# How it Works (Time-Depth Conversion)

In Insight, every location that needs to be converted will be converted individually based on the following steps. These apply to horizon points, volume traces or anything else that you want to convert between time and depth.

## Step 1: Retrieve a velocity trace for that location

• Based on the velocity model used for the conversion, Insight retrieves a velocity trace for that location. If there is no velocity trace at the location, Lagrangian interpolation method is used to produce one from the surrounding data.
• If there is insufficient surrounding data (due to a hole or being past the edge, etc), then no velocity trace will be produced. There will be no extrapolation done in the background. However, you can extrapolate the velocity or fill the holes with the Volume Interpolation/Extrapolation process.

Note: Insight honours every node in the velocity, allowing you to be as precise as you want. The standard practice in DUG's service division is to use a velocity sampled at around 100m x 100m x 24ms/20m unless there is a strong reason for more density. If you want to downsample, use the Velocity Conversion process to do the vertical resampling and export the converted volume onto a sparser IL/CL grid. Do NOT use the volume export to downsample. The Velocity Conversion process follows a special process to ensure that the output velocity is equivalent to the input. A naïve downsampling that is not velocity-aware will not produce an equivalent model. This is why you must never use other processes (such as Volume Resampling or a simple export) to resample a velocity model.

## Step 2: Convert velocity trace from interval, RMS, average, etc, to time/depth pairs

Based on the type of velocity, Insight converts the velocity trace to time/depth pairs. Depth values are calculated from time according to the velocity type:

• Average velocity
• Z =Velocity * TWT/2
• Interval velocity
• Z = 0
• Z =(Velocity * Sample Interval/2) + (previous depth)
• RMS velocity
• Dix Formula, to convert to Interval, with constraints on the interval result, then
• Z = Interval, as above.

Once the Time-Depth relationship is determined, linear interpolation is used to find intermediate values.

## Step 3: Use time/depth pairs to build a function that will convert any time value to its equivalent depth, or vice-versa

Any value above the seismic reference datum (0 TWT) or below the bottom of the velocity trace is not converted. Insight does not extrapolate.

## Step 4: Convert time-depth based on type of data

• For data with point sources like a horizon, polygon or fault picks, the single point of interest is converted and drawn at the appropriate location on the screen. Next, Insight returns to step 1 for the subsequent point.
• For volumes, regardless of whether or not they are being time/depth converted, these are interpolated onto the screen at the display resolution. This is an important detail because it means that data is not put through a lossy vertical resampling process. The time/depth conversion is exact to the limit of your screen resolution. Hence, as you zoom in closer, it will interpolate the converted data at ever increasing levels of detail. In contrast, if you export the volume, it must be actually resampled onto the requested export sample interval.