from assayofficebirmingham.com

The debate is very much present in Sanskrit sources because it is contiguous to the possibility of intellectual perception. In fact, if sense-perception can be constantly improved by practice, it seems plausible to assume that it could be improved until the point in which the intellect can perceive in its own right. And this opens the door to super-sensuous perception.
But let me come back to the case of expert knowledge. The standard example is that of an expert of gems, who can recognise a genuine gem through sense perception. There are interesting debates about what exactly can be perceived through sense-perception in this case. For instance, Veṅkaṭanātha in his Seśvaramīmāṃsā ad PMS 1.1.4 (English translation under preparation by me) says that an expert can recognise the different hues of colours of a gem, which are concealed to lay people due to their similarity. However, even an expert cannot sense-perceive the preciousness of a gem —he instead only infers it, although the inference is not verbally formulated. By the way, Michel X, while commenting on the the post which triggered the present one* expresses some skepticism concerning the experts’ perception in the case of works of art. He might be right, and I can easily imagine Veṅkaṭanātha claiming that judging about the authenticity of a work of art implies implicit inferences rather than sense-perception alone.

To sum up, authors like Yāmuna (in his Ātmasiddhi, the relevant passage is translated towards the end of this article) maintain that perception can undergo an indefinite progress, basically until everything can be directly perceived —so that this argument leads to an evidence for the existence of God. Veṅkaṭanātha, by contrast, claims that perception can be improved through exercise, but that such improvement has precise limitations. Just like different people can jump more and more, but no one can jump until the moon, so visual perception will never grasp what is intrinsically outside the precinct of application of sight, e.g., smell.

Another instance of the argument of the continuous improvement I could find is located in Śaṅkara’s Yogaśāstravivaraṇa (the attribution is doubted, and sense perception does not help, thus I am here following the opinion of an expert, Kengo Harimoto), which has been recently critically edited and translated by Kengo Harimoto (see here). The argument is found in the commentary ad Yogabhāṣya 1.25 and leads to the establishment of an omniscient God through the fact that knowledge can always be improved and that it needs to achieve a peak somewhere. Maṇḍana Miśra, who is believed to have been a senior contemporary of Śaṅkara, rejected the theological part of the argument (see Harimoto, p. 11). Harimoto does not say where, but I imagine that this might be in his Vidhiviveka, around 1.14, where the discussion on omniscience takes place.

UPDATE: A further instance of this argument is discussed here.
*This post has been stimulated by Helen De Cruz’ discussion of the topic, here, and by two comments I received about it (many thanks for them, by the way!).