Shadow detail/gamma enhancement for the analog sources - Page 2  

Ignorant question: Why does device do a better job of lower ire gamma adjustment than a HTPC video card? I ask because this is great if it works and I want to build one!

TSE, maybe you could offer a kit. That way you could make some well earned money for all your work.

Clarence,

Your post makes perfect sense to me!!!

YEAH RIGHT!!!

What the heck is all that stuff?

Chris, you're just suffering from radiation poisoning. :p

I'm not sure it DOES do a better job than an HTPC. But for those of us who don't HAVE an HTPC, this is an attractive option!

Actually techman,

judging by Chris' informed posts in the Digital Hi-End Projectors Forum, he's far from a digital virgin. I'll have to footnote the amnesiac effects in my thesis. :D :D

John

Hope I don't sound like a r@sshole, but might I suggest keeping the name calling confined to the other 2 gamma threads that spiraled off topic.

It looks like TSE did a great job with these circuits and I'd hate to see a third thread ruined.

I'll second that... I have the wherewithall to build these beauties and want to see this tse design to conclusion (or at least gerber files!)

Phil,
This circuit is a gamma adjustment circuit, it adjusts a part of the gamma curve. HTPCs adjust the whole gamma curve, although I'm sure algorythms like Rec.709 could be easily implimented in video card drivers to tweak the gamma curve too.

Clarence,
Just stirring, but your point is taken.

Cheers,

John

Thanks John. I like slinging OT jabs too, but I didn't want tse to get frustrated by the heckling.

Would you care to share more info on the new Moome card?

Thanks!!

Yeah, lots of parts... expecially when you list them separately for the RG&B circuits.

Can you verify that C17 is 100nF?
Clamp C16 100pF
Clamp C17 100nF
Clamp C18 100pF

Also, for what it's worth, I tried to get a WAG on the component costs
(not including the resistors and capacitors... are there any listed that would cost more than about a dime?)
Clamp U9A 74HC4538 $1.17 http://www.newark.com/NewarkWebComme...KU=38C1094&N=0
Clamp U9B 74HC4538 $1.17 http://www.newark.com/NewarkWebComme...KU=38C1094&N=0
Misc U3B/C/D 74HC86 $1.10 http://www.newark.com/NewarkWebComme...KU=38C1097&N=0
Clamp R14 2.2K $0.49 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Sync U3A 74HC86 $1.10 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Misc D3 1N4002 $0.03 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Clamp R21 50K variable $2.55 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_R U2 HFA1100 $3.45 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_G U2 HFA1100 $3.45 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_B U2 HFA1100 $3.45 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Sync U6 LM311 $2.60 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Misc F1 RXE040 $0.24 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_R D1 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_G D1 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_G D2 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_B D1 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
VID_B D2 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Sync D7 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
Sync D8 SD101A $0.05 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
SH_R U7A/B TL082 $1.50 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
SH_G U7A/B TL082 $1.50 http://www.newark.com/NewarkWebComme...&Ntk=gensearch
SH_B U7A/B TL082 $1.50 http://www.newark.com/NewarkWebComme...&Ntk=gensearch

I've taken the liberty of uploading these files onto the freedfm service at 4pcb... I should have a response whether their automated checker spots any troubles shortly...

--Edit: There it is:
--Edit2: The autochecker found and fixed silkscreen line width problems (they require 7 mil line widths) which may result in less readability.

 

TSE-001 Gamma Correction Circuit.pdf 28.1435546875k . file

Let me check with PCB Express... the $51/3 deal may apply with this board size, 2 layers.

The 4PCB boards arrived for my work project... They look good (no solder mask, which makes the silkscreen hard to read...)

At lunchtime I'll look into getting pcbexpress to quote what I've taken to calling the TSE-001 Gamma Circuit.

The ExpressPCB "ProtoPro" service may prove a better price/performance point... four production-style (soldermask and silkscreen included) boards for $179USD.
http://www.expresspcb.com/ExpressPCB...csProtoPro.htm
Sounds like cleaner, prettier boards, with a two day turnaround.

OK, ExpressPCB (using their focmula and volume discount structure from their web site) would be $8.80USD per board plus shipping for 100 pieces with a 10 day turnaround ("Production" boards.)
Quantities from 50-99 would be $9.21 each, and from 2-49 would be $14.76.. this is far better than 4PCBs' $23-$42 each in 5-10 quantities.

This board is too big for the as-advertised 3 for $51.00 price.

If anyone just wants a quick peek at some of the Gerber files:

http://img361.imageshack.us/img361/1369/tsepcb3hu.gif

http://img361.imageshack.us/img361/5452/tsegreen0io.gif

ExpressPCB requires the board layout be done in their software, which has no import/export functionality. Bizarre, and business limiting. Oh well.

Back to 4PCB... They called me for a follow-up to determine if an order would be forthcoming... I mentioned that a short run of prototypes would be first with a more quantities later.

Sales0chick mentioned two specials they are running (I get the impression that there is always something special going on here...)
1) 3 for 33 special... $99USD for three two-layer boards with soldermask and silkscreening.
2) 2 for 4 special... Four week price on two week turnaround.

The 3 for 33 each pricing is quite a bit better price and appearance than any of the proto costs I've seen thusfar.

How can I do something like this with my HTPC? I use FFDShow for gamma correcton right now, but it adjusts the whole gamma curve and not just part of it, correct? When I adjust it to get acceptable shadow detail I no longer get absolute black during a fade to black.

What can I get to help my HTPC do what this thing that TSE has made does? Certain video card? Modded video card? What?

I woud like to have absolute black again if I could, but right now the cost is just too much (black crush).

Thanks

1) Avoid using the FFDshow gamma adjustments, I believe it only operates in 8-bit. This will introduce banding problems.

2) depending on your video card, you should use that. Nvidia has good simple gamma adjustment, as well as I believe many have more advanced adjustments allowing you to make spaghetti out of your gamma curve if you so desire! It also makes these adjustments at I believe 10 bits at least, so this helps avoid banding.

When you are adjusting gamma, be sure to do it sparingly, and view bright scenes as well. Many times people will adjust gamma while looking at dark scenes, and keep seeing more and more shadow detail and raise their gamma curve far too much. What you are doing is crushing the contrast in the bright portions of the image and your depth will suffer and things will look postery and flat in bright scenes. The better way to do it is to move back and forth between a dark scene and a bright scene and find a good interplay between shadow detail and depth/contrast in bright scenes. Also, if you figure your CRT has a gamma of close to 2.5, you can just shoot for a target gamma maybe .2 or .3 less than that. If you have a non-LC machine, you will want to have a slightly lower final gamma number than an LC machine, because of the lower ANSI contrast (this means raising the gamma curve in the electronics more). An LC machine will maintain shadow details better, so there is less of a need to try to elicit low-end detail, and so you can be more conservative with your gamma adjustments.

G'day Folks,

following TSE's initial posting on using electronics to replace the typical video card gamma correction function, it got me thinking that a mate of mine who heads a local electronics design company here in Melbourne might be able to knock up something similar. I added a couple of 'wish list' items to the original concept and yesterday he showed me what has been working on for the last couple of weeks in his spare time.

It uses the same Intersil HFA1100's at its core that TSE uses in his design (he researched alternates but concluded that those OP-amps were the best chips appropriate), and has added functionality for variable gain, response curve and setup level. We set the prototypes up feeding a NEC 9PG Xtra that I maintain and the results were very promising. After a short time we were able to pretty well replicate the gamma function of the ATI Radeon 9600 card that normally feeds the projector. It was pleasing to note that the posterization and background noise that normally accompanies using high gamma values in the video card were reduced using the electronics solution. My mate is still fine tuning some of the component values on the board but it's looking like a viable alternative for those folks who don't use a HPTC to drive their projectors.

When I next catch up with him I'll try and remember to take some piccies of the cro traces showing the difference between input and output, but in the meantime here's a couple of piccies of the prototypes sprawled across his "testbench"

Cheers :)

Russ

http://www.users.bigpond.net.au/benn..._proto_01b.jpg
http://www.users.bigpond.net.au/benn..._proto_02b.jpg