Indigo speed test very inaccurate

[daflyboys]

Used to really like the Indigo speed test, but I have to think it's very unreliable now. Here's comparable results for my PC which is hard wired to my router. Clockwise, from upper left corner, these are from RCN, Indigo, Ookla and the God of War Wizard.

 

[RavRob]

Were they all from the same servers?

 

[makal]

For me it works well

 

[daflyboys]

I don't know what server Indigo is picking. You can only pick file sizes, no?

 

[RavRob]

Sorry, I don't believe in those speed tests. They are just too dependent on the server used. I believe on what I can play without buffering regardless of what speed any speed test tells me.

 

[vid]

You are correct, I have tried numerous testers, and the Indigo is widely different on the very low side.

 

[Shinzin]

Indigo uses the speed test (and gets its results) from another site. Speed tests use a lot of server bandwidth. You get better accuracy the longer it runs (the bigger the file). I was getting fairly accurate results when the download was taking 30 seconds or longer. The results were matching the results from the site itself (using my browser), other speed test sites, and what my provider advertises that I have. The faster your connection is, the bigger the file you need to test with to get past the 30 second threshold (I think its a math thing ).

[edit] Also ravrob is right. The results can vary depending on server load, internet routing, and network load.

 

[Shinzin]

I just tried it again without my vpn and the results were very close to what speedtest.net gives me in my browser.
Its close to what my provider says I should bev at also.
Going through the vpn is slower, just as it should be.
I used the 100M link which took just over 30 seconds for me.
It says 18.9. My provider says i should be 20 and speedtest.net gives me 21.
So its not off far enough to really worry about.
none of the speed test are going to be perfect.
If you find one that gives you the same speed over and over, then I would question if that site is really doing what it says

 

[Shinzin]

Another thing to keep in mind is that just because you get "good" results from a speed test server doesnt mean you wont get buffering from an overloaded source server. Buffering is usually the reason why people want to do a speed test and see what speeds they are able to get.
A more accurate speed test would use the source server that you are having trouble with. But that would also add to the loads on the source server, make a bad situation even worse, and would be self defeating.
Any speed test will only give you an idea of your capable speeds to that specific server, with certain internet routing, and at that particular moment. You get different routing (called hops) with every connection to the server and is why results vary even in the speed tests from the same site.
As with any speed test, use the one in indigo as a guide to get an idea of what your capable of. Not a be all end all indication of what every server on the net will give you.

 

[vid]

Another thing to keep in mind is that just because you get "good" results from a speed test server doesnt mean you wont get buffering from an overloaded source server. Buffering is usually the reason why people want to do a speed test and see what speeds they are able to get.
A more accurate speed test would use the source server that you are having trouble with. But that would also add to the loads on the source server, make a bad situation even worse, and would be self defeating.
Any speed test will only give you an idea of your capable speeds to that specific server, with certain internet routing, and at that particular moment. You get different routing (called hops) with every connection to the server and is why results vary even in the speed tests from the same site.
As with any speed test, use the one in indigo as a guide to get an idea of what your capable of. Not a be all end all indication of what every server on the net will give you.

Ideally it would be better to check the server speed one is streaming from, before streaming, but that will need a miracle, because then they would have no excuse on the poor speed.

 

[Shinzin]

Ideally it would be better to check the server speed one is streaming from, before streaming, but that will need a miracle, because then they would have no excuse on the poor speed.

Yeah thats what I meant by source server.
source server = server you are downloading or streaming from.
But like I said, doing it that way causes its own problems.
Like adding to the load on that server and slowing it down even more.
Even those results will vary from test to test because of how the internet works.
Its best to just recognize the speed test results for what they are.
The average speed that downloaded packets, from one file, take to get from the server doing the testing, to your computer, at that particular moment (including the relative distance from you and the number of connections to other servers, their loads and available bandwidth ect..., in between you and the server you are downloading/streaming from) and not the absolute speed that you will get from all servers, or any one server, on the internet.

 

[vid]

Yeah thats what I meant by source server.
source server = server you are downloading or streaming from.
But like I said, doing it that way causes its own problems.
Like adding to the load on that server and slowing it down even more.
Even those results will vary from test to test because of how the internet works.
Its best to just recognize the speed test results for what they are.
The average speed that downloaded packets, from one file, take to get from the server doing the testing, to your computer, at that particular moment (including the relative distance from you and the number of connections to other servers, their loads and available bandwidth ect..., in between you and the server you are downloading/streaming from) and not the absolute speed that you will get from all servers, or any one server, on the internet.

The reason I do a test is to check my system, ISP and VPN, and I so I want a test server that will not add or subject to the MBPS; so that any problems streaming it is not on my side. If I try 3 test servers and they indicate the average is over 30 MBPS for example, and Indigo tester is showing less than 6MBPS . Then I assume the Indigo tester is suspect and I can't test my side properly.

 

[Shinzin]

I want a test server that will not add or subject to the MBPS; so that any problems streaming it is not on my side.

Thats what I'm trying to tell you. They all do it by definition LOL. Its only the degree of which that will vary. There is no way around it.

If I try 3 test servers and they indicate the average is over 30 MBPS for example, and Indigo tester is showing less than 6MBPS . Then I assume the Indigo tester is suspect and I can't test my side properly.

I havent seen anything like that as long as i chose a download size that takes at least 30 seconds to download. Unless i try to compare vpn and no vpn.
The vpns slow it down considerably. Depending on the vpn and the level you pay for they can slow it down that much. When I use a vpn its no where close to what my ISP says i should get, but thats how it works. Its not the tester thats at fault. Its the vpn that doesnt allow the throughput at a higher speed unless you pay more for it (if they even offer it).
I have tried differant sites repeatedly with vpns,without vpns, inside of Kodi (indigo), and outside of kodi (from both internet browsers and python scripts). They All have had similiar results.