Video Dimensions Changed

I have a recording of a Formula 1 race from FOX that was broadcast in 16x9. I am unsure if it was a true HD source but Windows Media Player identifies the .tp file as 1280x720. When I try to edit this file with VRD it hangs at a point where some in car footage has severe pixelization. VRD pops up a MovieWindow Problem dialog box with the message "! Video Thread: Video Dimensions changed."

When playing this video in Windows Media Player I get momentary picture freezes (sometimes several seconds long) where the pixelization occurs but it will eventually play through. In BeyondTV (the application that created it) it also plays through with fewer and shorter picture freezes.

I have tried to run QuickStream Fix but it hangs with the yellow triangle exclamation point symbol.

edit: I am running what appears to be the current version of VRD TVS
Last edited:
Additional attempts with this file have revealed different behaviors when the QuickStream output is set to .mpg or .ts file formats.

When set to .mpg I get the yellow triangle and VRD output hangs at 1,354,240 KB (the original file is a .tp and 13,262,056 KB). I waited over an hour before pressing the Abort button.

With output set to .ts the output file gets to 1,373,184 KB and then VRD crashes. It never displays the yellow triangle. Here is the Error Signature I was able to copy from the Error Reporting Dialog Box:

Error Signature

AppName: videoredo3.exe AppVer: ModName: videoredo3.exe
ModVer: Offset: 0004bfd0

I tried to save a copy of the dump file but I could not save it while the Error Reporting Dialog was open and it deleted it when it finished with no sign of it in the recycle bin. I was able to copy the additional text file.


Super Moderator
Try using "Enable Filters" in QuickStream Fix.

Process the file through QuickStream Fix first, then load the fixed file for editing. Tools>QuickStream Fix

Check Enable Filters in QuickStream Fix and select the resolution you desire to output. If multiple resolutions are not listed, increase the number of samples. Usually a number like 128 is sufficient.

Here is a guide to assist you

Also, please try the latest Beta available here:

Running QSF with Filters on and specifying 1280x720 resolution processed the entire file successfully. I was then able to edit and save it with VRD.


New member
Thanks for the info...I ran into the same problem with a video I'd recorded for my daughter. I'd tried the QSF, but didn't know to turn on the filters. I still couldn't edit the resulting mpg file. Turning on the filters, and setting the resolution fixed my problem...Thanks again.


New member
In my case VideoRedo crashes when I keep holding down Shift or Ctrl + Left Arrow keys and one the aformentioned message pops up.


New member
Yes, it worked after fix. But why is stream fixing not enabled automaticaly for any conversion? Does VR need to start fixing from the top to the end?


Staff member
But why is stream fixing not enabled automaticaly for any conversion? Does VR need to start fixing from the top to the end?
Sorry, don't understand your question. Can you elaborate?


New member
Well, let's say I have a HD recording where I've made selections to be cut. In the middle of conversion I get some of those errors and I have to start from the beggining: fix the stream then select the cuts again. Why not have an option to automaticaly fix the stream as needed the first time.


Staff member
Why not have an option to automaticaly fix the stream as needed the first time
Its currently a technical issue. Memory is allocated all over the place based on the video dimensions when the stream is opened. If the dimensions change it throws everything off. This was never an issue until the last year or so when users switched from analog capture to direct digital capture.

We are working on a solution as we know the problem is annoying. In fact we are trying to decide now what we should do with the "changed dimensions". Attempt to display the different sized video, or simply display a warning that the dimensions have changed in the video window and skip the alternate dimensions when outputting the video.

I would be interested in hearing opinions.


My druthers for the COM interface are:

As a (perhaps optional) feature of QSF:
1. Do a quick scan of the file at perhaps 10 points equally spaced throughout and determine the "dominant" dimensions, i.e., the ones that occur most frequently.
2. Set the filter dimensions to the dominant ones and perform the QSF.

A separate function that just does the quick scan and returns a list of filter dimensions and "hit counts" would be useful but I don't know how the list could be returned from a COM function -- perhaps as a string to be parsed? Or, write a text file called VideoDimensions.txt or whatever. For this function it would be nice to have an argument for the number of points to check in the file.

The current functionality to QSF with no filter or with user-specified filter dimensions should be retained.


New member
Well, why not just assume automaticaly that the "dimension change" is an error, ask the user if the dimensions have really changed or just fix and continue? I am pretty sure the dimensions don't change it's just that I live in an urban area with high buildings all around me and my TV reception often suffers from multipath signals problem creating all sorts of weird stream errors.

The same goes for the Audio/Video Ring Buffer Overflows. I do get these from time to time. It would be nice to prevent these or fix them on the fly.
Last edited:


New member
Hi, Oliver40,

> In the middle of conversion I get some of those errors and I have to start from the beggining: fix the stream then select the cuts again.

There is a work-around that can reduce (not eliminate) the effort: when the output fails because of ring buffer overflow or something, go to the file menu in the upper left of the window and "save project". After QSF, copy the fixed file to replace the original source file. Then open the project. The cuts and bookmark will be there, although you will have to check them all (use the "jump to next/previous cut" buttons) because they may need to be moved a few frames to compensate for whatever frames the QSF deleted.

> Why not have an option to automaticaly fix the stream as needed the first time.

Back in January I asked for this, and although they made the QSF more robust, they didn't put that into the main editing mode. This was explained by DanR (thread 5963):

>>>The reason we didn't enable it for normal editing is that the fix is in a very sensitive part of the code and there's a slight possibility it could have some severe side effects during normal editing.
>>>We just don't have the bandwidth right now to fully test it.


setting filter on from COM

Is there currently a way to turn on the filters when doing the QSF via COM?

In my processing scripts, I need to copy my captured video from my video computer to my editing computer. I always do this via a QSF run on the editing computer gong between the 2 computers. It doesn't matter to me if this takes a wee bit longer than a simple copy, since it's all scripted, and having a cleaned up video is a nice bonus. If turning on the filters option can clean up the video even more, then I'd want to do this all the time as well.

I think one question would be, what is the downside to doing this enable filters all the time? But in any respect, I would think this should be COM scriptable, if it isn't already. I do my qsf by these 2 statements, and am unaware of any other way to do this.

$vrd FileOpenBatch $file

$vrd FileSaveAs $fixfile


New member
This effort on QSF is very good news. I vote to have it automated as much as possible (no surprise there). If you recall, we also discussed invoking QSF automatically upon the creation of a file (.dvr-ms for me, as I record digitally with HDHomeRun). That would put the icing on the cake. You could just ask up front what the acceptable dimensions are; I know what they are for the channels I record. You could just discard anything else insofar as I am concerned, since they are only commercials.


New member
I don't know if this is related to the video dimension problem but I don't understand why the capture image is bigger than the stream video size ( see the attachaments). In the image capture dialog, full size reported is 853X480 while the video program information dialog shows video size of 704X480 which is the correct size.


Last edited:


Staff member
The captured frame image is adjusted to the correct aspect ratio, which in your case is 16 / 9.
Top Bottom