Has anyone experienced getting drastically different speeds measuring either wifi or data with these two apps? With OpenSignalMaps, I'm getting a little over half of what I'm getting with SpeedTest.
Which one's more accurate?
clankfu said:
Has anyone experienced getting drastically different speeds measuring either wifi or data with these two apps? With OpenSignalMaps, I'm getting a little over half of what I'm getting with SpeedTest.
Which one's more accurate?
Click to expand...
Click to collapse
There are many different speed test apps or websites out there. May of them utilize different technologies like Java or Flash and therefore can give different values. This has been an oft debated issue in the internet world. Personally I believe SpeedTest because it usually gives me the speed value I expect. I'll make two points though:
1. You have to be sure you are talking about the same number, i.e. Megabits per second vs. Megabytes per second
2. Depending on the location of servers you are connecting to to run your test, or the sites connected in between you can get different numbers. i.e. One test connect to a server 100 miles away and the other server being 1000 miles away.
both apps give me basically the same results. make sure u account for bits vs bytes. Openmaps is way more advance. the tower map and compass is awesome. especially when trying out prls
raptoro07 said:
There are many different speed test apps or websites out there. May of them utilize different technologies like Java or Flash and therefore can give different values. This has been an oft debated issue in the internet world. Personally I believe SpeedTest because it usually gives me the speed value I expect. I'll make two points though:
1. You have to be sure you are talking about the same number, i.e. Megabits per second vs. Megabytes per second
2. Depending on the location of servers you are connecting to to run your test, or the sites connected in between you can get different numbers. i.e. One test connect to a server 100 miles away and the other server being 1000 miles away.
Click to expand...
Click to collapse
And don't always trust the server closest to you... I live in Tampa, and be it over the phone nor the PC, I get faster and better pings and speeds using the Atlanta server...
Sent from my PG86100 using Tapatalk 2
spaceosc said:
both apps give me basically the same results. make sure u account for bits vs bytes. Openmaps is way more advance. the tower map and compass is awesome. especially when trying out prls
Click to expand...
Click to collapse
They both measure by bits don't they?
clankfu said:
They both measure by bits don't they?
Click to expand...
Click to collapse
Yes, I believe so. If one is eight times faster than the other, then one is bits the other bytes. As for different speeds, the server makes a big difference. Speedtest usually uses the closest server so it's speeds tend to be faster for me. I have Bright house in Tampa and they host the local speedtest server, so my home speed always matches my advertised speed .
Save the Drama for your Mama with Tapatalk 2
coal686 said:
Speedtest usually uses the closest server so it's speeds tend to be faster for me. I have Bright house in Tampa and they host the local speedtest server, so my home speed always matches my advertised speed .
Save the Drama for your Mama with Tapatalk 2
Click to expand...
Click to collapse
Gotcha...makes sense.
Depends on methodology
Going to jump in here as one of the developers behind OpenSignalMaps.
When it comes to speed test there isn't really such thing as 'more accurate', its really just a question of the methodology you employ. There are all sorts of different choices you can make, such as, do you use the closest servers? how many http connections do you use? how large of a file do you download? do you discard any data points? The answers to all of these questions really depends on what exactly you are testing. The ethos behind OpenSignalMaps has always been that we are trying to measure the actual user experience as accurately as possible so all the decisions we've made in our speed testing methodology are in order to recreate what a user would experience in general usage of their device. I think the methodology behind Ookla's Speedtest.net app (which is fantastic) could be more accurately described as testing the raw, highest capacity of your network connection. Thus in general you should see higher speeds with the Speedtest.net app than the OpenSignalMaps app, but our aim is that our speeds would more closely mirror what you actually experience on your device.
One example is server choice: SpeedTest.net encourages you to use the nearest server, usually in the same city. We do use many different servers (to remove geographical bias), but in general the server will not be so close to you and we would argue that in general browsing of the web you aren't likely to be served by a server so close to you. We actually host our Speedtests on multiple popular CDNs in order to emulate a large proportion of general web traffic.
Another example is that SpeedTest.net will discard some of the result data before calculating the average speed as they argue its representing the TCP/IP algorithm rather than the raw HTTP throughput of the connection (see page 20 of an excellent paper "Broadband Speed Measurements" by Bauer, Clark and Lehr - google it). At OpenSignalMaps we don't care if the bottleneck is the TCP/IP algorithm or the actual network connection we just want to give you the most accurate representation of your connection speed.
Just to clarify I'm not trying to argue one is better than the other and we have great respect for Ookla and other speedtest providers, just that it depends on what you want to test. If you are looking to see if your ISP is providing you with the max speed that the are advertising, Speedtest.net is probably going to be the best tool for that. If you want to try and gauge what speeds you are actually seeing in day to day usage then we are trying to build OpenSignalMaps as the application for that. We still have a huge amount to do to achieve that but we have a lot of great features in the pipeline
bmdgill said:
Going to jump in here as one of the developers behind OpenSignalMaps.
When it comes to speed test there isn't really such thing as 'more accurate', its really just a question of the methodology you employ. There are all sorts of different choices you can make, such as, do you use the closest servers? how many http connections do you use? how large of a file do you download? do you discard any data points? The answers to all of these questions really depends on what exactly you are testing. The ethos behind OpenSignalMaps has always been that we are trying to measure the actual user experience as accurately as possible so all the decisions we've made in our speed testing methodology are in order to recreate what a user would experience in general usage of their device. I think the methodology behind Ookla's Speedtest.net app (which is fantastic) could be more accurately described as testing the raw, highest capacity of your network connection. Thus in general you should see higher speeds with the Speedtest.net app than the OpenSignalMaps app, but our aim is that our speeds would more closely mirror what you actually experience on your device.
One example is server choice: SpeedTest.net encourages you to use the nearest server, usually in the same city. We do use many different servers (to remove geographical bias), but in general the server will not be so close to you and we would argue that in general browsing of the web you aren't likely to be served by a server so close to you. We actually host our Speedtests on multiple popular CDNs in order to emulate a large proportion of general web traffic.
Another example is that SpeedTest.net will discard some of the result data before calculating the average speed as they argue its representing the TCP/IP algorithm rather than the raw HTTP throughput of the connection (see page 20 of an excellent paper "Broadband Speed Measurements" by Bauer, Clark and Lehr - google it). At OpenSignalMaps we don't care if the bottleneck is the TCP/IP algorithm or the actual network connection we just want to give you the most accurate representation of your connection speed.
Just to clarify I'm not trying to argue one is better than the other and we have great respect for Ookla and other speedtest providers, just that it depends on what you want to test. If you are looking to see if your ISP is providing you with the max speed that the are advertising, Speedtest.net is probably going to be the best tool for that. If you want to try and gauge what speeds you are actually seeing in day to day usage then we are trying to build OpenSignalMaps as the application for that. We still have a huge amount to do to achieve that but we have a lot of great features in the pipeline
Click to expand...
Click to collapse
Thanks for the response and explanation.
Related
Saw this on ppcgeeks, I was wondering what xda users' speeds were (rev a or not) on their titan...
www.dslreports.com/mspeed (1mb test)
1603k
opera mini
rev a. update
titan, sprint leaked rom
chicago (NW suburbs)
anywhere from 200k to 1200k
in ie
but, i can't seem to run the test again through opera mini, says to fast/unstable, do i need to clear some cache from opera? so many of these tests seem so off, like running from ie vs opera..
can anyone recommend good websites for testing from the phone (non flash) for accurate results?
I did the 1 meg test for the dcd's titan rom 2.0.0 on my Verizon Titan w/Rev A 3.27.00 Radio ROM
911 kbit/sec - 985 kbit/sec
0.293s to 0.198s latency
9.228s to 8.445s d/l time
I am getting 900k or better on a regular basis, same test site. I cant use speedtest.net for whatever reason cause the phone wont load properly.
If anyone is using other speed test sites, let me know.
I get 200 to 900 on Alltell in Springfield Missouri on the stock ROM. My results are all over the board. It bounces up and down plus or minus 300 kbits from one second to the next in the same physical spot. I check it often since it was my reason for switching to Alltel. It seems strange that it supposedly fluctuates so greatly from one second to the next as Im standing still.
Harrisburg, PA
First Run
837kbit/sec
0.837s latency
10.401s download time
Second Run
874bikt/sec
0.994s latency
10.014s download time
I'm getting around 900k, but those damned lacked of video drivers really kill the fun of EVDO when trying to use web browsers.
vaxick said:
I'm getting around 900k, but those damned lacked of video drivers really kill the fun of EVDO when trying to use web browsers.
Click to expand...
Click to collapse
I don't know too much about the lack of drivers; but what does the lack of drivers do to hinder the phone / web browsing?
On Telus, using Opera in my building with ****ty reception, 1039 kbit/sec.
0.329s latency
8.129s d/l time.
That is on a1jatt's custom Telus rom on the latest radio 3.27.00
Wow, a restest shows:
1996 kbit/sec.
0.22 latency
4.323s d/l time.
I'm in Ottawa, Ontario.
800-1100kbps
I used the mspeed a couple times but I prefer speakeasy.
http://www.speakeasy.net/speedtest/
You'll need flash
http://www.adobe.com/products/flashplayer_pocketpc/downloads/player.html
I cant BELIEVE I had forgotten about reinstalling Flash7
thx
haroonie said:
I don't know too much about the lack of drivers; but what does the lack of drivers do to hinder the phone / web browsing?
Click to expand...
Click to collapse
Puts a lot of stress on the CPU to have to both load and render the pages. You see the same effect on a PC if you have no video drivers installed.
There's a non-Java test at www.mobilespeedtest.com. With ROM version 2.09 and only PIE running, I got top EVDO speeds around 955kbps when testing five times in a row, with an average of about 865k when throwing out the one-time 245k result.
i'm in greensboro NC highest i've got around here is 12xx something
Sigh. Close to goin nuts here. Sprint Touch Pro EVDO Internet Speed vs. Mogul.
Trying to be clear and concise as possible here. I've searched forums here and at PPC and Google. I don't think I'm that much of a n00b ROMing and customizing my Mogul for the past almost 18 months.
Heres the issue: I have, for the last 3 days, Activated, Deactivated, and Reactivated my Mogul and my Touch Pro (Raphael) to test something I am experiencing.....
Severely sluggish Internet Speed Tests on my Touch Pro.
I am using http://www.dslreports.com/mspeed?t=1&s=1024&w= which is the 1MB file download.
Sitting in the same exact seat, either here at my office or at home I get around 900 down on the Mogul. IF, and only IF, the Touch Pro is able to completely download the dummy file for the speed test (it usually craps out right at 460) the test is rated in the 150 down range.
I'd like to also add, that when downloading the file, I will have 4 Bars and the In/Out Airows and then all the suddent they will dissappear and reappear... this is usually when the time out occurs.
I started tooling with this because I found I could not watch Youtube or Sprint TV like I used to be able to on the Mogul without it to either take a long time to buffer in the beginning or just play the first minute or so and then start buffering forever again.
I can not figure this out. Is anyone else experiencing EVDO performance issues versus their previous Sprint EVDO Rev. A device?
I have tried all of the different Radios, no change, and different ROMs from Stock, to Boreds, to MightyMouse and back to Stock.
Care to weigh in?
Tweaking your Sprint Touch Pro
Try downloading the Advanced Configuration Tool from:
http://www.touchxperience.com/en/advanced-configuration-tool.html
It is a free application and helps to tweak the bandwidth for download/upload and a lot of other settings. It also adjusts the sensitivity of the screen so you don't have to press as hard during scrolling. There are over 200 adjustments that can be made and some of them have "advised" settings to help you out. It makes a noticeable difference for internet speeds and other items.
Also get Net Framework 3.5 from Microsoft:
http://www.microsoft.com/downloads/details.aspx?FamilyId=333325FD-AE52-4E35-B531-508D977D32A6&displaylang=en
PHM Registry Editor will also be needed to adjust some settings depending on what Mods you download:
http://www.1800pocketpc.com/2008/09/04/phm-registry-editor-registry-editor-for-pocket-pc.html
I hope this helps some....best of luck
what settings need to be tweaked in the Advanced Configuration Tool and the Registry to fix this because my internet is also very sluggish at times too.
also having a problem. Mogul gets 1500-1900k, my TP gets ~300k. On sprint.
[HOTFIX] - Go under Advance Config Tool and under the network settings tap increase the tcp window size to the greatest value. My internet is flying now just had 1500 kb/s on the Sprint EVDO Network compared to 500k/b i used to have
jayvon2304 said:
[HOTFIX] - Go under Advance Config Tool and under the network settings tap increase the tcp window size to the greatest value. My internet is flying now just had 1500 kb/s on the Sprint EVDO Network compared to 500k/b i used to have
Click to expand...
Click to collapse
Is the TCP Window the only thing you changed? I changed mine and my speed went from 300 kb/s to 190 kb/s.
mtg1974 said:
Is the TCP Window the only thing you changed? I changed mine and my speed went from 300 kb/s to 190 kb/s.
Click to expand...
Click to collapse
Yea it was but make sure your tcp retransmissions are set to 8 or 10 times and make sure DHCP is enabled. Also keep taking the test it might take awhile for connections 2 be received. There might be something interferring with ur connection if like a tweak(cab) or registry edit u might have done.
I just used a program that I down loaded from a Russian site called tcptune. worked great. Don't use any of the presets just play with the tweaks. I have mine optimized, and I am getting about 900-1500 mbs. Works great. You may alos wan to go to DSLreports.com for some tips. I have attached the file
Just Did some Tweaks with the software I recommended. Here are the results from mobilespeedtest.com.
You should be able to get some great results. I have been looking for software to do this for a longtime.
I have attached a screen shot of the results. I would like to get some banter going on tweaks for internet speed re: HTC Touch Pro, and in particular Sprint EVDO rev A.
It's like nobody else does this? I have been tweaking internet speeds on my PC's for years to optimize speed.
So has anybody else been tweaking. I just tweaked my phone with a higher rwin, and have got some great speeds.
My speed test
I have a HTC Touch Pro on the way, should be here in the morning. I ran the speed test on mobilespeedtest.com and my result was 8445 kbps at 3MB on my mogul 6800. Why am I so much faster than what you guys are showing? I hope I can get my Touch Pro up to this speed.
What browser are you using?
I used the software that i provided (tcptune). After doing research, my final settings for Cellular line 1, and Cellular line 2, I set the settings up as following:
Rwin: 98304
Windows Scaling on
TTL 128
MTU Discovery on
Selective Acks on
Time Stamps off
I posted the one screen shot where I got over 2500 kbps on a lucky day
I regularly get 1200 to 1500 kbps with the tweaks I noted.
My previous speeds were 350-450 kbps regularly.
I am on Sprint, and let me tell you these tweaks make a world of difference, with no ill effects.
Slikgman:
Are sure about those numbers. 8445 kbps is 8.5 Mbps
I don't think that Verizon or Sprint even support those speeds.
You are talking like close to Ethernet (10 Mbps) speeds over wireless.
Are you sure you not talking 844 kbps?
TheDempster said:
Slikgman:
Are sure about those numbers. 8445 kbps is 8.5 Mbps
I don't think that Verizon or Sprint even support those speeds.
You are talking like close to Ethernet (10 Mbps) speeds over wireless.
Are you sure you not talking 844 kbps?
Click to expand...
Click to collapse
I know it sounds crazy, thats why I posted it. When I used DSLreport.com I got around 1000kbps and sometime 600kbps. With mobilespeedtest.com Im a Rock Star!!!! I just ran it again and it is 8108 kbps. I am using Opra Mini for my browser.
I ran the same test on my work computer (desktop) to see what it would be and I was reported running at around 3652 kbps. So my phone was faster than my work connection. LOL
I just got my Touch Pro, and I have never sync my mogule to my computer.
How do I get my contacts to my new PRO? I am running Vista and my mobile Device Center is not doing anything when I connect my 6800.
Are you using opera for the speed test?
If yes, you cant use opera.. Try internet explorer.
the 8.5m is not ur speed it is opera's proxy server speed
remember opera mini and skyfire use proxy servers, so all ur internet requests with those browser go from ur phone to there servers and then those servers have super fast internet to grab the page that u requested wich then they will make it a lot smaller and send it right back to u
Internet explorer 425kbps...... That sucks, thanks a lot guys!!!!!
LOL
TheDempster said:
I used the software that i provided (tcptune). After doing research, my final settings for Cellular line 1, and Cellular line 2, I set the settings up as following:
Rwin: 98304
Windows Scaling on
TTL 128
MTU Discovery on
Selective Acks on
Time Stamps off
I posted the one screen shot where I got over 2500 kbps on a lucky day
I regularly get 1200 to 1500 kbps with the tweaks I noted.
My previous speeds were 350-450 kbps regularly.
I am on Sprint, and let me tell you these tweaks make a world of difference, with no ill effects.
Click to expand...
Click to collapse
I set the rwin to 225000 and Im getting a little over 1500kbps right now. I cant wait to see what its like this evening.
When I go to dslreports I get slow speeds but when I go to mobilespeedtest I get phenomenal speeds. Which one should I believe?
I did a search and couldn't find anything definitive, so hopefully some of you can help me here
I've accepted the fact that my TFP won't get stellar wifi speeds but I guess what I can do is tinker with my router to ensure that it's maximizing its signal and throughput.
I have the Linksys WRT-320N router mounted high up on my wall in my 2 bdrm apt. When I'm next to the router, I get a pretty consistent 15mbps up/down but it trickles off to about 5 from my bedroom about 30 ft. away. My concern is that it's rated for a lot more than 15mbps, and my wired internet gets about 100mbps up/down, so surely it must be a router setting?
I've disabled QoS and tried various channels but besides that, I haven't done much. Are there particular recommended settings that I can play with to try to maximize my speeds? I also made sure my firmware is up to date on the router and have reset it a few times.
A question for you: why did you disable QoS?
How about using QoS to maximize wifi output/priority to your prime when its connected?
Why would you even use qos just for your prime? That's like a 6 lane highway and only allowing one car at a time. It's pretty useless.
I doubt you'll get any difference at all. Qos only prioritizes traffic so that more important traffic can pass first.
Try unlocking your prime and installing another rom, try another router, or move your router to a center location. You could try to update your firmware or even see if you can find any custom firmware for your router.
QoS has always been a pain in the ass so I would disable regardless.
But ASUS does need to publish their findings on recommended settings to optimize performance.
i disabled QoS based off another post about smoothing out the throughput from the router. either way, i've noticed neither a positive or negative result of this.
as far as custom firmware, i'm looking into the dd-wrt route, but would want to tweak the router with the stock firmware as much as possible to see if i have to resort to doing custom
Chances are the limitation will be the transformer's wifi chipset. If you flash your router with dd-wrt you might be able to get slighty better performance.
Don't know that its worth the effort. 5 mb/s is adequate for almost all web except downloading large files and vnc. My experience has been if you want performance your pretty much stuck with something that has gigabit lan.
Sent from my EPAD using XDA Premium HD app
I get about 0.5mb/s speed on the local Wi-Fi network. I get up to 8mb/s on the computer (802.11N card). What about you? Maybe it depends on kernel/rom..?
Sent from my GT-I9100
Little answer please
Sent from my GT-I9100
You didn't get an answer because this info is easily found via a Google search Look for the SGS2 entry at Wikipedia, then follow the links to the article on the relevant WiFi standards.
Oh yeah "802.11 a/b/g/n Wifi" is it? Do you know which speed each of these letters is supposed to give..? I do, so i'm going to tell you that with 802.11n (which my router fully support) you're supposed to get up to 450mb/s. In reality you get let less, but it's still up to about 250mb/s, now you can even divide that by eight, which result in 31,25, far from 0.5 (the speed i get, if not lower, which my 1st post mention, did you read it fully..?) isn't it? Now you might see things that others don't, if so just tell me.
Hmm..?
Sent from my GT-I9100
It depends on many things. Bug in the kernel, faulty hardware, the distance from the router (phones have smaller antennae than laptops due to the size constraint, so laptops will constantly get higher strength signal than phones), faulty routers are all factors, if your phone is the only wifi device that's not getting the speed it should.
I don't know what the real life max speed of GS2 is, but I've hit somewhere between 20~50Mbps on a N router that has 1000Mbps internet connection (~750Mbps realistically, and I wasn't the only one connected to the router). So it should be able to get 8Mbps that your computer is getting, assuming the GS2 is right next to the router.
It doesn't change depending on placing.. And I can go faster when downloading. So I don't think hardware is défective..
Sent from my GT-I9100
I just got a gigabit connection, and my 2x2 mimo laptop is able to get close to 600mbps down aI just got a gigabit connection, and my 2x2 mimo laptop is able to get close to 600mbps down and up on wifi, but my phone struggles to even pass 200 (with a link speed of 867). What are the maximum that you guys are seeing, and do you have an tips or settings to achieve that? My sisters' iPhone is able to get over 500mbps down, so my Moto Z should be able to as well.
Even my old LG G4 which has a link speed of 433 can download 270+ with no sweat.nd up on wifi, but my phone struggles to even pass 200. What are the maximum that you guys are seeing, and do you have an tips or settings to achieve that?
Why?
Umm, so I can maximize my speed and download faster? WTF kind of question is that?
ssmaster said:
Umm, so I can maximize my speed and download faster? WTF kind of question is that?
Click to expand...
Click to collapse
Sorry, but I can't imagine for what purpose you could use 200+ Mbit/s transferrates on a Smartphone. Regardless of my personal opinion, I have absolutely no clue how to "optimize" the settings to reach these rates. But I will test it for myself so I can provide you some comparative values. :good: