The images were very good, but a lot of them included other entitites (e.g. bacteria), so weren't homogenous in that they contained only the virus particles. Are you aware of any studies where this was actually done and demonstrated? I would like to read about it.
The images were not of viral preparations for infection testing.
As far as infection testing goes, I'm not a virologist, so I don't have that information to hand. I suggest you search Google Scholar or a similar resource.
... what I read was that in these experiments, the viruses aren't properly isolated, so that one can't determine that it is the viruses themselves causing the symptoms (rather than bacteria, toxins or other material).
It is not necessary to have a pure isolation of the virus, and it could be argued that using the virus alone would not give a true picture of infection in vivo, where this is never the case. The solution is the use of controls.
For example, a common method of virus isolation is filtration to remove all contaminants larger than the virus itself. This will eliminate bacteria, which (except for the rare 'giant' viruses) are at least an order of magnitude larger. Any remaining material, (cellular debris, metabolites, etc) can be controlled for by using identical preparation methods on uninfected cells and exposing a control group to that preparation. Control procedures of this kind are standard in scientific work.
When a particular viral culture, pure or otherwise, has been shown to produce the same disease symptoms and multiplication of the virus in the host, by many independent labs, using different hosts, and the virus has been identified infecting and multiplying in those host cells, particularly when the symptoms are consistent with damage to the specific host cells the virus infects, it becomes possible to say, beyond reasonable doubt, that the virus is the main or direct cause of the disease.
I don't work with viruses, but have been disappointed with scientists not using a rigorous scientific method in other fields, so would not be too surprised to learn it was so in this one. Would be happy to learn otherwise, though.
The use of controls in biological experiments is standard practice.
There is sloppy work in science, as in all other human endeavours, but seminal work in particular (such as the identification of a novel viral disease), undergoes extreme scrutiny and requires replication by independent groups.