top of page

Tech Battles: EHRS/UHRS Sparker Sources: Should we compress the wavelet beyond its native pulse?

Writer: Alastair MacLeod
Alastair MacLeod
5 hours ago
3 min read

Seismic data is the earth’s reflectivity series convolved with a source wavelet. By compressing that wavelet towards an impulsive spike, we aim to get closer to the earth’s true reflectivity and improve the resolution of the seismic image.


But is pushing wavelet compression always a good idea?


Team YES argues that the ideal source wavelet is an impulsive spike, and that processing should aim to get as close to this as possible. With good-quality acquisition, particularly from multi-layer sparker sources, wavelet compression can help realise the benefits of the available broadband energy.


Team NO argues that UHRS sparker acquisition is often inconsistent, and that pushing compression too far can introduce unwanted artefacts. In unconsolidated sediments, where Q values can be high, subsequent Q compensation can amplify noise rather than recover useful high-frequency information. This is particularly important when trying to identify small geological features such as boulders.


So, should we push wavelet compression beyond the native pulse, or is the risk simply too great?


This is Episode 2 of our #TECHBATTLE series, featuring RockWave seismic experts, Sandeep Bhamber and Greg Hulks, going head-to-head in a battle of YES vs NO, exploring some of the key considerations behind this question.


Who do you think wins? Let us know by engaging in the comments of our YouTube video below.


And BBC, if you're watching… we can make Sandeep available to present the next series of Countryfile. Those hand gestures! 😂


The full transcript can be found below the video:




TEAM NO

Before we argue about it, what do we actually mean?

What we see in seismic isn’t the geology directly.

It’s the earth’s reflectivity series - convolved with a source wavelet.

The closer that wavelet gets to an impulsive spike, the closer our seismic response gets to the earth’s reflectivity.


TEAM YES

Exactly! The ideal seismic source wavelet is an impulsive spike.

But a sparker naturally produces a pulse with positive and negate components, rather than one clean spike.

The only reason we struggle to inject these into the earth is due to mechanical issues –

Compressing the wavelet is an attempt to overcome those  —  making the wavelet narrower and narrower and…

 

TEAM NO

Woah woah, Greg, stop!

Compressing it further to be more spike-like might seem like a good way to increase resolution…

but it’s awfully risky.


TEAM YES

Why risky?

Anyone who works with seismic in oil and gas is probably thinking:

Why are we even asking this?

Exploration seismic processing doesn't worry about pushing wavelet compression...

so why should we in ultra-high resolution?


TEAM NO

Because you're comparing apples with lemons!

With UHRS, the sources are often less reliable,

and we’ve seen quite a bit of variation from survey to survey in terms of the source type and setup.


TEAM YES

You’re almost as square as our amplitude spectra will be after wavelet compression.

Low frequencies, mid frequencies, high frequencies…

all recovered equally.


TEAM NO

You’re not living in the real world are you!

And anyway… I’ve not even mentioned boulders yet!

As described in THIS paper by Nick Woodburn, compressing the wavelet can lower the dominant frequency.

📄 Paper: N. Woodburn, A. Hardwick, and R. Herring (2012), "Marine broadband seismic: Is the earth response helping the resolution revolution?" (SEG Las Vegas 2012 Annual Meeting) | DOI: 10.1190/segam2012-1111.1 | Full PDF: https://www.tgs.com/hubfs/Technical%2...  #seismic #geophysics #seismicprocessing 

Now… Q values can be huge in unconsolidated sediments.

You compress the wavelet... then apply Q correction to recover the high frequencies.

If the acquisition isn’t perfect, the only high frequency you’re boosting is noise that your client is going to hate!


TEAM YES

Very presumptuous of you!

Who’s to say the data wont be able to support it?

On good acquisition, compression can make a significant improvement.

And if you've got a good, multi-layer sparker source...

you may have more low-frequency energy there to begin with.

Compressing the wavelet can reveal the benefit of that broadband content.


TEAM NO

And that's exactly the problem.

If you've got a poor-quality sparker

Or all the real-world imperfections that come with acquiring seismic in marginal weather

The low frequencies you're trying to reveal might not be there in the first place.


TEAM YES

I don’t think that’s a reason not to look for them!

It sounds to me like we just have to know what we’re putting into the water...


 

Comments


bottom of page