Where did you get this from?
That's not quite my understanding of how HTP works - can you provide some more information and some links to where you are sourcing the information from please.
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
Keep the noise low : HTP and ALO "ON" or "OFF"
- Thread starter Ak pinxit
- Start date
That's not quite my understanding of how HTP works - can you provide some more information and some links to where you are sourcing the information from please.
birdbrain
Senior Member
Canon's own blurb does mention that using HTP may result in more shadow noise, so it sounds like they apply a 'tone curve', I think that's the right term, to then lift the shadows? Not sure about the dual ISO bit though.
--
Phil
I wondered why the ball kept getting bigger, then it hit me.
http://www.flickr.com/photos/philthebirdbrain/
--
Phil
I wondered why the ball kept getting bigger, then it hit me.
http://www.flickr.com/photos/philthebirdbrain/
Yep - that's my understanding too.
I thought it was simply, in effect, underexposed by one stop and then a different tone curve applied.
Not sure where TTMartin is getting his info from, or if he's just misinformed.
mojorisn
Leading Member
Chuck Westfall, from Canon, 2009 article on HTP, ALO and RAW (and some other topics)
http://digitaljournalist.org/issue0902/tech-tips.html
http://digitaljournalist.org/issue0902/tech-tips.html
Dual ISO may have been an over simplification. Just as stating that it is simply underexposing and applying a different tone curve is.
First ISO from the camera is created in two different ways. From what I understand the first is an analog amplification of the signal from the sensor. This analog amplification gives the ‘native ISOs’. The second is a digital modification of those signals after AD conversion, pulling 1/3 stop back, and pushing 1/3 stop forward from each native ISO, to give the complete range of ISO in 1/3 stop steps. So digitally modifying the signal after AD conversion is still technically changing its ISO.
So can you emulate HTP by simply using a lower ISO underexposing by one stop, then pushing the signal? Not really, HTP also redefines the EV range which the data exists. In addition the digital amplification/modification of the signal, in my tests HTP produced less shadow noise and banding, and more detail than under exposing at one lower ISO and pushing the image one stop.
Below are three RAW histograms from DPP from the consecutive image taken at the same large dynamic range scene.
All three were taken at the same Av and TV value and the ISO and exposure compensation are changed.
In the first image you see that was taken at ISO 100 and exposure compensation was set to -1.
Notice the defined EV range which the image exists is -9 EV to +3.6 EV.
In the second image the ISO was HTP 200 and no exposure compensation.
The defined EV range which the image exists is -8 EV to +5 EV.
The two histograms are basically identical, but, they exist in different defined dynamic range spaces.
The third image is the same first image with ISO 100 and exposure compensation -1, but, then the RAW brightness was increased by one stop. There is some visible degradation in the shadow portion of the histogram as well as the highlight portion is pushed off the top of the EV range resulting in clipped highlights.
So, while HTP isn’t two native ISOs (that's impossible), it’s also not simply underexposing by a stop and applying a different tone curve. As meantioned there is also a difference in the tone curve used by HTP to provide a smoother transition in the highlights.
You've been drinking too much Canon marketing Kool-aid
The only things that HTP does is shoot the image underexposed 1 stop and then set a 'HTP' flag in the RAW file metadata. The 1 stop underexposure is done by shooting at a lower ISO.
So when you choose ISO 200 with HTP on, the camera exposes for ISO 200; but shoots at ISO 100, then marks the file as 'HTP'. Compliant RAW converters (DPP, ACR, Lightroom) honour the HTP flag and increase the image exposure in the RAW conversion.
That's why you lose ISO settings under 200.
Try it yourself:
- Shoot a RAW image at ISO 200 with HTP on
- Shoot the same RAW image at ISO 100 with the same aperture and shutter speed; but with HTP off
- Open the two images in a RAW convertor that does not honour the HTP flag (e.g. RawTherapee). The images histograms will be the same .
http://www.guillermoluijk.com/article/htp/index.htm
http://www.luminous-landscape.com/forum/index.php?topic=21171
http://www.luminous-landscape.com/forum/index.php?topic=20984#entry153950
Phil.
--
http://www.pbase.com/phil_a_mitchell
Already clarified in preceding post.You've been drinking too much Canon marketing Kool-aid
The only things that HTP does is shoot the image underexposed 1 stop and then set a 'HTP' flag in the RAW file metadata. The 1 stop underexposure is done by shooting at a lower ISO.
So when you choose ISO 200 with HTP on, the camera exposes for ISO 200; but shoots at ISO 100, then marks the file as 'HTP'. Compliant RAW converters (DPP, ACR, Lightroom) honour the HTP flag and increase the image exposure in the RAW conversion.
That's why you lose ISO settings under 200.
Try it yourself:
Some links:
- Shoot a RAW image at ISO 200 with HTP on
- Shoot the same RAW image at ISO 100 with the same aperture and shutter speed; but with HTP off
- Open the two images in a RAW convertor that does not honour the HTP flag (e.g. RawTherapee). The images histograms will be the same .
http://www.guillermoluijk.com/article/htp/index.htm
http://www.luminous-landscape.com/forum/index.php?topic=21171
http://www.luminous-landscape.com/forum/index.php?topic=20984#entry153950
Phil.
--
http://www.pbase.com/phil_a_mitchell
Thanks for the detail reply - appreciate you taking the time to indulge me.
I think what you say (quoted) above is only true because the converter (in this case DPP) has recognised the HTP flag and has adjusted the histogram to indicate what it considers the 'corrected' EV range should be.
The way I see it, as does Phil it would appear, the HTP image truly is simply underexposed by one stop and then, by honouring the HTP flag in the raw file, the histogram has been 'adjusted' (as has the tone curve) to represent the 'corrected' range of the image.
It's interesting, particularly for people new to shooting Canon DSLRs, because it means that, if shooting raw, HTP is not really that useful.
It would seem to me to be better practice to simply apply compensation if or when necessary, as manually applying compensation will give the photographer slightly greater control (and will help them understand the concept and practise of exposure in more detail - rather than simply relying on what amounts to an automatic compensation setting).
birdbrain
Senior Member
I must admit I have always treated HTP as an 'out-of-camera-jpeg' function myself.
--
Phil
I wondered why the ball kept getting bigger, then it hit me.
http://www.flickr.com/photos/philthebirdbrain/
M
mcslsk
Guest
Thanks, Phil, very helpful.
ESfishdoc
Senior Member
If iso 100 is "noisier" than 160 I guess I'm just not seeing it.
I've only seen this "native iso" discussed here on this forum and not in other places...
I'm not saying it does't exist... just that it doesn't matter. I liken it to discussions about what is the best oil for my diesel truck.
To quote Rick Sammon's father: "If a picture is so boring you notice the noise, it's just a boring picture" (or something like that).
I think we would all agree that given the choice we would choose 100 over 400 and 200 over 800 and 400 over 1600 if given the choice. Go with the lowest practical iso and work on making a good picture. Nothing is a great substitute for the right exposure.
FWIW I don't use HTP.
Richard
--
http://esfishdoc.com/blog
My Flickr Photostream Slideshow
http://www.flickr.com/photos/33755787@N03/show/
My Images at Photo.net
http://photo.net/photos/esfishdoc
I've only seen this "native iso" discussed here on this forum and not in other places...
I'm not saying it does't exist... just that it doesn't matter. I liken it to discussions about what is the best oil for my diesel truck.
To quote Rick Sammon's father: "If a picture is so boring you notice the noise, it's just a boring picture" (or something like that).
I think we would all agree that given the choice we would choose 100 over 400 and 200 over 800 and 400 over 1600 if given the choice. Go with the lowest practical iso and work on making a good picture. Nothing is a great substitute for the right exposure.
FWIW I don't use HTP.
Richard
--
http://esfishdoc.com/blog
My Flickr Photostream Slideshow
http://www.flickr.com/photos/33755787@N03/show/
My Images at Photo.net
http://photo.net/photos/esfishdoc
I think it is misleading to say that you can 'simply apply compensation' the effect of HTP nearly impossible to recreate in DPP, and I expect at least difficult in other RAW converters.Thanks for the detail reply - appreciate you taking the time to indulge me.
I think what you say (quoted) above is only true because the converter (in this case DPP) has recognised the HTP flag and has adjusted the histogram to indicate what it considers the 'corrected' EV range should be.
The way I see it, as does Phil it would appear, the HTP image truly is simply underexposed by one stop and then, by honouring the HTP flag in the raw file, the histogram has been 'adjusted' (as has the tone curve) to represent the 'corrected' range of the image.
It's interesting, particularly for people new to shooting Canon DSLRs, because it means that, if shooting raw, HTP is not really that useful.
It would seem to me to be better practice to simply apply compensation if or when necessary, as manually applying compensation will give the photographer slightly greater control (and will help them understand the concept and practise of exposure in more detail - rather than simply relying on what amounts to an automatic compensation setting).
I find noise to pretty much be a non-issue for properly exposed images, whether I'm using HTP or not. From what I've seen manually under exposing an applying compensation in post results in more noise than when using HTP.
My advice for the best quality photographs is to use the tools you have (HTP, Evaluative Metering, etc) and properly expose the photograph.
This isn't directed at you or anyone in particular, but, my impression is that those that have the most 'issues' with Canon cameras are those that try to be 'smarter' than the camera or don’t trust the tools in the camera. i.e. not using evaluative metering and getting improperly exposed photographs, not using HTP and complaining about noise and banding when they try to protect highlights and underexpose then push the photo in post, etc.
davinci953
Leading Member
This analysis of HTP when introduced in the 1D Mark III may bear this out. It seems there may be more going on with HTP as well with shadows and detail. I don't think the implementation in later cameras has changed much. I'm curious to know how others interpret this analysis. I generally keep ALO and HTP off. But if anything, it's perked my interest to try to use HTP in some situations.
http://www.imaging-resource.com/PRODS/E1DMK3/E1DMK3EXPOSURE_HTPMODE.HTM
Scott Eaton
Senior Member
(1) Who uses DPP again?
(2) Who cares what DPP can / can't do. HTP can be easily replicated by under exposing and applying the proper highlight curve in LR or Photoshop. In fact, the later two products will do a better job.
(3) If Canon was so smart we would exclusively be using DPP.
HTP is a cheap hack for JPEG shooters living in denial that they need a dSLR to impress their Facebook fans. If you really want to see how BAD HTP is use it for video. It's like going back to VHS.
That analysis was done using JPEG images out of the camera - when HTP is enabled the camera uses a different tone curve when converting the RAW data to the JPEG.
So you do see differences in the resulting OOC jpeg images; but I'm pretty sure there is nothing going on with the RAW data other than the 1 stop underexposure I described in an earlier post.
Phil.
--
http://www.pbase.com/phil_a_mitchell
Proper exposure is a key component to any good photograph - no one's going to argue with that. If Evaluative Metering and HTP work for you that's great; but a good understanding of all the cameras features and options is also necessary.
Sometimes evaluative metering is not the right answer so it's important to know when and how to use center weighted average and spot metering too.
If you shoot JPEG then you might see better results with HTP in some situations; but it's not a panacea - e.g. Canon specifically recommend not using it in low light.
As I pointed out in my previous post, for RAW images, all that HTP does is an ' underexpose and push in post '. It's possible that DPP also applies a different tone curve to RAW HTP images; but I've compared an underexposed image pushed 1 stop in LR with the corresponding HTP image in DPP and I can see virtually no difference.
Personally I'd rather be in control of this and use the technique if and when I need to, rather than the camera forcing it on me. IMO learning how to interpret the histogram will serve you better than simply turning on HTP.
The important thing is to understand what the camera is doing and then you can make an informed decision on what to use.
I hope you don't really believe the camera is smarter than you are
Phil.
--
http://www.pbase.com/phil_a_mitchell
davinci953
Leading Member
When I first looked at the analysis, I thought he was looking at RAW files also. You're right though. It is only OOC jpegs. I'll have to take a look to see if the HTP flag in the RAW file has an effect on the JPEG when the file is converted during post.
If you are using DPP then everything that can be applied in camera to JPG output, can be applied to the RAW file. This includes newer features that were added to later cameras, like basic HDR and image compositing.
So in DPP the different Tone Curve is automatically applied.
John Sheehy
Forum Pro
HTP doubles the height of the read noise floor at ISO 200, and by the time you get to 3200, HTP does not raise the read noise floor.
I use HTP for auto-ISO in manual mode, since there is no way to use EC, but usually when I do that I am expecting high ISOs, or do not plan to show what's in the shadows.
ALO does not affect the SNR at all; it may simply brighten the JPEG, which may affect the visibility of noise, which would otherwise be too dark to see.
--
John
John Sheehy
Forum Pro
Keyboard shortcuts
- f
- Forum