Yes, and Rodinal reduces film speed anyway, even in normal development.
According to whom, and compared to what?
Let's back up and touch on the wider subject of "compensating development" - which applies to both Rodinal and "stand development," independently of each other.
Development is a chemical reaction, with two consequences (at least) beyond converting exposed silver halide into metallic silver.
1) The reagents in the developer get used up by the reaction.
2) The chemical reaction produces byproducts in the developer (bromine or bromine compounds, most usually), which themselves slow (restrain) the developing reaction.
This means that developer in contact with parts of the film with a lot of exposure (highlights) weakens at a given rate, and developer in contact with parts of the film without much exposed silver to react with (shadows) remains potent for a longer time.
The purpose of agitation is to remix the "used" and "fresh" developer touching various parts of the image fairly often, so that the developing is consistent across the whole image.
However, there are scenarios where the differential changes in the local strength of the developer are useful - most commonly, in extending the tonal range (but also creating edge effects between shadow and highlight areas, about which more later). This is called "compensating development" since it can compensate for a high-contrast-range subject or lighting.
On the whole, compensating development tends to produce the highest film speed possible for a given emulsion, in that it gives "full-time" development to the shadows, while the highlights use up the reagents and development comes to a relative stand-still. Film speed is, of course, measured by shadow density for a given exposure - the point where exposure density exceeds unexposed film-base-plus-fog by 0.1.
There are multiple ways to achieve compensating development:
1. "Stand development" - no agitation means the highlights use up the developer in contact with the film and then cease developing, while the shadows only slowly use up their "local developer" and get more actual developing time.
2. Split development - essentially stand development by chemical means. Example, Diafine. The film is put into a bath A of developing compounds minus the required alkali environment, and soak up as much as the gelatin can hold.
Film emulsion, although thin on a human scale, can soak up a LOT of developer molecules as the gelatin swells. It can be noted that Diafine gives a higher speed boost to thick-emulsion films like Tri-X than it does to thinner emulsions like TMax or Tech Pan, that can't soak up as much developer and thus can't benefit as much from compensating techniques.
Then the film is placed in an alkali bath B without additional developing compounds, and the alkali starts the development. With no new source of developing compounds, the highlight development quickly comes to a standstill, while the shadows keep developing in relatively "fresh" developer for maximum film speed.
3. Water-bath development. A favorite of Ansel Adams. Carries the split-development idea further. Put the film in developer and process for a partial development time. Then put the film into plain water without agitation. A certain amount of developer remains soaked up in the emulsion gelatin, and works away, again dying faster in the highlight areas than in the shadows. The developer/water steps can be repeated several times.
4. High-dilution developers. Example, Rodinal at 1:50 or 1:100. In effect, a developer containing its own water bath. With high dilution there are relatively few developer molecules available per ml, so they get rapidly get used up in reactive highlights, and development ceases until the next agitation. While the shadows develop "full-time."
Edge effects are a byproduct of compensating development. Where a strong shadow and a strong highlight are side by side, weakened developer from the highlight diffuses slightly through the gelatin into the edge of the shadow, reducing development, and less-used developer from the shadow diffuses into the edge of the highlight, increasing development. The result is like a Photoshop "unsharp mask" - a contrast increase along the edge that is perceived as a "sharper" edge. Fuji specifically added chemicals to the emulsion layers of Velvia to enhance edge effects and apparent sharpness in regular E6 processing.
What WILL decrease film speed are "fine-grain" developers (Microdol-X, Perceptol) - developers containing extra silver solvents, often sodium sulfite . At the same time that the developer is creating metallic silver, the solvent is eating away at it, keeping the grains small (but also reducing the amount of silver overall, thus a thinner negative, unless more exposure was given up front = lower effective film speed). And also reducing edge sharpness as the edges are chewed up by the solvent.
Rodinal is the diametric opposite of "fine-grain" developers. It contains no solvents at all, thus producing full grain, full speed, and high edge sharpness. Some people even add sodium sulfite to Rodinal to soften its acute grain tendencies. Extremely high dilutions of Rodinal may mean the developer poops out in BOTH highlights and shadows between agitations (not enough molecules) - but that just means a longer time or a bit more agitation should be used, not that Rodinal itself somehow reduces film speed.
Most developers (D-76, ID-11) are compromises that fall in between Microdol-X and Rodinal, with a little sodium sulfite to keep grain smooth, without losing too much speed or acutance.
Edited by adan, 08 December 2012 - 19:27.