Luminance meaning

[jetstream]

I'm wondering what luminance means in the context of astrophotography and what it is comprised of.

Thank you,Gerry

[steppenwolf]

Luminance in the context of astro-photography is the collection of all the wavelengths of light that your telescope and camera combination can accept. Images are normally (but not exclusively) captured using a 'Luminance' filter which normally has an Infra Red (IR) cut feature as IR is not required for amateur imaging.

Luminance (L) data is normally used in combination with data captured through red (R), green (G) and blue (B) filters to produce an LRGB image in which the colour is produced by the formation of an RGB image comprising the R, G and B 'channels' but the detail is provided by a painstakingly processed L channel.

[DHEB]

30 minutes ago, steppenwolf said:

Luminance in the context of astro-photography is the collection of all the wavelengths of light that your telescope and camera combination can accept. Images are normally (but not exclusively) captured using a 'Luminance' filter which normally has an Infra Red (IR) cut feature as IR is not required for amateur imaging.

Luminance (L) data is normally used in combination with data captured through red (R), green (G) and blue (B) filters to produce an LRGB image in which the colour is produced by the formation of an RGB image comprising the R, G and B 'channels' but the detail is provided by a painstakingly processed L channel.

Good explanation, thanks. What is not clear for me is if all luminance calculations in the context of astronomy are done by the same standards, that is, is using the same coefficients in the equation to transform the data. Are you, or anyone else, aware of this?

Thanks in advance

[steppenwolf]

I tend to think of Luminance as being a fluid collection of data with no specific standards - you don't even need a specific Luminance filter to capture the data as an IR filter or a good quality LP filter like a Hutech IDAS LP filter can be used to capture the raw data. I think it wise to include IR cut in the channel but that's the only constraint that I'd put on it. The processing of the Luminance channel is very much dependent of subject, sky conditions and personal taste!

[DHEB]

5 minutes ago, steppenwolf said:

I tend to think of Luminance as being a fluid collection of data with no specific standards - you don't even need a specific Luminance filter to capture the data as an IR filter or a good quality LP filter like a Hutech IDAS LP filter can be used to capture the raw data. I think it wise to include IR cut in the channel but that's the only constraint that I'd put on it. The processing of the Luminance channel is very much dependent of subject, sky conditions and personal taste!

I see, so in principle there are no "standard" coefficients to transform three dimensional (RGB) to one dimensional data?. Is it largely a subjective matter?

[steppenwolf]

Quote

Is it largely a subjective matter?

It is for me but others may take a different view!

[ollypenrice]

Subjective for me, too. Luminance is, for me, the spectrum which goes from blue, through green, to red. (I use a Baader L filter.) What does it add to RGB? Speed. In theory an hour of L ought to be roughly equivalent to an hour of red plus an hour of green plus an hour of blue. In fact an hour of L goes far deeper than that. My experiments suggest that the best you can hope for from 3 hrs RGB is about 75% of an hour of luminance and that is being optimistic. Once you start to process your image this becomes highly significant. You can process harder without breaking through the noise floor. That means you can stretch the faint stuff and sharpen the bright stuff without breaking down into noise.

I'm a great believer in luminance. I recently conceded that my friend Tom O'DOnoghue was right to go for L, even on a starfield shot. 

In the past I've taken the view that L can impede the capture of a tight, colourful starfield. Maybe not.

Olly

[Stub Mandrel]

On 31/07/2016 at 17:32, Cinco Sauces said:

I see, so in principle there are no "standard" coefficients to transform three dimensional (RGB) to one dimensional data?. Is it largely a subjective matter?

I'm sure this will get poor Olly tearing his hair out, but...

I was persuaded early on about the benefits for splitting my RGB images into two, processing one for the best colour, perhaps using slight blurs to help colour up over-exposed patches and another 'synthetic luminance' layer processed to give the best detail and contrast with acceptable noise levels, without worrying about the colour balance, this is then used as a luminance layer on top of the RGB layer.

I sometimes find that a my RGB images show better contrast in the green and blue channels than the red one so my 'luminance' layer can look decidedly horrible.

Of course this is a VERY subjective process as two people may choose to tweak even a synthetic luminance layer to bring out different features.

 

[ian_bird]

23 hours ago, ollypenrice said:

In the past I've taken the view that L can impede the capture of a tight, colourful starfield. Maybe not.

Olly

Interesting! I remember you erasing all the luminance data in the star field from a certain M31. Would you do that differently now?

And if so - what made you change your mind? Enquiring minds and all that.....

Cheers

Ian

[ollypenrice]

1 hour ago, ian_bird said:

Interesting! I remember you erasing all the luminance data in the star field from a certain M31. Would you do that differently now?

And if so - what made you change your mind? Enquiring minds and all that.....

Cheers

Ian

No, for a galaxy starfield I'd still go for an RGB-only approach. The interest is in the galaxy and small field stars really help to put the eye's attention there. However, Tom and I recently did a 'stars only' image of M22 (globular) in a rich MW starfield and, when the interest is only on the stars, I had to concede that the L layer brought more 'pop.'

Olly

[ian_bird]

19 hours ago, ollypenrice said:

 

20 hours ago, ian_bird said:

Interesting! I remember you erasing all the luminance data in the star field from a certain M31. Would you do that differently now?

And if so - what made you change your mind? Enquiring minds and all that.....

Cheers

Ian

No, for a galaxy starfield I'd still go for an RGB-only approach. The interest is in the galaxy and small field stars really help to put the eye's attention there. However, Tom and I recently did a 'stars only' image of M22 (globular) in a rich MW starfield and, when the interest is only on the stars, I had to concede that the L layer brought more 'pop.'

Olly

 

Ah! That makes sense.

Thanks.

Ian