The Analysis of Dissertation Results and Discussion: Video Streaming Experiment

2021-06-14
7 pages
1849 words
University/College: 
George Washington University
Type of paper: 
Dissertation results
This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

I conducted an experiment with three laptops placed at different intervals from the source of the wireless networks. These computers A, B, C were placed at intervals of 10metres, 50metres and 100metres respectively. Computer A was running windows 8.1 operating system, which required an update. Computer B had windows 8.1 OS, which was not up to date either. Computer C OS was windows 10 updated. During the test other devices such as a phone and a tablet were using the same wireless transmitter simultaneously. These variations provide a diversity of results of the experiences of clients while for streaming video, to determine how high interference or mobility may impact the available bandwidth during streaming session, which leads to an unexpected buffer underflow event.

The experiment used; three computers with wireless network adapter cards and wireless transmitter; IEEE 802.11g with an Ethernet connection. The software used to be streamtest.net, an online stream testing tool that measures the audiences experience; the results generated are used to troubleshoot any problems with the end users favorite stream. The tool was used to examine the effect of realistic viewing to the performance, the test analyses and samples the relevant test parameters and data points that constitute the individual end-user streaming experience. Eminem's music video, Am Not Afraid, URL was retrieved from YouTube,

Then posted it to the "Enter Stream URL" option box, and pressing the "Start Test" button.

The tool runs the tests, and real-time pictures are taken to be used later for analyses.

 

Several other variables were adapted, and the test runs until the data targeted results gathered. The relative speed of the internet was swaying between 6Mbps and 39Mbps.

Results

First Experiment

The regional streaming test recorded a 75% streaming rate success, in all the tests.

The first test was run when the Computer A was 10 meters away from the source, with a 360p video streaming quality.

The data recorded an unstable streaming with the bitrates going up and down, although no buffer events gathered. In this scenario, the bitrates per second had relatively minimal fluctuations compared to the other data recorded. This test also recorded the highest amount of bytes received within the 30 seconds the test run. Total bytes accounted to 72221.62MB. The bitrate, bandwidth and buffer events graphs were as below:

The second test, with Computer B placed 50 meters away from the source, provided the following results:

The results recorded shows a relatively slower pace than the previous, although no buffer events were observed. The network speed was as shown below.

The bytes received totaled up to 69642.10MB.

 

The last computer C that was placed 100metres away from the source results were way lower as indicated below. No buffer events were observed either.

The total bytes received were 30951.78MB

 

All the above results used the same streaming quality, which is the default YouTube quality of 240p while some used 360p, which has minimal variation in terms of quality.

Second Experiment

I decided to run several other tests using different streaming qualities and watch the difference in the kbps, the amount of bytes received or if there were any buffer events. This test used the laptop that was 100 meters away from the wireless transmitter. The qualities recorded were 1080p (HD), 720p (HD), 480p and 360p.

The experiment had these results in general;

1080p (HD)

720p (HD)

480p

360p showed similar results as 240p and 144p therefore providing little or no effect to this research.

Data Analysis: First Experiment

In the first test the results revealed that there was a certain percentage lost as the distance increased away from the transmitter/router. The percentage between 10metres and 50metres was 2%, while that between 10metres and 100metres was 15%. It should be noted that the below 10metres there were no loss in the streaming quality as it was 100% in total. Besides the three computers there were other computers and device (phones and tablets) that were using the same wireless transmitter although the specific number was not evaluated.

The difference in the bitrate measure in the three different tests showed the diverse video quality that the users experienced. The higher average bit rate means that the quality image, in the first experiment 68020kbps and the second experiment recorded a higher rate at 68785kbps, the third experiment recorded it bitrate rate at a disappointing 15245.75kbps. If the results were repeated the test showed different bitrate rated revealing that the network was constantly fluctuating at different levels. From the graphs listed the bitrate graph did not record as smooth curves at any give level during the experiment. The bitrate indicated how many bits of video that was transmitted over the specific period of time, which was 30 seconds.

Bytes received (brandwidth graph), which is the rate at which traffic was carried to the network, this was varied by the actual internet connection speed, which was varying from 6Mbps to 39Mbps, this speed was not constant as it was going up and down as the streaming continued. From the graphs in all the three cases the curve is wavy and produces a straight curve at no specific time. That is why the momentary test of bandwidth could not indicate bandwidth availability throughout, if the music video was fully run. The reasons why the bandwidth fluctuated is that; the users were streaming other programs while at the same time running the test. For instance, in one occasion the user was streaming the same video on YouTube and running other tabs that required data flow, doing several activities or streaming other videos on the same computer while running the stream test reduced the amount of bandwidth available.

The second reason was; the three computers were used simultaneosly, as were many other networks that were connected to the same wireless transmitter, therefore, this caused the amount of bandwidth to go down and up depending on how many networks were sharing the same transmitter. This is a common occasion, where a lot of users causes the network to slow and viceversa. Lastly, the wireless transmitter offers network connection to several other users, therefore, the satellite provider equally divides the amount of available bandwidth to me and other users as the demand increases, thats why the bandwidth rate or bitrate were not stable.

A deeper look at the effect of distance in the streaming strength, according to the literature review, research showed that wireless devices have limitations when it comes to their range. This was proved in this research when the strength of the signals fluctuated with the increased distance from the router. As we moved further from the router the signal strength became weaker. In addition, since the research was conducted in an indoor setting, physical obstruction might have caused the effects that were recorded. These could have been instances, such as cabinets, ceiling, thick wall among many others. Any obstacle that was present between the wireless device and the wireless router created a signal loss (Troubleshooing Guide, 2013).

Other factors could have been wireless interference, since the region is hubbard with several wireless transmitters, this might have caused a troll in the streaming speed and quality. The region is endowed with several institutions and big businesses that run the same frequency as mine, which is 2.4GHz, might have caused intrusion and upset the wireless connection and signal, other wireless might have been the microwave oven in the house which at some point was being used or Bluetooth devices which is a common destruction. Each of these interferences reduce the streaming quality at a negligible amount and when added together the interference is astounding.

Unfortunately, no buffer events were recorded, meaning there was no waiting time before the video actually played. The quality of the video streamed was 240p or 360p, YouTube default settings, which sometimes goes as low as 144p sometimes, however this quality is very low and provides blurry images that any user dislikes, although the streaming speed might be astounding on the other hand the experience is not impressive and sooner or later the user will become disinterested. This is precisely why there was no buffer fill recorded, this also attributed to the high rate of data collected. Therefore, the second test was to specifically show how different video qualities affected the streaming speed.

Data Analysis: Second Experiment

This experiment explored the strength of streaming at different levels of the video quality. Those qualities were as shown in the diagram above, 1080p (HD), 720p (HD), and 480p. Each of the event recorded a buffering event. 1080p (HD) had a low bitrate (24938.73kbps), data received 33531.29MB and the buffering took close to 2 seconds of the streaming time, after the buffer the bitrate was relatively smooth, see figure above. 720p (HD) had a bitrate of 14819.59kbps, the data received was 30951.78MB and the buffering and lag took close 3 seconds. 480p had a bit rate of 5168.80kbps and the data received as 30488.44MB, buffering took 3seconds of the streaming time.

From the results above it is quite hard to determine what was the precise cause of the observation, since several variable caused the effect, although like the other experiment several of the factors were into consideration. A higher video quality like High Definatition quality(HD), a lot of data is collected thus explaining why the data collected was greater for 1080p than in the other lower levels. Thus, at a higher video quality more data is required, thus leading to lag length while streaming. In this experiment the lag length was short but this can be attributed to the higher bitrate. When the buffer was filled the video continued playing, from that point the download rate kept up with the bit rate, which is quite smooth, thus the stream was in one smooth playback.

However, this is not always apparent in all mediums, sometimes the buffer is drained and without enough video to continue, playback halts like in the case of 480p when the buffer appeared in the 18th second of streaming, thus three more seconds were added to the streaming time due to the lagging. If in any case the streaming would have continued for a longer period of time the results would have provided a clear check on the delayment caused by buffering in each case. This experiment shows that lower qualities use relatively low bitrates which is equivalent to the bytes collected, therefore the streaming is much smoother, since less data is needed to run the stream.

Any video streaming platform offers different video quality to the user while streaming such as Netflix and YouTube. Video titles are available in multiple levels of video quality. Some sources provide a platform that automatically determines the level one receive by analyzing the current Internet speed. However, one can select the video quality level, and the level may change depending on actual network conditions. The provision of different qualities to the user allows the user to evaluate what quality they want in regards to their current needs, some streaming does not require a high level of quality. In addition, the user can review their network connection speed and the suitable quality appropriate for their streaming.

Have the same topic and dont`t know what to write?
We can write a custom paper on any topic you need.

Request Removal

If you are the original author of this essay and no longer wish to have it published on the collegeessaywriter.net website, please click below to request its removal: