Not impressed with MTF program line...

**MJSfoto1956** · 01-25-2014, 02:25 PM

Just for kicks, I wanted to see how "smart" the K-3 algorithms are for different program lines.

Now, as a software architect, I'm not a fan of "hard and fast" rules -- I like software that is "flexible". (better yet, I like software that is user-configurable!!)

Given that, my expectation was that MTF *should* mean: bias the settings towards obtaining the most sharpness out of the current lens but also taking into account all the other factors that affect "sharpness". In other words, I expected it to behave analogously to the way that "Depth of field priority" biases the lens to open the lens up, but doesn't blindly open it up to the maximum and keep it there always.

Instead, what the MTF program line appears to do is chose the lens' known ideal f/stop then never deviate from that setting. I find this nearly useless (and not in keeping with the "priority" behavior of other program lines), especially on a prime lens. I would have expected the algorithm to attempt to keep as "close" as it could to the best MTF but allow it to move a little in either direction depending on the current situation. E.g. if the best MTF for a given lens is at f/2.8 then allow the MTF program line to vary (as in a bell curve) between f/2 and f/4. After all, what's the purpose of blindly sticking to f/2.8 but then auto specifying ISO3200? Clearly a "sharper" image would be had at f/2 and ISO1600 in the same situation, but the way Pentax implemented MTF doesn't appear to work that way. It basically is no different than using Av mode with the "A" part automagically chosen for you -- that's nice, but it simply doesn't go far enough. Clearly I can see this as a convenience feature aimed primarliy at non-professionals who may not really understand how their complex lens behaves at different focal lengths, and especially designed with super zooms in mind, but given the way it currently is implemented, I seriously doubt I could ever imagine using it. I would much rather have something that was completely manual versus an algorithm that I feel is patently flawed. I would rate this feature as "close but no cigar". Ah well. On to the next test...

YMMV

Michael

Adam · 01-25-2014, 04:10 PM

I think its been like that since the mtf program line came out many years ago. The info comes from the lens itself and the camera just rolls with that.

I agree that a "smart" program line would be nice. Though that's sort of what autopict does.

Teal · 01-25-2014, 04:35 PM

I'm actually keen to know how many people use MTF and has it worked for them?

SpecialK · 01-25-2014, 05:19 PM

Originally posted by MJSfoto1956

...also taking into account all the other factors that affect "sharpness".

Nope, it reads the chip in the lens which gives the "general" best setting (determined by engineers) for that type of lens - but not even that particular lens. That is all. Never used it :-) If I'm trying to get there (rarely) I just stop down 2 stops and go with it. I'm not a infinite pixel peeper :-)

**Ex Finn.** · 01-25-2014, 05:27 PM

Originally posted by Teal

I'm actually keen to know how many people use MTF and has it worked for them?

I have been running in modulation transfer function for several years. Seems to work fine for my shooting needs.
Granted, there were couple times I should have intervened and opened up the aperture, for desired effect.

**MJSfoto1956** · 01-25-2014, 06:54 PM

Ok, I did a little bit more digging and I was wrong in my analysis. The MTF program *DOES* open up/stop down above and below the "optimal" aperture for the given lens, but it seems to be somewhat "sticky" -- you really need to have extreme low-key or high-key situation to force the program to consider anything other than the "optimal" aperture. So in that way, it does behave somewhat analogous to the other modes, with the proviso that it "appears" to be more biased than the other modes. It would be awesome if they would publish the lookup tables they use the program lines use to make the decisions. It would be even better if we could "tweak" these lookup tables to our own purposes. Suffice it to say I take back my initial criticism and now have a more favorable view of what they have implemented. Still not sure I will use it though. Need more testing. Hopefully, Pentax engineers will agree that letting end users customize their systems is in everyone's best interest.

Cheers,
Michael

**BWG** · 01-25-2014, 09:08 PM

I like it for walking around. I leave ISO on auto with the range appropriate for the lighting conditions. It usually chooses an appropriate aperture and if I want something different I choose something different. The Pentax program line really is an excellent feature.

sterretje · 01-25-2014, 10:39 PM

Originally posted by MJSfoto1956

Now, as a software architect, I'm not a fan of "hard and fast" rules -- I like software that is "flexible". (better yet, I like software that is user-configurable!!)

Nice to see a colleague in the software field with a similar mindset to mine. There are definitely not enough of us.

Originally posted by MJSfoto1956

Given that, my expectation was that MTF *should* mean: bias the settings towards obtaining the most sharpness out of the current lens but also taking into account all the other factors that affect "sharpness".

Wrong expectation

My thought is that DOF is perceived sharpness, not absolute sharpness.

**MJSfoto1956** · 01-26-2014, 04:56 AM

I like DxO's concept of "visual acuity" better than a single mathematical MTF "sharpness" value. Acuity in this sense takes into account "the system": i.e. noise, sharpness, vignetting, etc. Consequently, if the "MTF" program line meant instead "try to obtain the best acuity out of the system" instead of "use the single sharpest aperture setting for all photos unless there are no other options", then I would be much happier with it.

Michael

stormtech · 01-26-2014, 08:26 AM

I use the MTF program line on my K-5IIs for a certain reason.

I have my camera with my DA*300 mounted and always at the ready on the top of my book shelf. We always get wildlife coming by the house. There have been many times where we only had a couple seconds to grab the camera and shoot before said critter disappears. With the MTF program set, I can be partially confident that I might get a shot. If the time allows, it is so easy just to flick the control wheels to change the aperture or shutter speed setting. Also, with my wife who shares my hobby but is not interested with anything to do with camera settings, she also can just grab the camera and shoot without having to ask me to change any settings.

So for my kind of shooting, the MTF program is just a basis for a quick point and shoot and so easy to change with the flick of a control wheel.

FOX · 03-10-2014, 07:24 AM

thanks Michael and all for the interesting details. Learning a lot here

jon404 · 03-10-2014, 11:45 PM

Found that on my f/2.8 200mm lens, that it set the aperture at f/4. But this lens is sharpest at f/5.6 to f/8... and I use f/8 a lot for depth-of-field. In practice, I don't think MTF matters much.

asp1880 · 03-11-2014, 03:30 AM

Originally posted by MJSfoto1956

It would be awesome if they would publish the lookup tables they use the program lines use to make the decisions.

It's tempting to think of a program line as a lookup table. But it's more like a function of a ton of inputs, yielding an output triplet of (shutter, aperture, iso).

Inputs beside the metering reading is also stuff like shutter speed limitations (flash sync etc), iso limitations (autoiso, highlight protection), variable aperture zoom settings and probably tons more.

It's quite complex in its full level of detail. That's probably why you can't just program a simple table with one line per light level.

Regards,
--Anders.

bdery · 03-11-2014, 12:33 PM

Originally posted by MJSfoto1956

Instead, what the MTF program line appears to do is chose the lens' known ideal f/stop then never deviate from that setting. I find this nearly useless (and not in keeping with the "priority" behavior of other program lines), especially on a prime lens. I would have expected the algorithm to attempt to keep as "close" as it could to the best MTF but allow it to move a little in either direction depending on the current situation. E.g. if the best MTF for a given lens is at f/2.8 then allow the MTF program line to vary (as in a bell curve) between f/2 and f/4. After all, what's the purpose of blindly sticking to f/2.8 but then auto specifying ISO3200?

If you don't use auto ISO, I believe it actually works better. I find that it tries to stick to around 1/50, 1/60 speed and if it can, then it goes to the sharpest aperture. Otherwise, it will open up the aperture. This way it never closes it down, but it might open it.

I confess that I trust it only so far, I use P mode as a "av on demand" and frequently change the aperture value.

Originally posted by Teal

I'm actually keen to know how many people use MTF and has it worked for them?

I do, always.

01-25-2014, 02:25 PM	#1
MJSfoto1956 Site Supporter Join Date: Dec 2013 Location: Boston Posts: 1,305	Not impressed with MTF program line... Just for kicks, I wanted to see how "smart" the K-3 algorithms are for different program lines. Now, as a software architect, I'm not a fan of "hard and fast" rules -- I like software that is "flexible". (better yet, I like software that is user-configurable!!) Given that, my expectation was that MTF should mean: bias the settings towards obtaining the most sharpness out of the current lens but also taking into account all the other factors that affect "sharpness". In other words, I expected it to behave analogously to the way that "Depth of field priority" biases the lens to open the lens up, but doesn't blindly open it up to the maximum and keep it there always. Instead, what the MTF program line appears to do is chose the lens' known ideal f/stop then never deviate from that setting. I find this nearly useless (and not in keeping with the "priority" behavior of other program lines), especially on a prime lens. I would have expected the algorithm to attempt to keep as "close" as it could to the best MTF but allow it to move a little in either direction depending on the current situation. E.g. if the best MTF for a given lens is at f/2.8 then allow the MTF program line to vary (as in a bell curve) between f/2 and f/4. After all, what's the purpose of blindly sticking to f/2.8 but then auto specifying ISO3200? Clearly a "sharper" image would be had at f/2 and ISO1600 in the same situation, but the way Pentax implemented MTF doesn't appear to work that way. It basically is no different than using Av mode with the "A" part automagically chosen for you -- that's nice, but it simply doesn't go far enough. Clearly I can see this as a convenience feature aimed primarliy at non-professionals who may not really understand how their complex lens behaves at different focal lengths, and especially designed with super zooms in mind, but given the way it currently is implemented, I seriously doubt I could ever imagine using it. I would much rather have something that was completely manual versus an algorithm that I feel is patently flawed. I would rate this feature as "close but no cigar". Ah well. On to the next test... YMMV Michael

01-25-2014, 04:10 PM	#2
Adam Administrator Site Webmaster Join Date: Sep 2006 Location: Arizona Photos: Gallery \| Albums Posts: 51,594	I think its been like that since the mtf program line came out many years ago. The info comes from the lens itself and the camera just rolls with that. I agree that a "smart" program line would be nice. Though that's sort of what autopict does. Adam PentaxForums.com Webmaster (Site Usage Guide \| Site Help \| My Photography) PentaxForums.com server and development costs are user-supported. You can help cover these costs by donating or purchasing one of our Pentax eBooks. Or, buy your photo gear from our affiliates, Adorama, B&H Photo, KEH, or Topaz Labs, and get FREE Marketplace access - click here to see how! Trusted Pentax retailers:

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Program line: "MTF"?	WMBP	Photographic Technique	37	11-04-2015 03:58 PM
MTF program mode, w00t !	spystyle	Photographic Technique	7	07-16-2013 02:53 PM
Not impressed k5 pics not sharp	harleynitelite	Pentax K-5 & K-5 II	49	09-08-2012 12:55 PM
MTF program line overexposes?	Adam	Pentax DSLR Discussion	12	06-09-2010 12:49 PM
Not impressed with wide angle lens	molko	Pentax SLR Lens Discussion	31	05-17-2010 08:25 AM

01-25-2014, 04:35 PM	#3
Teal Forum Member Join Date: Dec 2007 Location: Sydney Posts: 55	I'm actually keen to know how many people use MTF and has it worked for them?

01-25-2014, 05:19 PM	#4
SpecialK Pentaxian Join Date: Dec 2006 Location: So California Photos: Gallery Posts: 16,480	Originally posted by MJSfoto1956 ...also taking into account all the other factors that affect "sharpness". Nope, it reads the chip in the lens which gives the "general" best setting (determined by engineers) for that type of lens - but not even that particular lens. That is all. Never used it :-) If I'm trying to get there (rarely) I just stop down 2 stops and go with it. I'm not a infinite pixel peeper :-)

01-25-2014, 05:27 PM	#5
Ex Finn. Site Supporter Join Date: Nov 2007 Location: Southern Maryland. Espoo. Kouvola. Photos: Gallery \| Albums Posts: 7,975	Originally posted by Teal I'm actually keen to know how many people use MTF and has it worked for them? I have been running in modulation transfer function for several years. Seems to work fine for my shooting needs. Granted, there were couple times I should have intervened and opened up the aperture, for desired effect.

01-25-2014, 06:54 PM	#6
MJSfoto1956 Site Supporter Join Date: Dec 2013 Location: Boston Posts: 1,305 Original Poster	Ok, I did a little bit more digging and I was wrong in my analysis. The MTF program DOES open up/stop down above and below the "optimal" aperture for the given lens, but it seems to be somewhat "sticky" -- you really need to have extreme low-key or high-key situation to force the program to consider anything other than the "optimal" aperture. So in that way, it does behave somewhat analogous to the other modes, with the proviso that it "appears" to be more biased than the other modes. It would be awesome if they would publish the lookup tables they use the program lines use to make the decisions. It would be even better if we could "tweak" these lookup tables to our own purposes. Suffice it to say I take back my initial criticism and now have a more favorable view of what they have implemented. Still not sure I will use it though. Need more testing. Hopefully, Pentax engineers will agree that letting end users customize their systems is in everyone's best interest. Cheers, Michael

01-25-2014, 09:08 PM	#7
BWG Loyal Site Supporter Join Date: Feb 2007 Location: Monterey County, California Posts: 165	I like it for walking around. I leave ISO on auto with the range appropriate for the lighting conditions. It usually chooses an appropriate aperture and if I want something different I choose something different. The Pentax program line really is an excellent feature.

01-25-2014, 10:39 PM	#8
sterretje Veteran Member Join Date: Dec 2008 Location: Roodepoort, South Africa Photos: Gallery \| Albums Posts: 3,561	Originally posted by MJSfoto1956 Now, as a software architect, I'm not a fan of "hard and fast" rules -- I like software that is "flexible". (better yet, I like software that is user-configurable!!) Nice to see a colleague in the software field with a similar mindset to mine. There are definitely not enough of us. Originally posted by MJSfoto1956 Given that, my expectation was that MTF should mean: bias the settings towards obtaining the most sharpness out of the current lens but also taking into account all the other factors that affect "sharpness". Wrong expectation My thought is that DOF is perceived sharpness, not absolute sharpness.

01-26-2014, 04:56 AM	#9
MJSfoto1956 Site Supporter Join Date: Dec 2013 Location: Boston Posts: 1,305 Original Poster	I like DxO's concept of "visual acuity" better than a single mathematical MTF "sharpness" value. Acuity in this sense takes into account "the system": i.e. noise, sharpness, vignetting, etc. Consequently, if the "MTF" program line meant instead "try to obtain the best acuity out of the system" instead of "use the single sharpest aperture setting for all photos unless there are no other options", then I would be much happier with it. Michael

01-26-2014, 08:26 AM	#10
stormtech Veteran Member Join Date: Jan 2011 Location: In the boonies (NW Penna) Photos: Gallery \| Albums Posts: 5,744	I use the MTF program line on my K-5IIs for a certain reason. I have my camera with my DA*300 mounted and always at the ready on the top of my book shelf. We always get wildlife coming by the house. There have been many times where we only had a couple seconds to grab the camera and shoot before said critter disappears. With the MTF program set, I can be partially confident that I might get a shot. If the time allows, it is so easy just to flick the control wheels to change the aperture or shutter speed setting. Also, with my wife who shares my hobby but is not interested with anything to do with camera settings, she also can just grab the camera and shoot without having to ask me to change any settings. So for my kind of shooting, the MTF program is just a basis for a quick point and shoot and so easy to change with the flick of a control wheel.

03-10-2014, 07:24 AM	#11
FOX Senior Member Join Date: Jan 2013 Location: Ottawa Photos: Gallery Posts: 132	thanks Michael and all for the interesting details. Learning a lot here

03-10-2014, 11:45 PM	#12
jon404 Veteran Member Join Date: Dec 2012 Photos: Gallery Posts: 696	Found that on my f/2.8 200mm lens, that it set the aperture at f/4. But this lens is sharpest at f/5.6 to f/8... and I use f/8 a lot for depth-of-field. In practice, I don't think MTF matters much.

Thread Tools	Search this Thread
	Search this Thread: Advanced Search

03-11-2014, 03:30 AM	#13
asp1880 Veteran Member Join Date: Jul 2010 Location: Lyngby, Copenhagen Photos: Albums Posts: 742	Originally posted by MJSfoto1956 It would be awesome if they would publish the lookup tables they use the program lines use to make the decisions. It's tempting to think of a program line as a lookup table. But it's more like a function of a ton of inputs, yielding an output triplet of (shutter, aperture, iso). Inputs beside the metering reading is also stuff like shutter speed limitations (flash sync etc), iso limitations (autoiso, highlight protection), variable aperture zoom settings and probably tons more. It's quite complex in its full level of detail. That's probably why you can't just program a simple table with one line per light level. Regards, --Anders.

03-11-2014, 12:33 PM	#14
bdery Pentaxian Join Date: Mar 2008 Location: Quebec city, Canada Photos: Gallery Posts: 9,357	Originally posted by MJSfoto1956 Instead, what the MTF program line appears to do is chose the lens' known ideal f/stop then never deviate from that setting. I find this nearly useless (and not in keeping with the "priority" behavior of other program lines), especially on a prime lens. I would have expected the algorithm to attempt to keep as "close" as it could to the best MTF but allow it to move a little in either direction depending on the current situation. E.g. if the best MTF for a given lens is at f/2.8 then allow the MTF program line to vary (as in a bell curve) between f/2 and f/4. After all, what's the purpose of blindly sticking to f/2.8 but then auto specifying ISO3200? If you don't use auto ISO, I believe it actually works better. I find that it tries to stick to around 1/50, 1/60 speed and if it can, then it goes to the sharpest aperture. Otherwise, it will open up the aperture. This way it never closes it down, but it might open it. I confess that I trust it only so far, I use P mode as a "av on demand" and frequently change the aperture value. Originally posted by Teal I'm actually keen to know how many people use MTF and has it worked for them? I do, always.