I disagree.
A file, be it a raw, jpeg or tiff is just a collection of data.
In each case you need to know something about the data before you can extract any useful information from it.
- In the case of a raw file, that might be the necessary demosaic algorithm so you can assign the right colour values to the stream of quantized voltages.
- in the case of a jpeg, that might be the decompression algorithm necessary to unpack the compressed data.
In both cases though, neither file can be viewed in any form in its native state (try opening a jpeg file in notepad), but both files can rightly be considered image files in my opinion because that is the nature of the information that have encoded within them.
A key difference between raw and jpeg is the number of steps that need to be taken before the data within would represent a fair reflection of what was seen by the photographer at the time they pressed the button.
- In the case of a raw, you need to apply a white balance, colour levels, contrast levels etc (all the things you can change in your basic raw processor).
- In the case of a jpeg, you will have either applied these before and resaved, locking in your decisions or perhaps the camera has chosen them from you based on a preset.
But in my opinion, they are still both image files, they simply have more of less steps baked in at the point of saving.
As to what constitutes pixel editing, I think that's changed over time.
- Pixel editing used to mean grabbing that brush tool in your favourite editor and overwriting part of the image in some way. ie. a very localised change which was then baked into the saved file so it could not be undone.
- Then came layers in photoshop for example, this meant you could apply destructive changes, but have the option to undo some of those at a point in time in the future.
- Then came the none-destructive editing we see today in the likes of Lightroom and Capture One.
The speed of computers today is such that, in the majority of cases, it's perfectly acceptable to apply these edits on the fly without impacting the user experience, so in said tools, all edits are simply stored as a 'to do' list dynamically applied.
In these tools, its perfectly possible to affect local pixels in the image in a very specific way that could only have been achieved previously by 'pixel editing', but there comes a point however when you need to perform such a large volume of changes (or complex changes more effectively than the tools allow), and this is when you may wish to bake in a number of those changes and save the changes into a new version of the file.
I'd personally make a distinction between changes applied in a none destructive manner in say Lightroom, and changes baked in to an image in something like Photoshop (not withstanding layers etc). I might refer to the latter as pixel editing, but it's a bit of misnomer really.