Linux-SWAT
Forum Addict!
- Joined
- Feb 13, 2010
- Messages
- 9,194
So maybe another chip should be used ?
How to get rotation working
- Thread starter EvilDragon
- Start date
VNC is actually pretty good now adays. I'd say it's as good as some commercial products like LogMeIn (LMI used to be way faster back in 2008ish timeframe) but either VNC has improved or net connections have improved enough to make it similar. It's not as good as X2Go (when that works well) or RDP, but there's little that's as good as those two IMHO.
J
JakDanyelz
Guest
I'm so happy to see how things are coming along. You guys are doing some really great work and I appreciate how you keep us all in the loop on your progress. What a great team!
directive0
Very Active Member
I wish all small scale tech projects had as much transparency and momentum as the Pyra.
Hello all,
I have a first batch Pandora and Im following the developement of the Pyra.
But at this thread I have a question.
Why does the Pyra need a rotator chip?
Who is holding a device like the Pandora or Pyra sideways or heads down?
I cant imagine that rotating the picture is something usefull. I would hate it if my Pandora would rotate the image just because i lie on my side.
Or am I completely wrong and a rotator chip does something other that rotating the picture?
Thanks
I have a first batch Pandora and Im following the developement of the Pyra.
But at this thread I have a question.
Why does the Pyra need a rotator chip?
Who is holding a device like the Pandora or Pyra sideways or heads down?
I cant imagine that rotating the picture is something usefull. I would hate it if my Pandora would rotate the image just because i lie on my side.
Or am I completely wrong and a rotator chip does something other that rotating the picture?
Thanks
slaeshjag
¯\_(ツ)_/¯
The display is a portrait one, because pretty much all displays are.
The rotator is needed so that you don't need to hold the Pyra sideways.
The rotator is needed so that you don't need to hold the Pyra sideways.
fusion_power
Advanced Member
Actualy, the chip is mainly needed so you are not depending on software-solutions (Android, iOS...) to rotate the screen. So it can be independent on which OS you have. 
The OMAP5 has a Vivante GC320 in it (http://www.vivantecorp.com/en/technology/composition.html). Does anyone know why we don't / can't use that to do the rotation? Would it use too much memory bandwidth or something?
slaeshjag
¯\_(ツ)_/¯
^ Actually, we most likely do. Unlike the SGX
Exophase
Nothing good will ever come of Exophase.
We do. AFAIK it wasn't clear that it'd be a realistic option back when the rotation chip was selected.
There are still some disadvantages, namely:
- Takes away bandwidth, about 420MB/s for full screen at 60Hz, so would have a bit of a performance hit
- Requires about 3.5MB of extra RAM to be allocated
- Takes away time where the 2D accelerator could be used for other thing like compositing windows or sprites in a 2D game
And there could be software complications getting it to work transparently. But it also doesn't add the compression artifacts that the rotation chip adds at full resolution. And I don't think the chip has yet been configured to work w/o tearing yet.
Don't know which solution uses more power.. the rotation chip has its own memory, presumably SRAM which could use less power than going out to main memory. But it has to perform the compression and decompression, and it has to recreate the MIPI signals on the other end.
WizardStan
Mega GP Mania
- Joined
- May 24, 2008
- Messages
- 16,731
I seem to recall that the rotator chip can be disabled and the LCD driven directly, and that it did not cost very much (like $5? I can't find a source on this so might be misremembering) so if a "better" solution presents itself there isn't really a major loss to having it.
Thats great for applications that use the Frame-buffer directly so you only have to swap x and y to get rotation for them.
I, too, have been pondering the rotator chip. Doesn't the display firmware govern the orientation of the screen? I'm guessing it can't ( easily ) be changed. I'm guessing also it's been mentioned somewhere and I should read the forums more
Strikes me that a dedicated rotator chip is something you'd use if you had multiple desired orientations ... and not something you'd use to fudge a video signal 90 degrees to suit display firmware written for mobile devices.
I have a dim awareness that display firmware is a massive pain though- so I'd understand if it couldn't be fixed. We have several sample displays and driver boards at Pim ( for Picade mostly ) many of which are curiously configured!
Strikes me that a dedicated rotator chip is something you'd use if you had multiple desired orientations ... and not something you'd use to fudge a video signal 90 degrees to suit display firmware written for mobile devices.
I have a dim awareness that display firmware is a massive pain though- so I'd understand if it couldn't be fixed. We have several sample displays and driver boards at Pim ( for Picade mostly ) many of which are curiously configured!
fusion_power
Advanced Member
"Compression Artifacts"? WTF never heared of that issue. Me don't want! If the rotator chip really have these disadvantages like tearing/artifacts at he end, then we may better drive with standard solutions. I can't imagine that basicly only switching X and Y really cost so much effort to rotate a screen nowadays.
levi
Still fresh, damnit!
The tearing is an issue with was being resolved, last time we heard - it needed a hardware and a software change to make it work, I think, so I've not heard positive news that it's been fixed yet, but I'm hopeful that we'll hear something soon.
In terms of the artefacts, some test images were sent through a rotation chip before it was selected, and we couldn't easily spot any changes it had made, so it was deemed to be acceptable.
In terms of the artefacts, some test images were sent through a rotation chip before it was selected, and we couldn't easily spot any changes it had made, so it was deemed to be acceptable.
- Joined
- Jan 18, 2010
- Messages
- 11,418
This was a known thing when this was first discussed... This thread I linked below has the reasons why ED and Nikolaus went with a rotator chip, also somewhere buried in that thread I recall ED sharing actual image diffs of source and post compression images.
http://boards.openpandora.org/topic/16040-beautiful-landscapes-are-prettier-than-portraits/
EDIT: Found the link to the compression comparisons: http://boards.openpandora.org/topic/16040-beautiful-landscapes-are-prettier-than-portraits/page-11#entry328594
Last edited by a moderator:
