Flixwagon v.0.4.45 iPhone 2G Network Video Settings Tests

Introduction

Only hours after yesterday's post, I received notification that Flixwagon had been updated. In this version, the application includes new options for video settings. There are four in total*.

1. Low quality, EDGE friendly. Low quality mode: suitable for broadcast over slow networks such as EDGE (176x144, 5 fps)
2. Smoother (Recommended). Smoother mode: Makes moving objects look even better. Suitable for most filming scenarios. (176x144, 12 fps)
3. Sharper. Sharper mode: Adds more details to the the** video image, but less effective in capturing motion. (236x176, 7 fps)
4. Still objects (See "Details" below). Still objects mode: Highest resolution. Lowest FPS. Only suitable for broadcasting still objects such as printed text and pictures (300x224, 3 fps)

Methods

Initially, I set out to perform a quality test, just to see if there was a significant difference in the output between the various settings. However, as often occurs with experimentation, I ended up with unexpected results. Accordingly, my tests from this morning became more of an evaluation of the setting descriptions themselves rather than the video output quality. Below are the basic circumstances surrounding my tests.

• Time: All trials took place between 7:58 and 8:45am EST.
• Network: All trials were completed over the AT&T EDGE 2G network. The signal strength was at four bars for the vast majority of the testing. There were a few intermittent spikes where the signal would drop to three bars or rise to five bars.
• Duration: On each trial, I stopped the recording when the timer hit 14 seconds, although some videos still reached the site at 13, 15, or 16 seconds.
• Application Version: 0.4.45

For each trial, I tapped the bobble-head doll to get it moving, and began recording. After 14 seconds, I stopped recording, and started an online stopwatch (http://www.online-stopwatch.com). Once the video had uploaded to 100%, I paused the stopwatch and recorded the time.

Results

Setting	Trial Number	Upload Time (m:ss)	Video Link***
Low quality	1	2:32	Watch
Low quality	7	1:44	Watch
Smoother	2	3:03	Watch
Smoother	8	2:17	Watch
Sharper	3	2:01	Watch
Sharper	4	4:25****	Watch
Still objects	5	1:38	Watch
Still objects	6	1:39	Watch

The two fastest times occurred using the "Still objects" setting. Next came the "Low quality" setting, which had a comparable time to "Still objects" in the second trial, but a longer one in the first. Due to the outlying trial, the data from the "Sharper" setting is inconclusive. "Smoother" had the slowest upload times of all groups.

Conclusions

Upload Performance

The different settings produced surprising results in terms of upload time. Interestingly enough, "Still objects" had the fastest times. While this setting features the highest resolution available, it also employs the lowest frame rate. One might expect "Low quality" to be in the first position, but it was instead a close second, perhaps due to the increased frame rate. Although the "Sharper" data was inconclusive, judging by the normal trial and frame rate, I suspect that this setting would follow "Low quality" as third fastest in the long run. "Smoother" was the slowest, but not by an excessive margin. It also features the highest frame rate. These findings, while not definitive by any means, seem to suggest that frame rate is a vital factor in the time it takes to upload a video via the iPhone and EDGE network.

Another point of interest surrounds the overall application upload performance in comparison to previous versions. Even the outlying trial took only four and a half minutes. In my original use of the application, videos would take at least five minutes, and up to ten or thirty, to upload. This indicates that the technology is steadily improving. It is exciting to see videos uploading in only one to two minutes, even though they were only 14 seconds long. If the performance can continue to increase, perhaps to the point where a video takes as long to upload over the 2G networks as it does to record, users will have a very practical, useful, and powerful application at hand.

Video Quality

After reviewing the videos, I have determined that there is no "best" setting for the iPhone. A large factor in the type of setting that should be used is again, the frame rate. Moving subjects will look much better on the "Smoother" setting than any of the others, thanks to the higher frame rate. The difference that frame rate makes can be seen even in the light nodding demonstrated by the bobble head doll in these trials. On the other hand, the fidelity of video appears to be largely unaffected by the mode that is chosen. To the eye, each setting seems to output roughly an equal picture. Yet, the "Low quality" and "Smoother" options did not handle details quite as well as "Sharper" and "Still objects." All things considered, the situational context should be the primary concern for videographers, rather than the desired output quality.

Error

There were a number of possibilities for error in this experiment. I attempted to control the conditions as best as possible, however the results seem to indicate that there were either external factors affecting upload times (most likely) or that the difference between the video settings is not significant over the 2G network. One uncontrollable factor that I believe may have had an impact on my results is the ability of Flixwagon's servers to receive my videos. The rate of transfer seemed to start out a little slower, increase significantly for trial 3, drop off at trial 4, and then increase again for the remaining attempts. Likewise, only being able to judge by the bars on my phone and not by the exact kb/s means that the EDGE network could have had varying transfer rates between the recordings, even though it remained consistently in the four bar range. Naturally, the number of trials and conditions are the other form of error. This test took place within one hour on a single day and with only two attempts per setting.

To Users

While the difference between 1:44 and 3:03 may be significant from a statistical standpoint, it mostly likely is not in a practical sense. Thus, from a user perspective, upload time is not likely to be a factor that influences which setting one uses. Likewise, there is not a noticeable difference between the video output quality for different settings. I recommend that videographers think first about what they wish to record and choose a setting accordingly. The descriptions give an accurate view of what to expect in the final video (i.e. use "Smoother" for recording motion and "Still objects" for static subjects). It is still worth considering the network connectivity, however. When in a weak 2G signal area, it may be best to use the "Low quality" or "Still objects" setting in spite of subject matter.

Notes

*Descriptions taken directly from the Flixwagon iPhone application v.0.4.45

**A minor typo in the description

***Will open in a new window

****I view this trial as an outlier. There was a 5-10 second delay during the upload, whereby it suffered from "No Network Activity" and subsequently had to reconnect to the video server. Although this delay does not account for the large difference in time, I believe that the servers probably had an overall loss of strength during the trial, which caused a slower upload for this particular video.

Round Two: Flixwagon And Qik On iPhone Over 2G Network

After receiving comments on my last post that refuted the findings from my 8/14/08-8/16/08 tests of Flixwagon and Qik's iPhone video streaming services, I decided to retest the applications.  Both have been updated multiple times since their initial release.  For this experiment, I decided only to examine how well the programs work over the 2G ATT EDGE network.  It is assumed that both can operate acceptably under optimal conditions (i.e. a strong wi-fi signal) and thus, quality would be the only concern, not functionality.  I aimed to see how well the applications run under the most common situation that an iPhone user would find in the U.S.  That is, a 2G cellular network.  My trials were notably limited, however, the findings were consistent with those that I previously experienced. Note that I used the suggested minimal settings from my previous post's comments for Qik and the most comparable setting available ("smoother") for Flixwagon in both trials.

Trial 1: Static Subject in Strong 2G Signal

Flixwagon

Qik

Results: These videos were taken indoors, at the center of a developed campus area. I had full 5-bar signal strength on my iPhone. As can be seen from the videos above, both Qik and Flixwagon are capable of producing video when in range of a strong 2G signal. Each application took time to cycle through after recording. Qik calls it "flushing buffer" and counts down towards 0kb, whereas Flixwagon calls it "uploading" and counts up towards 100%. Either way, a 30 second video will take about 5 minutes to complete after the recording is stopped.

Trial 2: Moving Subject in Modest 2G Signal

Flixwagon

Qik

Results: These videos were taken outdoors, in a local community park. My signal strength was between 3 and 4 bars. In this situation, Flixwagon performed much like it did in the first trial and any other time that I have used it over a 2G network. Qik, however, was unable to completely flush its buffer and got stuck at 4k, which I predict is why the video did not go through.

Conclusions

Flixwagon and Qik appear to have improved their services over the past few weeks. I took notice to the fact that I tend to experience fewer random crashes, freezes, or network time outs than I did originally. This means that streaming is more reliable and predictable, which is a good thing for users. The performance of the applications was not altered by whether the subjects were static or in motion, as long as the camera was kept stationary. Camera movement is a vital factor in the quality of video that can be achieved on the iPhone. It is best to set the shot up and refrain from moving afterwards. Audio quality was well handled in all cases. It is difficult to draw conclusions as to the service capabilities of these applications from such limited testing, but I have confirmed that both are able to broadcast over 2G. Yet, I will point out that one facet of my overall experience has consistently recurred. That is, Flixwagon appears to be much more resilient to poor network connectivity than Qik. Barring non-reproducible errors, anything broadcast through Flixwagon will eventually end up online. On the other hand, the static Qik trial is the only complete 2G video that I have successfully broadcast to date.

My recommendation to both companies in terms of network capabilities would be to continue to optimize for low-end use conditions. Most potential users are going to be on 2G networks. Furthermore, the ideal wi-fi condition significantly limits the range of useful content that can be broadcast (i.e someone's living room, a coffee shop, an office presentation, ...what else?). Personally, I would rather be seeing majestic landscapes, on the spot interviews, and mass gatherings. At times, these scenarios will be entirely devoid of network connections, but in the event that one is available, users should have the opportunity to broadcast the compelling content that these moments offer.

Round One To Flixwagon | Next: Video Recorders

Notice: The testing that produced these findings took place between 8/14/08 and 8/16/08.  Both applications have been updated since.

On the heels of the iPhone video streaming software releases nearly two weeks ago, I immediately set out to test their low-end use capabilities on my trip to Las Vegas.  I was armed with my iPhone 2G and was, for the most part, without wi-fi access.  Here are some of my basic findings:

1) Qik will not work without wi-fi.  This essentially took Qik out of the running on my trip, hence the lack of videos on my profile there.

2) Flixwagon will work without wi-fi (to a certain extent).  Flixwagon does at least attempt to encode a video over the regular 2G phone network.  I was not able to get anything longer than about 70 seconds to encode without timing out or crashing and what did go through took an extremely long time.  This is actually not bad, however, considering the amount of data transfer necessary for video uploading.  There was also a bit of luck involved in the process.

3) Qik allows users to download videos, whereas Flixwagon does not.  I believe that this is an essential feature for this type of service and that it should not be forgotten.

4) Flixwagon and Qik perform roughly the same in terms of streaming/encode/upload when using a strong wi-fi signal.

5) Qik and Flixwagon videos are cumbersome to add meta data to.  Both services essentially require users to log in after their videos are broadcast to add tags, titles, descriptions, etc.  This is due to the fact that one is typically not going to type away on the phone for five to ten minutes to set everything up before broadcasting.  When I took 30+ videos on my trip, it was quite a burden to come back and clean up the data for all of them.  Furthermore, it was somewhat after the fact, since much of the point behind the technology is that things are being broadcast live versus recorded and uploaded (i.e. YouTube).  I am not sure of the exact solution to this problem, but having more meaningful default meta data or "smart" previously used tags that users can quickly choose from may help.  Note that nearly every recent video on either site is yet another "untitled."

6) Both services make videos public by default and then require a setting to be changed each time the application is run.  If users want to make private videos, they should only have to change the setting once and it should be saved until further notice.

7) The iPhone mic is very sensitive to loud noises, including music and speech.  Many of my videos came out with frequent scratchy flare-ups, which were often caused by my own speech.  It is worth noting for other videographers that special attention should be paid to the volume of the objects that are being recorded.  Shielding or muffling the iPhone speakers may be necessary in environments with loud ambient noise levels.

On a side note, I did make one audio recording with the QuickVoice application and it came out very well.  There were many points where it suffered from the same high volume spikes as my videos, however if I had handled that better, the overall quality would be very good.

Note:  In contrast to the video streaming programs, the video recorders were finally released in the past week.  Again, there appear to be two main options.  One is the videorecorder3g, made infamous on the original jailbreak iPhones, and the newer Cycorder.  I plan to test out both of these technologies, but may wait until they are on more equal ground.  Both appear to be in early beta and, according to reports, are full of bugs and/or missing features.  However, I am eager to test these out on my iPhone.

Hall & Oates Concert 8/15/08

The Joint is an intimate, yet comfortably sized, 1200 person venue located inside the Hard Rock Hotel and Casino in Las Vegas, Nevada. On Friday, August 15, 2008, it was home to my favorite musicians of all time - Daryl Hall and John Oates. My seat was located at the fourteenth row banister, which overlooked those who were seated on the floor six feet below. It was slightly left of center stage.

I knew when I ordered tickets for this show, in this location, that it was going to be big and it truly was everything our hearts desired. According to the interactive text message board that displayed fan's messages before the show, travelers had journeyed from all across the world to attend. A number of fans traveled from as far Brazil and the Philippines, whereas many others came from all directions of the United States.

The crowd was one of the things that set this concert apart from others. It certainly was not filled with casual onlookers. Daryl was frequently able to take breaks in the middle of a song, just to hear everyone filling in the words together, which made for an engaging and entertaining experience.

Furthermore, the set list contained a unique blend of older Hall and Oates material, as well as the major hits. The 1970s were well represented by a number of songs that the duo has revived in the past year, such as When The Morning Comes, Las Vegas Turnaround, Uncanny, and Had I Known You Then. Cab Driver was the only post-80s song featured during the show. See the bottom of this post for the complete set list.

Judging by the amount of energy surrounding me, One On One was a favorite of attendees from the opening set. Hall, Oates, and the band had the crowd on its feet, dancing and singing, from the end of Maneater through the first encore. The double encore of Private Eyes that followed was particularly amazing and the most notable part of the entire show. Imagine the entire audience singing and participating in the song's infamous clap, clap-clap chorus. It made for an incredible sounding interactive experience that goes far beyond listening alone. It is no surprise that they have saved the song only for special encores in recent concerts.

It was evident throughout the night that Daryl, John, and the band members enjoy their work every bit as much as their friends and fans. Their timeless sounds have spanned four decades, with a fifth rapidly approaching. In addition to their passion, they have also maintained their performance skills and sound every bit as good as in years past. I never cease to be impressed by Hall and Oates. Although this was my third time seeing them in concert, they played a whopping 7 (of 15) songs that I had heard in my previous visits. Hall and Oates' unique combination of heart, talent, and soul, combined with an ability to create new experiences from songs four decades young is a hallmark of their timeless nature.

Complete Set List

1. Everything Your Heart Desires
2. When The Morning Comes
3. Las Vegas Turnaround
4. Family Man
5. Say It Isn't So
6. Uncanny
7. Had I Known You Then
8. She's Gone
9. Cab Driver
10. One On One
11. Sara Smile
12. Maneater
---------- Encore 1 ----------
13. I Can't Go For That
14. Rich Girl
15. You Make My Dreams
---------- Encore 2 ----------
16. Private Eyes

P.S. I will be attending my fourth Hall and Oates concert in Windsor, Ontario on Friday, November 7, 2008 and am already looking forward to it.

Qik? Flixwagon? Video Streaming B[l]ogged Down

To update you on my attempts at broadcasting live video from the iPhone, it is evident that the rest of the world has heard the news since my initial tests last night.  As of right now, both the aforementioned Qik, and its competitor, Flixwagon, appear to have completely bogged down servers.  It is common for new services to get overloaded at initial release and these are no exception.

For the sake of recording this weekend, I hope that they are able to get their services up and running at a level that can support the massive amount of users.  If not, I may be out of luck in terms of live video streaming.

Regardless of how it goes in the next few days, I will be testing this technology further as time goes by (and hopefully the initial surge of novelty users lose interest).

If any of you want to follow my videos, here are my current profile addresses:

Qik (http://www.qik.com/quick)

Flixwagon (http://www.flixwagon.com/quickflix)

UPDATE: Oddly enough, the only thing that I have gotten to work so far today is to upload my video via Flixwagon on the (incredibly slow) AT&T EDGE network, although the application is prone to crashing or losing its connectivity during the process.  This is what I will most likely try to do for at least a song or two at the concert tomorrow night (assuming the services do not solve their server and wi-fi access problems by then).  It is at least worth noting that Flixwagon is capable of uploading videos over normal wireless phone networks, whereas Qik is not (or perhaps does not allow users to attempt such a feat).

UPDATE II: I have successfully gotten Flixwagon to work over wi-fi now, however Qik still is not placing my videos online, though data transfer appears to be happening.  It looks like Flixwagon will be my only shot at capturing some video this weekend.  It works very well/fast over wi-fi, but is (unsurprisingly) slow over the AT&T phone network.

Qik At Last On The iPhone (And Just In Time)

Qik is a video streaming service that allows mobile phone users to create and share live broadcasts.  I have been waiting a few months for this technology to release to the public, especially because I was hoping it would come out before my trip to Las Vegas.  Well, less than twenty four hours before my plane will arrive, Qik has been released for the iPhone.  With no time to experiment before the trip, I am not sure how well it will work (especially in regions lacking wi-fi access, which likely will include the concert hall), but I'm going to give it a shot.

I will be attempting to make a live broadcast from the Hard Rock Casino and Hotel in Las Vegas, which is hosting the Hall and Oates concert on Friday, August 15, 2008.  You are welcome to join me from my Qik profile at 8pm PST (11pm EST)!

I plan to throw a few test videos up before the big night, so if you happen to catch one of the early ones, feel free to send me a message about how well they came out, any problems you encountered, etc.  If we can get some early troubleshooting done, it will lead to better videos later on.

The thing that attracts me to live mobile phone streaming so much is the capability that we have to capture things that would otherwise be entirely forgotten.  Concerts, interviews, unexpected events, disasters, and moments with friends or family, are just a few of the things that can be remembered through this technology.

UPDATE:  I called the hotel and they said that they do not have wi-fi access available across the facility.  This most likely means that my Qik videos will not come through while in the concert.  However, I will still try to take video during the show and at any wi-fi points that I pass through while on the trip.  Also, I have the QuickVoice Recorder application downloaded, which allows my iPhone to create audio recordings that can later be transfered to the computer.  I will try to capture some songs and sounds with it.  Lastly, as always, I will attempt to sneak my photo camera in and snap some good pictures.  If I succeed at getting it inside, my photos will naturally be the highest quality output from the trip.

Why Time Does Not Exist

An interesting topic of conversation arose amongst friends in the past week. We asked the question, "does time exist?" In the moment, my initial instinct was to say, "no," however I did not have an on-the-spot explanation. Now that I have had a few days to think it over, I believe that I have come to a reasonable conclusion and explanation for the concept.

Why Time Does Not Exist

Imagine a universe where no change occurs. All entities are static and persist without any alterations. In such an environment, I believe that it would be impossible for any human to conceive the notion of time. This is so, because time is merely a contrived measurement. It was purposefully created to act as a simplified labeling system. In the aforementioned static universe, time cannot exist, for there is nothing that can be measured by it. What humans are really describing when they speak of time, is change. In the ordinary universe, change is so abundant as to be incomprehensible in terms of its individual events. Hence, the notion of time is a singular catchall that represents the innumerable occurrences that are perpetually taking place. It was invented for the purpose of comprehending, discussing, and manipulating change. Time itself is an empty container. It is not until the indistinguishable infinity of events are taking place that the concept of time is given meaning. Thus, contrary to common understanding, it is not time that causes change, but rather, continual change that allows for the perception of time.

An Example: Human Aging

The saying “age is just a number” or one very similar to it is often heard by those describing their ages or the differences between them. Ironically enough, such statements are correct in their marvelous simplicity. What the populous takes for granted as an elementary, all-inclusive, three-letter word, is actually the perpetuation of complex changes that take place in the body across its lifespan. It is the continual generation and deterioration of cells, creation and transmission of ideas, and ebb and flow of emotions, amongst countless additional occurrences, that constitute one’s age, not the mere passage of an arbitrary metric.

Implications For Manipulation

Considering that time does not exist in and of itself, the prospects for manipulating time are entirely absent from the universe. Anyone who might be seeking to command the notion that is commonly know as “time,” has two foreseeable options.

The first would be to alter how humans perceive time. To do so would require psychological methods. For example, it is a common experience amongst individuals involved in enjoyable activities, that time seems to pass quickly. The reverse is true for unenjoyable proceedings. That is, they are discerned to last much longer. Effects on the perception of time are unlikely to be physically significant, however may have many useful mental applications.

Since time is not an entity in its own right, but the catchall description of countless events that simultaneously occur, the second way that time could foreseeably be altered is to exploit the occasions that compose it. For instance, if human aging were attributed solely to cell growth and decay, for one to halt, reverse, or incite aging would require the ability to halt, reverse, or incite cell growth and decay. Alas, cell manipulation is no simple task. Furthermore, aging cannot be entirely pinned down to a single, few, or even comprehendible number of interactions. Likewise, a similar condition precludes the practical alteration of time in other cases. However, all options considered, this method is much more likely than the former to have tangible, physical effects. Yet, it is seemingly impossible in many instances, due to the overwhelming quantity of unknown variables involved.

Final Thought

At this juncture, it may be beneficial to reevaluate your conclusions on the concept of time. Ask yourself what happened while you read this paper. Did ten minutes simply pass by? Perhaps your eyes moved from left to right across the screen, sending signals to your brain, which were translated into meaning - all the while an infinite number of events simultaneously took place, both in your immediate surroundings and the universe. Maybe something else happened - possibly nothing. Whatever your take on time, feel free to share it with me and anyone else who may read this article by posting your comments.