We may agree that during the combustion, regardless if the one of wood, gasoline, or in your example, hydrogen gas, is an exothermic reaction. In a simplified view, products of these reaction are gaseous water; in the processes burning organic matter varying relative amounts of carbon monoxide and carbon dioxide are to be considered, too.

The gaseous products formed take up a volume, much more than their starting materials in the condensed state (liquid, or solid). Because of the local heat, these gases aim to expand rapidly. In contrast to burning wood for a bonfire, the volume offered by a test tube is confined except for the mouth of the test tube, where a decompression to reach ambient pressure may take place. This small scale explosion initially occurs at the interface of air's oxygen at the mouth of your test tube filled with hydrogen, yet this frontier actually propagates up to the closed end of the test tube, too. Sometimes, especially with lengthier test tubes, this may be seen from the outside, too.

Gasoline driven combustion engines work by the same principle; spark plugs locally ignite a mixture of finely dispersed fuel and air, triggering a rapid combustion (explosion) where the reaction products again aim to expand. As they operate at a larger scale as the test tube experiment, the generated sound is much more intense, and these vehicles need a silencer, too.

While your bonfire equally yields $\ce{CO2}$ and water, these products may expand easily; there is no pressure generation, hence there is no sudden expansion (at least, not at this scale), and hence we may perceive the bonfire as (in comparison to the other examples) as peacefully calm.

I suggest to take a view on videos about explosions of hydrogen gas, and an appealing demonstration how it looks in the inner of such an Otto engine here. In the later example (the display of the combustion starts around 2:51 min and again at 6:00 min), gasoline was replaced and a high speed camera looks through a special lid, yet the example still shows nicely the propagation of this frontier of explosion.

We may agree that during the combustion, regardless if the one of wood, gasoline, or in your example, hydrogen gas, is an exothermic reaction. In a simplified view, products of these reaction are gaseous water; in the processes burning organic matter varying relative amounts of carbon monoxide and carbon dioxide are to be considered, too.

The gaseous products formed take up a volume, much more than their starting materials in the condensed state (liquid, or solid). Because of the local heat, these gases aim to expand rapidly. In contrast to burning wood for a bonfire, the volume offered by a test tube is confined except for the mouth of the test tube, where a decompression to reach ambient pressure may take place. This small scale explosion initially occurs at the interface of air's oxygen at the mouth of your test tube filled with hydrogen, yet this frontier actually propagates up to the closed end of the test tube, too. Sometimes, especially with lengthier test tubes, this may be seen from the outside, too.

The combustion of hydrogen is worth a special note, because -- as correctly spot by airhuff -- formally there is volume contraction along

$$ \ce{2H2 + O2 -> 2H2O}$$

which is more than compensated by the thermally caused expansion of the gaseous water.

Gasoline driven combustion engines work by the same principle; spark plugs locally ignite a mixture of finely dispersed fuel and air, triggering a rapid combustion (explosion) where the reaction products again aim to expand. As they operate at a larger scale as the test tube experiment, the generated sound is much more intense, and these vehicles need a silencer, too.

While your bonfire equally yields $\ce{CO2}$ and water, these products may expand easily; there is no pressure generation, hence there is no sudden expansion (at least, not at this scale), and hence we may perceive the bonfire as (in comparison to the other examples) as peacefully calm.

I suggest to take a view on videos about explosions of hydrogen gas, and an appealing demonstration how it looks in the inner of such an Otto engine here. In the later example (the display of the combustion starts around 2:51 min and again at 6:00 min), gasoline was replaced and a high speed camera looks through a special lid, yet the example still shows nicely the propagation of this frontier of explosion.