, then back to buffertown

My original theory was that the TV is somehow forcing a lower standard of connection, maybe dropping to 'n' and capping throughput accordingly . But I've locked the settings in the 80Mhz band with no change. Moreover, even with TV on the network, the laptop-laptop speed remain high. So as far as I can tell, the TV should be able to accept speeds just as fast.


I've cross checked this now using diffeent channels, trying both laptops as DLNA servers, other MerlinWRT versions and stock firmware, and even an entierly new router (a newer Sagemcom 4353 bundle-DSL type unit) with similar results. The TV seems to be the problem, I just don't know why.

What have I missed, folks?

Thanks!
 
I can't say this surprises me at all. In my experience with Sony and Hisense TVs they just don't have the processing power to handle that volume of throughput via DLNA. YMMV of course.

Can you plug the TV directly into the router via Ethernet? That could help to identify whether or not this a WiFi related problem.
 
I can't say this surprises me at all. In my experience with Sony and Hisense TVs they just don't have the processing power to handle that volume of throughput via DLNA. YMMV of course.

Can you plug the TV directly into the router via Ethernet? That could help to identify whether or not this a WiFi related problem.


Can, and have! Its a 10/100 port (at least thats what the router reports).

I originally thought the same but like I said above - if it has the processing power to accept it when I have the laptop plugged into the router by ethernet (which it demonstrably does, I've clocked throughput at 130Mbps in that config), then why does it slow down when the laptop connect by wifi? The Laptop - Router connection surely has no impact on the processing power of the TV, so why the slowdown?
 
What you haven't covered are the wireless specs of the TV itself. Can the TV handle/consume an 80MHz channel width? I have clients that get finicky when I enable 80MHz, so I stick to 40MHz usually.

What is the wireless performance from the TV for other things other than DLNA? Is the TV just in a bad place to consume WiFi efficiently? Have you confirmed the link speed of the TV to the router?
 
Side note: One thing to bear in mind when interpreting "throughput" numbers for DLNA and other streaming formats is that they are always misleading (to a lesser or greater extent).

The client will get (and buffer) "chunks" of the stream at a time. When the buffer is sufficiently low it will get another chunk from the server. So the data isn't being transferred in a continuous manner. So in reality you should be seeing a short burst of data followed by a gap. As the demand for data approaches that of the server/network to provide it the size of the gap diminishes. Once the gap disappears you start to get buffering.

So the sampling rate of whatever you're using to monitor the throughput effects the accuracy of what you're seeing.
 
What you haven't covered are the wireless specs of the TV itself. Can the TV handle/consume an 80MHz channel width? I have clients that get finicky when I enable 80MHz, so I stick to 40MHz usually.

What is the wireless performance from the TV for other things other than DLNA? Is the TV just in a bad place to consume WiFi efficiently? Have you confirmed the link speed of the TV to the router?


I assume so? Merlin WRT reports a 702-866Mbps Tx and Rx link rate in the Client list, at idle. This varies under load - sometime showing breifly as low at 6.5Mbps , but that seems to be that case normally with all connections, including the laptop connections which are as fast as expected. The router is locked to the 80Mhz setting and the TV wifi connects and holds quite reliably for lower-bitrate streaming. If if use 40Mhz, the performance worsens.

Thats the most I have to go off here - Hisense doesn't publish any detailed specs of the card itself. I havent got anything else that comes close to the 4k files provoking the issue, Youtube/Netflix content on the built-in apps is all very low quality/bitrate by comparison. I'd love to run an iPerf3 server on it and get a real benchmark but I've no idea how.
 
Last edited:
Side note: One thing to bear in mind when interpreting "throughput" numbers for DLNA and other streaming formats is that they are always misleading (to a lesser or greater extent).

The client will get (and buffer) "chunks" of the stream at a time. When the buffer is sufficiently low it will get another chunk from the server. So the data isn't being transferred in a continuous manner. So in reality you should be seeing a short burst of data followed by a gap. As the demand for data approaches that of the server/network to provide it the size of the gap diminishes. Once the gap disappears you start to get buffering.

So the sampling rate of whatever you're using to monitor the throughput effects the accuracy of what you're seeing.

Yes, I understand this mechanism. What you describe is the exactly the behavior I see on the bandwidth monitor (sampling every 0.5s) when playing DLNA streams by wifi between laptops - 'chunks' of data sent in ~200Mbps bursts. The higher the bitrate, the shorter the pauses, until with movies around the 80gb + filesize, it's just a continuous stream of 200Mpbs transmission. As to be expected - cache being played as fast as it fills.

It's also what I see playing far lower bitrate files on the TV - only the chunks are sent in much slower ~ 70Mbps bursts, and in this case, any movie larger than 15gb results in a constant stream and buffering. If I switch to an ethernet link from laptop to router, then the 'chunks' are sent at 130Mbps, and I dont seem the same beahviour until around 25GB-sized movies. Sadly most of my rips are done personally off BluRay discs so theyre easily 80gb.

Interestingly, as I said in the OP, if I pause the playback, the Rx speeds up as it fills the cache, by about 70Mbps for around 30s, then drops to 0Mbps once the cache fills. What doesn't make sense is why the Rx rate only increases when the playback is paused. Which brings us back to the core question:

The TV is demonstrably capable of receiving data by WiFi at 130Mpbs while playing media (and more when paused), when the source laptop is hooked to the router by cable. The router is proven capable of streaming DLNA at 130Mbps (and more), over wifi-wifi alone. So if the router can provide it, and the TV can handle it, why then, does the TV limit to 70Mbps during playback when the source laptop connects by WiFi?

This inconsistency to me says something is amiss. I just dont know what
 
Bumpity Bump.

Nearly a week and I'm still unable to make sense of this behaviour.

Is anyone able to shed any light on why the TV would choose to receive at 70Mbps on 5G Wifi, when it is able to do at least 130Mpbs, when the source is also connected to WiFi and capable of 300Mbps +?
 
It really sounds like your WiFi in general isn't clean....at least when the TV is involved. I see you posted multiple use cases with different clients and connections, but the real slow use case is two WiFi connected devices with one of them being the TV. I am still wondering if you have a signal quality challenge at the TV. If the laptop is wired, it isn't so bad. But when the laptop is wireless, performance drops. That sounds like signal issues to me. Especially since you report that the link drops to 6Mbps at times....no quality link should be dropping that low. That to me really points to signal quality or interference issues.

Now, why does the laptop to laptop test not suffer the same? No idea, but guessing different RF characteristics maybe? Is the 2nd laptop in the exact same location as the TV? Is it further or closer to the AP?

When you know you have a 120Mbps'ish limit for the TV on WiFi, then you go and add another WiFi device into the conversation, it isn't out of the realm of possibility to think that your 120Mbps'ish limit gets cut in half when it is talking to another WiFi device on the same channel. If you have this kind of high bandwidth requirements on a streaming type setup, WiFi isn't going to be your friend here.

Just re-read your notes about the TV bursting higher when playback is paused. I really do wonder if you are hitting CPU limits of the TV? When paused, the CPU isn't busy rendering so it can transfer. When playing, the CPU is busy processing that data and can't accept data? No idea if that is how it works...this is pure guessing here.
 
THanks for the reply! Good thinking material there.

It really sounds like your WiFi in general isn't clean....{snip}

That sounds like signal issues to me. Especially since you report that the link drops to 6Mbps at times....no quality link should be dropping that low. That to me really points to signal quality or interference issues.


Hmmm! Honestly I dont know. I just assumed because everything was so close, signal must be good. Test layout it literally everything in one room - laptops on onw side of me on the sofa, router/s on the other side on a coffee table, TV 10ft in front. Other wifi in the house on or off makes no difference, nor does channel. And again - in the same circumstances, laptop-laptop wifi is blitzin. To me, that suggests it's not an environmental/signal. But I dont know

What should my layout be to minimize self-inflicted interference?



As for the reported link speed drops, they are only momentary, and never during data transfer. I admit I havent watched closely to map an exact pattern (I'll do this later, now!) but the usual behavior is when the connection is idle, it reports 800+, with occasional blipls down to single digits, then when streaming it holds at some constant value, broadly reflective of the streaming speed (ie link a bit higgh than actual transmission).

Youve got me thinking - I'll do a test and watch how link speed changes with different device transmission.


Now, why does the laptop to laptop test not suffer the same? No idea, but guessing different RF characteristics maybe? Is the 2nd laptop in the exact same location as the TV? Is it further or closer to the AP?

When you know you have a 120Mbps'ish limit for the TV on WiFi, then you go and add another WiFi device into the conversation, it isn't out of the realm of possibility to think that your 120Mbps'ish limit gets cut in half when it is talking to another WiFi device on the same channel.

Yeah, the laptop-laptop behavious is the real spanner in the proverbial works. Lesser rates with double wifi legs makes total sense...on its own. But that would normally be a case of router/signal being the limiting factor, no? IN which case, for the laptop to be so much faster contradicts that. Very perplexing


Just re-read your notes about the TV bursting higher when playback is paused. I really do wonder if you are hitting CPU limits of the TV? When paused, the CPU isn't busy rendering so it can transfer. When playing, the CPU is busy processing that data and can't accept data? No idea if that is how it works...this is pure guessing here.

This was (is?) my secondary theory. But it doesn't explain the disparity between speed when laptop is on wifi vs ethernet. Ie if the CPU can keep up with continuous playback @ 130Mbps (which is can with laptop on eth) then why can't the CPU keep up wth both on WiFi?

My head hurts.
 

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top