Millisecond Forums

+x

 

seandr - 1/12/2022

I'll add that we did A LOT of download optimization work in Inquisit 6. Download times, especially for large numbers of files, should be a fraction of what they were with Inquisit 5. You might consider upgrading, if that is an option.

-Sean

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

+x

 

Dave - 1/12/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

> However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant.

The way you've implemented the selection, every file has to be downloaded. If you can move the selection outside Inquisit, you can treat each version as its own condition and run only that condition based on the selection. Then only the files applicable to that version would have to be downloaded. I.e.

<batch cider>
/ subjects = (1 of 9)
/ groupassignment = groupnumber
/ file = "DART_cider.iqx"
</batch>

<batch beer>
/ subjects = (2 of 9)
/ groupassignment = groupnumber
/ file = "DART_beer.iqx"
</batch>

...

<batch whiskey>
/ subjects = (9 of 9)
/ groupassignment = groupnumber
/ file = "DART_premix_whiskey.iqx"
</batch>

with



where

https://mili2nd.co/cj5b?version=1 would launch the cider version.
https://mili2nd.co/cj5b?version=2 would launch the beer version, and so forth.

+x

 

rhollett - 1/13/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

+x

 

Dave - 1/12/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

> However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant.

The way you've implemented the selection, every file has to be downloaded. If you can move the selection outside Inquisit, you can treat each version as its own condition and run only that condition based on the selection. Then only the files applicable to that version would have to be downloaded. I.e.

<batch cider>
/ subjects = (1 of 9)
/ groupassignment = groupnumber
/ file = "DART_cider.iqx"
</batch>

<batch beer>
/ subjects = (2 of 9)
/ groupassignment = groupnumber
/ file = "DART_beer.iqx"
</batch>

...

<batch whiskey>
/ subjects = (9 of 9)
/ groupassignment = groupnumber
/ file = "DART_premix_whiskey.iqx"
</batch>

with



where

https://mili2nd.co/cj5b?version=1 would launch the cider version.
https://mili2nd.co/cj5b?version=2 would launch the beer version, and so forth.

Thanks for those suggestions Dave.

Participants will arrive at this experiment from Qualtrics (previously from Sona) and then back to Qualtrics/Sona for auto course credit allocation (passing the ID along the chain). I have found this to work well for other experiments.

I suspect I can handle the selection in Qualtrics using embedded data for the groupnumber and pass that along with their ID number to facilitate this batch option that you suggest?

Cheers,

Ross

+x

 

Dave - 1/13/2022

+x

 

rhollett - 1/13/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

+x

 

Dave - 1/12/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

> However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant.

The way you've implemented the selection, every file has to be downloaded. If you can move the selection outside Inquisit, you can treat each version as its own condition and run only that condition based on the selection. Then only the files applicable to that version would have to be downloaded. I.e.

<batch cider>
/ subjects = (1 of 9)
/ groupassignment = groupnumber
/ file = "DART_cider.iqx"
</batch>

<batch beer>
/ subjects = (2 of 9)
/ groupassignment = groupnumber
/ file = "DART_beer.iqx"
</batch>

...

<batch whiskey>
/ subjects = (9 of 9)
/ groupassignment = groupnumber
/ file = "DART_premix_whiskey.iqx"
</batch>

with



where

https://mili2nd.co/cj5b?version=1 would launch the cider version.
https://mili2nd.co/cj5b?version=2 would launch the beer version, and so forth.

Thanks for those suggestions Dave.

Participants will arrive at this experiment from Qualtrics (previously from Sona) and then back to Qualtrics/Sona for auto course credit allocation (passing the ID along the chain). I have found this to work well for other experiments.

I suspect I can handle the selection in Qualtrics using embedded data for the groupnumber and pass that along with their ID number to facilitate this batch option that you suggest?

Cheers,

Ross

> I suspect I can handle the selection in Qualtrics using embedded data for the groupnumber and pass that along with their ID number to facilitate this batch option that you suggest?

Yes, absolutely. Since your participants start off in Qualtrics anyway, that is the ideal solution.

+x

 

Dave - 1/13/2022

+x

 

rhollett - 1/13/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

+x

 

Dave - 1/12/2022

+x

 

rhollett - 1/12/2022

Hi,

I have a very image-heavy experiment with over 1200 files uploaded to my Inquisit web license in total. However, not all the images are used for each administration. I use a selection script and a batch file to allow participants to select their preferred version of the experiment, each of which involve less than 150 jpg images (around 40KB each - or around 6MB). However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant. I think the image files are already sufficiently compressed but the download time in testing seems ridiculous given my internet speed (literally waiting 10-15minutes in Chrome).

Is there any way I can improve the efficiency of this delivery? Obviously I cannot deploy this experiment the way it is and using multiple web licenses seems inefficient (and expensive) given how many versions there are. 

Thanks in advance.

> However, the load time is insanely slow, probably because the player is downloading every single file in the web experiment, rather than just the one's needed for the version selected by the participant.

The way you've implemented the selection, every file has to be downloaded. If you can move the selection outside Inquisit, you can treat each version as its own condition and run only that condition based on the selection. Then only the files applicable to that version would have to be downloaded. I.e.

<batch cider>
/ subjects = (1 of 9)
/ groupassignment = groupnumber
/ file = "DART_cider.iqx"
</batch>

<batch beer>
/ subjects = (2 of 9)
/ groupassignment = groupnumber
/ file = "DART_beer.iqx"
</batch>

...

<batch whiskey>
/ subjects = (9 of 9)
/ groupassignment = groupnumber
/ file = "DART_premix_whiskey.iqx"
</batch>

with



where

https://mili2nd.co/cj5b?version=1 would launch the cider version.
https://mili2nd.co/cj5b?version=2 would launch the beer version, and so forth.

Thanks for those suggestions Dave.

Participants will arrive at this experiment from Qualtrics (previously from Sona) and then back to Qualtrics/Sona for auto course credit allocation (passing the ID along the chain). I have found this to work well for other experiments.

I suspect I can handle the selection in Qualtrics using embedded data for the groupnumber and pass that along with their ID number to facilitate this batch option that you suggest?

Cheers,

Ross

> I suspect I can handle the selection in Qualtrics using embedded data for the groupnumber and pass that along with their ID number to facilitate this batch option that you suggest?

Yes, absolutely. Since your participants start off in Qualtrics anyway, that is the ideal solution.

+x

 

rhollett - 1/14/2022

Hi Dave,

OK so I have shifted the selection into qualtrics and this appears to be working fine using the following URL to pass the subjectid and groupnumber (instead of "version") from Qualtrics. Both these data are also correctly recorded in the Inquisit web files afterwards.

mili2nd.co/cj5b?groupnumber=${gr://SC_9M3MPkfzJ5g9f02/Score}&subjectid=${e://Field/subjectid}

However, the Inquisit web player still seems to be downloading excessive files (somewhere around 600, according to the loading window) before starting up.

Each "file" (script/version) only requires about 140 files to run so it is still appears to be downloading a lot of files before starting up.

I am wondering if it is still running through multiple batches before starting up? Perhaps my batch script is not set up correctly?

+x

 

Dave - 1/14/2022

+x

 

rhollett - 1/14/2022

Hi Dave,

OK so I have shifted the selection into qualtrics and this appears to be working fine using the following URL to pass the subjectid and groupnumber (instead of "version") from Qualtrics. Both these data are also correctly recorded in the Inquisit web files afterwards.

mili2nd.co/cj5b?groupnumber=${gr://SC_9M3MPkfzJ5g9f02/Score}&subjectid=${e://Field/subjectid}

However, the Inquisit web player still seems to be downloading excessive files (somewhere around 600, according to the loading window) before starting up.

Each "file" (script/version) only requires about 140 files to run so it is still appears to be downloading a lot of files before starting up.

I am wondering if it is still running through multiple batches before starting up? Perhaps my batch script is not set up correctly?

> However, the Inquisit web player still seems to be downloading excessive files (somewhere around 600, according to the loading window) before starting up.

The number you see in the loading window is not the number of downloaded files. Rather, it's something akin to number of preparatory steps taken, which includes -- among other things -- parsing the elements, downloading the files, processing the files into memory, and so forth. If you set things up as sketched out at https://www.millisecond.com/forums/FindPost33664.aspx , then only the files pertaining to the applicable condition are downloaded and processed.

+x

 

seandr - 1/17/2022

Interesting, I'd expect at least a modest improvement from reducing file sizes. Inquisit 6 is definitely much faster with experiments involving large numbers of files. If it would help, I can set you up with a trial license for Inquisit Web so you can test it out. Email sales@millisecond.com if you'd like to try that.

-Sean