DL2 vs. DL1

Matt,
You addressed the exact things I would want to know, thanks. Of course, I am concerned about a few things.

[quote=schiefels]
...DL.2 has a much more robust pan bearing...


As an eventual point of failure, I would anticipate a need to replace pan bearings when movement and indexing showed inconsistencies that were no longer within the user's acceptable tolerance.
Albeit replacing a bearing every couple or years is perhaps reasonable. I still remember the headaches of changing pan belts on Spot 250s simply because it was tedious.

[quote=schiefels]
...Operationally the big difference is the encoders...
...DL.1s use analogue...
...DL.2s use much higher resolution digital encoders...


Knowing how analogue encoders work, I know that their long-term ability to index to the same location is weak. In the application I am using, Pan will be the most critical indexing function simply because all of the installations are to use complex edge blending where the projection surface is close enough to the audiences where they will be sensitive to deviations.

[quote=schiefels]
This is not to say that the movement accuracy and repeatability on the DL.1s is bad, only that the DL.2s are better.


Yes, just from what you pointed out, the DL.2 would be better but only in a side by side comparison could we know how significant.

Just a thought...
I did a few calculations to evaluate possible tolerance.
At 45.9' of Throw, at a wide Zoom, the DLs have a projection width of 25.3'.
This is a beam angle of 15.4 degrees (arctan [(25.3/2)/45.9)] = 15.4).
Since the projection is 1024 pixels wide, this would yield an angle of 0.0150 degrees per pixel.
Of course since most of the Axon stock clips are DV NTSC, it means we're dealing with 720 x 480. Let's pretend that in scaling the image, the media server doesn't recognize anamorphic pixels and therefore we're talking about an image 640 x 480.
At a projection of 640 wide, this would yield and angle of 0.0241 degrees per pixel.

Pan is a 16 bit configuration, it means that at highest possible resolution, the DL.2 would have would be 65,536 steps. Since the Pan range is 400 degrees, this means that there are 163.84 possible steps per degree.
The reciprocal of that is 0.0061 which is about a forth of a pixel in accuracy using the 640 pixel or about half of a pixel accuracy at 1024.

Now if we consider the DL.1 and your guess on accuracy, we see a different outcome. We'll call the DL.1 6,600 steps which is 1 tenth that of the maximum resolution from 16 bit. Using the same equations, here are my results:
A 640 Pixel accuracy is about three pixels.
A 1024 Pixel accuracy is about four pixels.

On a specific application, I have throw distance widths of 13' 4".
Roughly translating, that means each 1024 pixel is about 1/8" So it means I'll have a possible slop of indexing equal to about 1/2 an inch.
My audience is sitting about 10' to 30' from the screen. So as you can imagine, this is noticeable, like reading 1/2" wide letter text from 10' to 30' and most people can easily do that without glasses.

It may be a mute point but I am hoping to demonstrate in my calculations that the possible difference between a DL.1 and a DL.2 is significant for those of us who intend to edge blend.

Here are my questions...
Can I order my DL.1s with the same digital encoders and pan bearings that come on a DL.2?
Or Can I order DL.2s without the Media Server at a reasonably proportionate cost to the DL.1?

I am eager to get feedback. Oh, and check my calculations. I've been know to make mistakes when crunching numbers off the top of my head.

Matt,
You addressed the exact things I would want to know, thanks. Of course, I am concerned about a few things.

[quote=schiefels]
...DL.2 has a much more robust pan bearing...


As an eventual point of failure, I would anticipate a need to replace pan bearings when movement and indexing showed inconsistencies that were no longer within the user's acceptable tolerance.
Albeit replacing a bearing every couple or years is perhaps reasonable. I still remember the headaches of changing pan belts on Spot 250s simply because it was tedious.

[quote=schiefels]
...Operationally the big difference is the encoders...
...DL.1s use analogue...
...DL.2s use much higher resolution digital encoders...


Knowing how analogue encoders work, I know that their long-term ability to index to the same location is weak. In the application I am using, Pan will be the most critical indexing function simply because all of the installations are to use complex edge blending where the projection surface is close enough to the audiences where they will be sensitive to deviations.

[quote=schiefels]
This is not to say that the movement accuracy and repeatability on the DL.1s is bad, only that the DL.2s are better.


Yes, just from what you pointed out, the DL.2 would be better but only in a side by side comparison could we know how significant.

Just a thought...
I did a few calculations to evaluate possible tolerance.
At 45.9' of Throw, at a wide Zoom, the DLs have a projection width of 25.3'.
This is a beam angle of 15.4 degrees (arctan [(25.3/2)/45.9)] = 15.4).
Since the projection is 1024 pixels wide, this would yield an angle of 0.0150 degrees per pixel.
Of course since most of the Axon stock clips are DV NTSC, it means we're dealing with 720 x 480. Let's pretend that in scaling the image, the media server doesn't recognize anamorphic pixels and therefore we're talking about an image 640 x 480.
At a projection of 640 wide, this would yield and angle of 0.0241 degrees per pixel.

Pan is a 16 bit configuration, it means that at highest possible resolution, the DL.2 would have would be 65,536 steps. Since the Pan range is 400 degrees, this means that there are 163.84 possible steps per degree.
The reciprocal of that is 0.0061 which is about a forth of a pixel in accuracy using the 640 pixel or about half of a pixel accuracy at 1024.

Now if we consider the DL.1 and your guess on accuracy, we see a different outcome. We'll call the DL.1 6,600 steps which is 1 tenth that of the maximum resolution from 16 bit. Using the same equations, here are my results:
A 640 Pixel accuracy is about three pixels.
A 1024 Pixel accuracy is about four pixels.

On a specific application, I have throw distance widths of 13' 4".
Roughly translating, that means each 1024 pixel is about 1/8" So it means I'll have a possible slop of indexing equal to about 1/2 an inch.
My audience is sitting about 10' to 30' from the screen. So as you can imagine, this is noticeable, like reading 1/2" wide letter text from 10' to 30' and most people can easily do that without glasses.

It may be a mute point but I am hoping to demonstrate in my calculations that the possible difference between a DL.1 and a DL.2 is significant for those of us who intend to edge blend.

Here are my questions...
Can I order my DL.1s with the same digital encoders and pan bearings that come on a DL.2?
Or Can I order DL.2s without the Media Server at a reasonably proportionate cost to the DL.1?

I am eager to get feedback. Oh, and check my calculations. I've been know to make mistakes when crunching numbers off the top of my head.