Canon EOS RP RAW editing flexibility?

Sgt_Strider wrote:

dpfan32 wrote:

You have visible noise, when you lift the shadows too heavy. Look at my example above

I apologize if I'm being rude, but we're not even talking about lifting shadows as there is universal agreement that there is noise. What we are talking about is exposing to the right or at least getting the exposure right and then dialing back the settings like exposure. The consensus thus far is that noise won't be an issue in this scenario?

The noise may become an issue depending on how much dynamic range you need for the given scenery and how that compares to the dynamic range of the camera.

Higher dynamic range of your camera means you need to do less exposure blending/HDR. The camera covers greater range of contrasty scenes in one shot.

dpfan32 wrote:

This is the image I shot

Unedited

You want to see the ugly noise?

Lights 100% down and Shadow 100% up in Lightroom

This is a Sony A7 photo

Lights 100% down and Shadow 100% up in Lightroom + 1EV Exposure on top needed.

So if you want to do this type of shots you better bracket or buy a Nikon camera!

Your Sony ICLE-7 does have a couple of steps better DR at base ISO than the RP.  But, the Canon EOS R5 and R62 both have a bit more DR than either the Nikon Z72 or Z9, according to the dynamic range charts at Photons to Photos, so I don’t think the OP needs to move to Nikon in order to get those kind of shots.  The new EOS R8 has the same sensor as the R62, and therefore should do really well in low light and in high contrast scenes.

• Sgt_Strider wrote:

• If you shoot landscapes and properly use ETTR (exposure-to-the-right) with your RP, you may get good results comparable to someone who doesn't use ETTR with an R or even R5. Although you'd get much better results had you used an R5 with ETTR.

  Lowering 'exposure' in Lightroom after ETTR doesn't increase visible noise. Sitting at base ISO and using ETTR is the way to squeeze the possible dynamic range out of your camera.

  But if you can't use ETTR (say because it's too dark or you can't use long enough shutter speed), yes you'd better use higher ISO 3200 right in the camera rather than ISO 800 plus 2-stop exposure lifting in Lightroom. That's because the RP is not 'ISO-invariant'.

• Thank you! That's what I was asking!
• I have to admit, I'm still confused with the ISO-invariant terminology.

So called ISO-invariance is just as described above: your shots at some high ISO have the same visible noise as shots taken at lower ISO plus 'exposure' lifted by corresponding amount in a photo editor.

The R5 is practically ISO-invariant from ISO 800, but the RP is never ISO-invariant:

https://www.photonstophotos.net/Charts/PDR_Shadow.htm#Canon%20EOS%20R5,Canon%20EOS%20RP

The invariance starts where the graph goes flat which unfortunately is never for the RP.

That means, with an RP, it's best to lift the ISO in camera. However note the dynamic range generally reduces by 1 stop when you raise the ISO by 1 stop.

I'm also aware that the R and the R5 have superior sensors. I'm just trying to understand how to extract every possible performance out of the RP sensor.

I mainly shoot on a tripod. Can't I just go on bulb? How well does the RP handle long exposure noise when shooting at base ISO?

That's a bit different from what those charts may show. When shooting long exposure you may get non-random thermal noise, plus hot pixels. Tbh I haven't seen long exposure comparisons across different cameras. But I guess the RP shouldn't be too bad in that regard - you may do your own tests to see if it provides satisfying results.

Sgt_Strider wrote:

I have to admit, I'm still confused with the ISO-invariant terminology.

With a sensor that is ISO-invariant, you can shoot at ISO 100 and increase the exposure by say four stops, and the resulting file will be identical to the file you would shoot at ISO 1600.  So technically, you could leave your camera ISO at 100, and make all your exposure adjustments in post processing, and the results would be the same as if you had set the camera to the appropriate ISO at the time you took the shot.

danferrin wrote:

Sgt_Strider wrote:

I have to admit, I'm still confused with the ISO-invariant terminology.

With a sensor that is ISO-invariant, you can shoot at ISO 100

ISO 100 is a bad example. Most of the modern cameras (Canon, Nikon, Sony, Fuji) are ISO-invariant from about ISO 640-800.

Above I've posted a link to Photonstophotos chart where you can check ISO invariance range for your camera.

Quarkcharmed wrote:

danferrin wrote:

Sgt_Strider wrote:

I have to admit, I'm still confused with the ISO-invariant terminology.

With a sensor that is ISO-invariant, you can shoot at ISO 100

ISO 100 is a bad example. Most of the modern cameras (Canon, Nikon, Sony, Fuji) are ISO-invariant from about ISO 640-800.

Above I've posted a link to Photonstophotos chart where you can check ISO invariance range for your camera.

So the RP, per that chart, is ISO invariant at base ISO and then from 300-400 onwards it is better to just raise the ISO versus underexposing and recovering in post? Am I reading this right?

Fjzk wrote:

Quarkcharmed wrote:

danferrin wrote:

Sgt_Strider wrote:

I have to admit, I'm still confused with the ISO-invariant terminology.

With a sensor that is ISO-invariant, you can shoot at ISO 100

ISO 100 is a bad example. Most of the modern cameras (Canon, Nikon, Sony, Fuji) are ISO-invariant from about ISO 640-800.

Above I've posted a link to Photonstophotos chart where you can check ISO invariance range for your camera.

So the RP, per that chart, is ISO invariant at base ISO and then from 300-400 onwards it is better to just raise the ISO versus underexposing and recovering in post? Am I reading this right?

Per that chart, the RP is nowhere ISO-invariant for practical use.

You can see flat groups like ISO 100-120*, then 160-200-250 etc. - that's because the camera uses digital gain (multiplication) between those ISO values, which is the same as digital 'exposure' slider in a raw photo editor. But it's not practical to use ISO-invariance within such short ranges covering less than a stop. You might as well make the camera use 1-stop increments for ISO, instead of default 1/3-stop increments.

* ISO 50 is marked as extended ISO in that chart (and in the camera manual) - in practice that means the raw data at ISO will be the same as ISO 100, just different ISO tag and middle grey point. So ISO 50 doesn't really help invariance.

dpfan32 wrote:

Yeah this was an interesting point.

ISO 50 setting make a +1EV overexposure right?

Well in Canon cameras the extended ISO 50 introduces a metering bias, which effectively makes the camera overexpose compared to the same metering mode at ISO 100 (note that 'overexposure' is relative to the current metering mode, no more no less).

In connection to ETTR. When you hint the right side of the Histogram this doesn't mean that it's the end.

This only means that this is the last point where the JPEG wouldn't blow out!

Yes, that's why you might want to use something like UniWB (so called universal white balance) if you really want to maximise the DR of your camera.

The RAW file can handle up to 2 EV overexposure, I read somewhere.

The question is, this overexposure - it's 'over' against what exactly? If you're on standard evaluative metering, you may already be blowing out some highlights because metering makes mistakes, so overexposing +2 ev against evaluative metering can easily make things much worse.

If you don't want to go for all the hassle with UniWB, there's another simple technique to ETTR: use spot metering on the brightest part of the scene (e.g. white clouds), then use exposure compensation of +3 stops (the exact amount has to be tested with your particular camera). This may bring you pretty close to ETTR. Unfortunately, with Canon cameras, spot metering is bound to the dead centre of the frame, which means you'll have to recompose every time you change the exposure, which makes this method impractical if you're shooting from a tripod.

Another method - use RGB histogram and ignore red and blue channels, ETTR by the green channel only (suggested by Iliah Borg IIRC). Again, unlike UniWB, it's not 100% reliable, but it may work in many daytime situations.

The reason I actually shot at some point wit ISO 50 on my EOS 6D Mark II was actually a EOS RP review I watched on YouTube, where the reviewer showed, that the ISO 50 setting had much less noise in the shadows that ISO 100!

Of course he didn't shoot a high dynamic range scene.

The result had been the same if he overexposed the ISO 100 RAW 1 stop while shooting and made the image darker afterwards.

Am I right?

Yes.

But this is IMHO a good practice for ETTR:

In my opinion, no, shooting at extended ISO 50 is not ETTR. There was a similar very interesting debate on DPR recently, I don't remember who it was with exactly.

So the issue is, as above, that with ISO 50 you effectively use a constant metering bias against your current metering mode, and it doesn't guarantee you're getting a constant 2.5-stop headroom. And even when you are, you're still 1.5 stops away from the 'ideal' ETTR.

But yeas I still use the RP an my main camera and please tell me if my thoughts are correct about the ISO 50 with the ETTR I did at some point or did I misunderstood this ISO 50 thing

I'm not sure about the RP, but in the R5 the histogram may be showing no clipping at ISO 50 but you actually get clipped highlights in raw.

Yes in many cases ISO 50 brings you 1 stop closer to 'ideal' ETTR, but it's not very reliable so occasionally you will be getting blown out shots.

Quarkcharmed wrote: [...]

If you don't want to go for all the hassle with UniWB, there's another simple technique to ETTR: use spot metering on the brightest part of the scene (e.g. white clouds), then use exposure compensation of +3 stops (the exact amount has to be tested with your particular camera). This may bring you pretty close to ETTR. Unfortunately, with Canon cameras, spot metering is bound to the dead centre of the frame, [...]

On 1-series bodies you do get AF-point-linked-spot-metering. And only on 1-series bodies, the R3 lacks it as well.

koenkooi wrote:

Quarkcharmed wrote: [...]

If you don't want to go for all the hassle with UniWB, there's another simple technique to ETTR: use spot metering on the brightest part of the scene (e.g. white clouds), then use exposure compensation of +3 stops (the exact amount has to be tested with your particular camera). This may bring you pretty close to ETTR. Unfortunately, with Canon cameras, spot metering is bound to the dead centre of the frame, [...]

On 1-series bodies you do get AF-point-linked-spot-metering. And only on 1-series bodies, the R3 lacks it as well.

Which I don't understand, it would be so easy to implement in mirrorless.