I'm curious if any other Twitch broadcasters are experiencing this issue with Rogers. I get 250 down / 20 up consistently and my upload to Twitch will start to drop frames/packets if I exceed around 3000 bps, which is not even a 1/4 of my upload.
I've been troubleshooting the connection issues streaming to twitch, for almost a year now with support. Modems have been replaced. Techs have come and replaced wire and tested to the curb etc. Speedtests are fine. The issue persists across multiple machines/cables and every Twitch ingest server that I can reasonably test location wise.
At some point, someone suggested I encapsulate my data by testing with a VPN to the local cities of the ingest servers I was testing. This is when I realized, that Rogers may in fact be throttling my RTMP uploads by payload. When I use a VPN, and test to any of the ingest servers I had issues with, I get no packet loss or dropped frames to Twitch. I can push the Twitch cap of 6000 bps without issue with the VPN enabled. ( testing to Toronto, NY, Chicago ingests primarily)
I did another test. I streamed my feed to restream.io ( which also uses RTMP) to rule out a rule specific to twitch completely. I saw the exact same issue. Without the VPN, I would get frame/packet loss, but with the VPN on, I would have no issue.
I've been running with the VPN for some time now as a workaround, but it's an additional cost of course, and one I shouldn't need to incur. Ater a few months of solid streaming through the VPN, I tested again without it. As expected I immediately started to lose frames/packets intermittently through the broadcast, due to insufficient network when I disabled the VPN.
I'm at the point where I think I'm going to need to change providers. Calling front line support doesn't seem to get me anywhere. They just want to replace a modem or close my ticket the next day without any feedback/consent or resolution.
Is anyone else out there experiencing these types of issues uploading to twitch via rogers? Advice on how I can escalate this to an appropriate resource within the support team who will look at it seriously?
150u Ignite user here w/ CGNM-3552 bridged with an Archer C7 router. I have been pretty satisfied with the promised speeds and consistency in everything but Twitch.
Like everyone else here, I am getting drops in bitrate to Twitch. Dropped frames tend to hover around 15% @ 5000kb/s. I should note that Twitch streams tend to buffer fairly often.
Using NordVPN did not affect the results.
I tried streaming to other platforms today. YouTube has a 24-hour "preparation" period, but I was able to find success with Facebook Live. I was able to stream a consistent 5000 bitrate with 0.2% dropped frames over a 24 minute period (small time frame, I know). The URL Facebook provides has RTMP in it, so I'm assuming it is a RTMP stream.
For anyone interested in testing via Facebook, there's a sort of preview page; nothing will show up on your feed until you actually press Go Live.
I'll post my results with YouTube as soon as I get the chance to.
If this isn't fixed soon, I'm switching to Bell (whether or not this is an intentional throttling issue, routing issue, or poor choice of modem). They put up fiber lines in my neighbourhood recently and have been aggressively calling and doing door-to-door. Their prices are usually only good for a year though.
I just wanted to let you know about my findings.
I just tested the bitrate to YouTube. From my findings, both Facebook and YouTube livestreams use RTMP and they work spectacularly (<1% dropped frames for Facebook, 0.1% for YouTube).
I'm fairly certain that Rogers is throttling Twitch traffic, but there's no way to be completely sure. I understand that everyone's set-up and circumstances are different, and that this does not rule out any Puma6 related issues.
If you would like to see a video, I can record the OBS client with a few minutes on each platform (Twitch, YouTube, Facebook). Although the sample size will be small, you can immediately tell the difference.
I am currently on the 150 ignite plan, with a CGNM 3552.
I am mainly experiencing 2 problems:
1. Latency spikes in TCP-based applications (namely, Blade and Soul), which I believe to be a Puma 6 issue
2. Twitch streaming (both inconsistent upload bitrate, and download buffering), which I believe to be throttling, or a routing issue, as Facebook and YouTube streaming is impeccable and without problems
I am seeking advice on how to remedy these issues, I have tried every troubleshooting step reasonably possible, and don't see any other things to try other than installing a trial firmware, or swapping modems.
As far as I know, the CGNM 3552 is incompatible with the 184.108.40.206 firmware that appeared to fix a lot of latency issues for others (although in my case, the changelogs clearly state that the improvements in earlier revisions affect ICMP and NOT TCP). If anyone has any suggestions on which firmware trials to test, I'm all ears.
In regards to switching modems to the non Puma 6 modem (CODA 4582), I gave technical support a call, and they said it's possible to get the CODA on the 150u plan, but that would depend on the store. I called my local Rogers Plus store and they said that the only way I can get the CODA is by overriding the system via the loyalty department. I give the loyalty department a call and they said that the only way I would get the CODA is by paying a $50 service fee so that a technician can come and install the modem (as a way to "bypass" the system).
Technical support concluded that the steps I have taken to troubleshoot my problems were sufficient and offered to send a technician to check my home of "dead-spots" (despite my repeated explanations that I am on ethernet). Additionally, they said that I would be charged $50 if they were to find any problems with splitters and the like. To my knowledge I don't have any of those cable splitters installed, but I chose to postpone the visit because this is likely a modem issue, and a throttling/routing issue. At the end of the day, the technician will do their business, and tell me that everything is fine on Rogers end and to contact Twitch. That, and I don't want to be baited into random fees and service charges, given that most of us are stuck with defective modems and all.
So yeah, I'm really at my wit's end here. I apologize if this sounds like a rant, I don't know what else to do. If anyone has any suggestions on what else to try, I would greatly appreciate it.
I am not going to go through the trouble of uploading my 'proof' but my problems are pretty much the same as many others here. Rogers has tried very hard to keep me happy. But their efforts have been limited to making sure my connection at home is fine (which it now is).
This whole Twitch nonsense is not worth their time.
If Bell was more than 50/10 in the L6T area I'd be gone.
Oh I also read a post earlier in this thread where Youtube streaming is fine (streaming to YT instead of Twitch).
I was like well let's just give it a shot because I have nothing but free time.
I set it to 12,500 Kbits/s (more than double of what Rogers is incapable of doing on Twitch). Worked well for 3 hours yesterday 0.01% Packet Loss which is very decent.
Tried it again later. And now streaming to YT is almost as bad as Twitch.
DSL reports rates my connection as over 800Mbits down and over 25Mbits up, and an A or A+ rating.
But it would appear my third world internet doesn't want believe in live streaming of games.
What is even more interesting is if I use a bit rate that Rogers doesn't like. Say anything over 3000Kbits/s it will fluctuate wildly and often even "time out" (OBS will report an upload of 0/ Kbits/s). Now if I use third world quality, of 1,500Kbits (significantly less than what Rogers is billing me for). It never goes down to 0.
Not sure how it all relates. But it sure looks like I'm not getting what I'm paying for. And after yesterday it actually looks like Rogers is actively out to ruin my experience.
@scratchieepants, your comments really point out the perplexing issue, streaming at 12,500 kb/s, and then being unable to stream later on. Yesterday was Sunday, which can see a much lower or higher load at the CMTS depending on what time you were streaming. It makes me wonder if some of these issues are caused by the CMTS load? Level II Tech Support would have access to the CMTS load data.
Out of curiosity, can you load an application called Core Temp, which shows the CPU core temps and the CPU core load in %. I'm curious as to what the difference in CPU core load happens to be between something like a 1500 kb/s stream versus a 12500 kb/s stream. Can you run Core Temp when you run a stream and let me know what the numbers look like. Core Temp can be found here:
Are you a twitch partner? I understand that allows you to run higher upload rates.
Are you running the 4582 modem in Gateway mode with a direct connection to the modem or in Bridge mode with a router behind it?
One question that needs answering is whether or not Rogers peer directly with Twitch and if so, what the throughput rate, or, does Rogers peer with Torix, who then peers with Twitch. That might be part of the problem as well.
I sound immature and angry. That's not necessary.
But it's been a source of frustration for weeks now.
To recap Rogers' people have been excellent every step of the way, I am only commenting specifically on what I am experiencing.
I don't have CoreTemp, but I am sure you've heard of HWiNFO 64, it's pretty good.
at 12,500 streaming, CPU maxes out at 82C, but usually sits in the low 70's. (It's a Dell small form factor PC dedicated to streaming, i7-4790 (non K). CPU utilization stays below 50%.
At 6000, about 10C less on all fronts.
I'm not partnered on Twitch, but now 'everyone' can stream at 6000 kb/s if their connection can support it.
And with regards to that poor streaming performance and time of day. I even messed with it this morning before work. 845 AM in Brampton, Ontario... should be fairly decent?
Also with the Twitch bandwidth test, I can sometimes score well with the default short (10 second) test. But the longer version (30 seconds), not a single server will score over 10 (you need 80-100 consistently for it be a good connection). Like I'll get the 10,000+ bandwidth rating to QC and Toronto's Twitch ingest... but a rating of 0.
I absolutely understand the frustration. I'm not a twitch user or uploader, but, these complaints have been posted in the forum for months now and there doesn't seem to be any impetus to resolve the problem. I'd like to do more to help, but, as a site volunteer, REs don't have any system access or the knowledge to use it, even if we had system access.
Can you try one thing when you have time. Run a short test stream to youtube at a high rate, maybe at 12,500 kb/s if that works, then turn around and run a short test stream to twitch. I'd like to see what happens at 3kb/s and 6 kb/s. Then turn back around and run another stream to youtube, just to confirm that you're still in a time window where high rate streams will run. Its good to know that streaming to youtube can work, I'm just wondering what you will see with near simultaneous streams. Is twitch always a problem is the question that I'm wondering?
From what I understand, both twitch and youtube use RTMP for uploading. The fact that streaming to youtube can work very well blows my theory of modem protocol incompatibility out of the water. So, this points to some issue at the CMTS or beyond. It may be a path issue from your modem to the ingress server. Ergo the question on peering between Rogers and Twitch.
Just as a follow-up to this..
I've been streaming on Twitch everyday @ 6k with ~0.5% dropped frames on Twitch for about three weeks now. I didn't want to ruin it and speak too soon. Occasionally, I have seen the dropped frames jumps to 2-4% dropped frames (maybe once a week or less), and very rarely the dropped frames reaches absurd amounts (20%+) for a short amount of time (has only happened twice very briefly).
I still have no idea what fixed it (for the most part).
I talked to a technical support supervisor and they sent a technician out because they saw that my upstream channels were "degraded". When I asked the technician what he did, he told me that my upstream was failing the test on his mobile tester thing, and that he changed "all" the lines. I also installed wifi drivers around the same time (was advised to do so by another streamer on Rogers network), but I doubt that's what fixed it..
In any case, slight derail, but is it normal to see higher pings and spikes to my modem than the other Rogers hops? The last couple of servers are in Dallas, and yet the spikes are still higher.
My modem is a CGNM 3552, and I noticed this way before signing up for trial firmware. I've been having latency problems in TCP games since I've been with Rogers (I used to have the same issues on CGN2ROG).
YT is still mostly not working. But around 1AM this morning I give it a shot. Half an hour with less than 0.5% dropped frames... I'd call that good enough. 6000kb/s stream.
Quickly switched over to Twitch, same OBS settings (720p/60, high profile, faster CPU preset).
6000 kb/s failed miserably well over 20% dropped frames. And uploads dropping below 1000 kb/s.
Tried 1500 kb/s, nearly 100% perfect.
Tried 3000 kb/s, about 97% good ... not ideal but it's OK.
What I still find curious man, 1500 is perfect. But how come 3000 and 6000 in the same time frame... will drop all the way down to absolute zero?
I've even streamed a black screen (no image/small cpu/gpu hit), and the same will occur. It does seem connection related.
CS has suggested swapping my modem. But is it also possible for me to get on the trial firmware program?
Thanks for taking the time to do the WinMTR.
It's interesting to see that we both have higher average ping and spikes to the modem, even though you're on a CODA-4582 and I'm on a CGNM-3552. Perhaps it's because they share similar Puma technology?
It truly is mind boggling how Rogers restricts us to faulty hardware (Hitron modems).
It just seems like the whole trial firmware project is just an ongoing search for a band-aid for a (very) poor business decision (I'm guessing Rogers is probably locked into some sort of agreement with Hitron). Don't get me wrong, I really do appreciate all the effort and time invested into helping Rogers customers, but isn't there something fundamentally wrong here?
@scratchieepants, personal opinion, your modem is fine. If you can stream at 12.5kb/s with next to nil losses, then the modem doesn't have an issue. It was interesting to see the switch from youtube to twitch. Beyond the CMTS you would have a different path to the ingress servers. Its possible that the same path is used beyond the CMTS, to some point before they branch off.
In any event, the fact that you can stream to youtube with OBS shows that the modem and CMTS can support streaming at higher rates. The question is, what happens beyond the CMTS?
Are you able to tell if OBS is using IPV6 at all? Looking online it would appear that OBS is IPV6 capable by now, just wondering if its used in both cases, Youtube and Twitch?
So, when you see that streaming doesn't work, that tells me that there is either a noise issue with the upstream channels when they arrive at the CMTS, or, the CMTS or upstream server suffers from congestion.
Here's something that you can do, when streaming works and when it doesn't:
1. Call tech support and ask the Customer Service Rep what the upstream channel Signal to Noise ratios and Signal levels are? You can see those as they aren't reported back to the modem. The CSR can see both sides. So, the question is, is there a large difference in those values from periods where streaming doesn't work, to periods where streaming works? Noise in the cable system could be an issue. Usually that also shows up in the downstream channels as well, but, its worth determining if there is a large difference at the CMTS.
2. Ask the CSR to pass you onto a Level II tech. The question for that tech is what it the current load on the CMTS in terms of a % load? Once again, the question is, is there a large difference in those same time periods? Is the CMTS highly loaded when streaming doesn't work? That would point to a need to spit the CMTS load among two or more CMTS systems.
So, if the answers are that there isn't a noise issue, and that the CMTS isn't carrying a large load, then the answer lies further upstream, either within Rogers network, or wherever Rogers connects to Youtube and/or Twitch. That might be a multi ISP path.
So, the hunt continues .... little by little.
@dtx1, @Makaveli99, the second hop outbound is problematic, either in the case where the modem is running in Gateway mode, or where its running in Bridge mode with a router behind it. In this last case, the modem acts as an invisible bridge to the CMTS. The second hop in both cases is the Casa Systems CMTS.
In both cases, an IPV4 or IPV6 ICMP ping will result in ping spikes, usually up into the 80 to 90 ms range. That's an internal modem issue that applies to more than one modem of Hitron modem. I've asked to have that corrected as it once was with the CGN3ACSMR or CGNM-3552, don't remember the details now. So, it was corrected in another modem, it shouldn't be hard to carry out that same correction for current modem firmware, personal opinion. You will also see false packet loss indications from the modem and CMTS if you carry out a ping test beyond the CMTS with Pingplotter. A command line ping test will confirm or dispel the false packet loss indications from the modem or the CMTS.
The end result is that you can't rely on a ping test to the CMTS to look at response times and latency between the modem and CMTS. You can still use a ping test to the CMTS IP address to check for packet loss, and if you let that run for a 24 hour period, you should only see somewhere around 10 to 20 packets lost throughout that time period.
From my own testing, that indicated ping spike from the CMTS doesn't have any effect on results to targets beyond the CMTS.
@dtx1, in terms of the TCP issues that you indicated, Hitron and Rogers started to issue firmware updates to resolve latency issues with the Puma 6 modems in Sept 2016. That was an IPV4 ICMP fix only. Since that time, there have been further updates to reduce other protocol latency, but, no one, including Intel, Hitron or Rogers has come out with any proof in the form of multi-protocol ping plots which indicate that Intel has finally put the Puma 6 latency issue to bed. I've been running the Puma 7 CODA-4582 since Dec 2016, so, I haven't had any opportunity to determine if the Puma 6 latency that I noticed much earlier that year has finally been resolved.
I can say that the 4582 does not suffer from the latency issues that the Puma 6 modems had. The one exception is the rather annoying ping latency to the CMTS. That was introduced in version 220.127.116.11, and its been there ever since.
To determine if there are issues remaining with latency thru the Puma 6 modem, that would require someone to run the following in both IPV4 and IPV6 for both Gateway and Bridge modes:
1. ICMP ping test to the Rogers DNS
2. TCP/IP ping test to the Rogers DNS
3. DNS query to the Rogers DNS to generate UDP responses which can be clocked and plotted.
That's another discussion, with a set of instructions that I need to post for anyone to carry out.....
I will try your suggestions when I have more time to focus on it.
But currently, anytime of day.
The Twitch Bandwidth Test is an absolute failure with any of the logically ideal ingest servers.
like 0's for QC and Toronto. Under 50 everywhere else (when running the medium test, short test does yield some awesome results on occasion, but repeating it again and again shows that it doesn't work properly).
I've been a small streamer for over 2 years now, and have been having this issue on multiple modems. rogers is a joke.
I have learned to not expect anything to get better, instead, to get worse, and to pay more over time.
Quite literally the worst customer service, and worst performance per canadian dollar.
There are no solutions.
@delos, can I draw your attention to the following thread:
Could you post your circumstances in a post within that thread so that @RogersAndy can have a loot at it.
So i was having the same issue as everybody else, when streaming to RTMP upload would drop to 0 every few seconds.
If i stream FTL to mixer upload holds steady. RTMP to same server, upload drops to 0 every few seconds.
However after a call to office of the president my issues magically went away. No more streaming problems.
However i just upgraded to ignite tv and now my problems are back.
The plot thickens, this is on the technicolor broadcom modem.
Thanks so much for sending us a PM, we'll be happy to assist and continue there in identifying the issue. Feel free to post your traceroute results that show the issue in effect here though when able. (Be sure to remove any personal information such as IP or MAC address). This way others who may have the same issue can have a stepping stone for identifying their latency.