This is a more important and difficult question than I think you realize. Considering how much of Roman politics was oral and that Roman law so heavily emphasized libertas and the role of the populus the degree to which public oratory was actually "popular" (that is, directed at the people) is a major problem of our understanding of the late Republic. Fergus Millar argued some time ago that the Republic, previously seen as mainly a network of patron-client relations with those at the very top deploying their political capital with little interest or input from the populus, was truly "democratic" in that the plebs were truly in control. Millar's suggestion has not been received particularly well, but it has forced a revisiting of the question of the "democratization" of the Republic. Central to this issue is whether at speeches ad populum or contiones people could attend in large, representative numbers and could hear and understand what was going on.

It's clear enough that public addresses were a prized source of entertainment in the city. Catullus, for example, mentions a man in the crowd during one of Calvus' speeches against Vatinius exclaim di magni, salaputium disertum! The meaning of salaputium is disputed, but it probably means something like "great gods, an elegant midget!" The Latin is obviously vulgar, and the surprise suggests someone who doesn't know Calvus. This is clearly someone who stopped to watch a forensic speech while strolling through or near the Forum, as we are told repeatedly people did. Moreover, not only was this entertainment, it was the way that news spread through the city. Rome had no newspapers, and while the acta senatus were posted daily starting in 59 they were not really a viable means of getting information through (and the acta, from what we can tell, did not really give useful information). Public speeches were an important means of allowing the plebs, for one thing, to know what was going on in the senate. Cicero delivered the Third Catilinarian directly after the meeting of the senate at which the Catilinarian conspirators were exposed, summarizing to the people what had happened and what was known about the threat.

There are several problems here. The first is space. The traditional places for address to the people, the Forum, the Capitoline, and the parts of the Campus Martius used for voting were too small to accommodate a majority of the urban population. Mouritsen has pointed this out largely with respect to voting, that the spaces used for election and legislative voting, theoretically involving the entire Roman populus, were far too small for more than a few thousand people to vote. The same can be said of public addresses. Moreover, the plebs could not simply convene at any time and place at will. Rural inhabitants were mainly farmers, whose mobility was limited. I think Mouritsen somewhat overstates his belief that only wealthy rural landowners could move in and out of the city for the vote, since we know that the seasonal transport of rural laborers was a major economic activity (Vespasian was rumored to be descended from a man who had made his fortune transporting workers from Umbria to Sabine country). Small-time rural farmers, either owning or renting plots of a couple jugera, did likely have the opportunity to travel for short periods to the city, since agricultural work can leave relatively long stretches of free time in off seasons. Still, rural inhabitants couldn't just up and head to the city every time there was an important trial or contio. Urban inhabitants, on the other hand, had to work. The view of the urban population as the "idle plebs" living on the largess of the state and wasting their time at the arena or making trouble in the streets has long been discarded. The plebs worked to survive. If they were lucky enough to have a trade they worked as skilled artisans, shopkeepers, or craftsman. These individuals, owning their own shops, renting permanent lodgings, and secure in their ability to purchase food could afford to close up their shops for a day (or leave them to their workers or slaves, if they had a few) to head into the Forum to hear a speech or attend a contio. But this so-called plebs media made up a minority of the urban population. Most urban inhabitants were day-laborers without permanent lodgings, without permanent work, and with little but their own bodies and those of their families. From what little evidence we have of a worker's pay and the cost of living in the city it's pretty clear that these people had to work pretty much all the time simply to eat. A day of work lost because there was simply no day-work to be had was bad enough--a day willingly lost to listen to a speech was worse yet. It should not be supposed, however (and for some reason Mouritsen totally overlooks this, although Morstein-Marx's Mass Oratory corrects most of Mouritsen's errors with respect to public speaking), that these day-laborers were entirely absent from forensic life. The Forum and the areas surrounding it, despite the effort to push most business outside of the Forum in the second and first centuries, appear to have been a good place to pick up workers (from what little we have). Laborers unable to find work for the day were likely all around the forum, and probably often found themselves in the crowds of forensic speeches. Indeed, they may well often have found themselves some money in the process. Political participation "for hire," in a manner of speaking, is well-attested in the late Republic. A little too well-attested, in fact, since it was a common accusation even posthumously to accuse people of hiring their supporters. Certainly money was exchanged often. Our strict modern idea that any exchange of money in politics is some form of bribery is not one that would have been totally current in the Roman world, much of which functioned by the exchanges of goods and favors between patrons, clients, and peers. Moreover, the fact that political participation (of which attendance at public speeches was one part) meant a day out of work for the lower classes was well recognized--a Athens the radical democrats established that attendees of the Assembly would be compensated in pay, allowing more of the poor to participate in the democracy.

So that complicates the picture a little bit, and I don't want to dwell on it excessively since it's a very lengthy argument that's not necessarily totally relevant. What's clear, however, is that under normal circumstances forensic speeches and contiones did not necessarily have especially large attendance. What attendees were present were probably largely educated and/or supporters of the particular speaker--the organization and function of the contio in politics seems to suggest this (however, for a full treatment don't rely solely on the chapter on the contio in Mouritsen's Plebs and Politics, but go instead to Morstein-Marx's Mass Oratory, easily the most complete treatment available). Public speeches could, however, and did get hard to hear. Roman speakers spoke with no amplification, and unlike in the Athenian Assembly, which had decent acoustics and was an alright platform for public address, the public spaces for speaking in Rome were woefully inadequate. Orators trained to project their voices and make it clear through hand motions and rhetorical devices like repetition and tricolons what they were saying, but nevertheless in a packed crowd probably "telephone" was the only way the furthest members of the crowd could hear everything. This, provided that the audience was quiet and peaceful. Quite frequently public speeches could get rowdy, either because of an unpopular speaker or because of some critical matter. When Cicero spoke at the trial of Milo he was unable to be heard and was interrupted. What precisely happened here is unclear--Plutarch says that the sight of Pompey's soldiers scared Cicero so much that he was unable to finish the speech, but Asconius says that though the hostile crowd shouted him down he finished his speech even though presumably he couldn't be heard--but at any rate disturbances of this sort were hardly unknown in Roman oratory. Not infrequently were orators forced to speak vainly into a crowd that was making so much noise that they could not be heard. At public debates this threat was even greater. When Clodius debated Pompey over the cura annonae he embedded his supporters into the crowd and had them react according to instructions to signals he gave, eventually resorting to a sort of call-and-response tactic to insult Pompey and prevent him from speaking properly. Eventually Milo's armed partisans dispersed the crowd with force, once Pompey was unable to keep the crowd in check.

So yes, in a very real way the acoustic limitations of oratory were a major point in Republican political life. I highly recommend you first look at Mouritsen's Plebs and Politics, mainly because it set the stage for Morstein-Marx's Mass Oratory, which is quite thorough and definitive.