Upload & Sell: Off
I learned flash using two single power flashes "on the fly" shooting weddings. It's actual very simple if it is approached systematically because distance = power and it's just a matter of finding the distance needed for each f/stop when flash is used directly.
For example, 8ft is a good distance for people photos because it makes the near/far perspective of the face in the photo similar to the impression by eye. Shoot closer and the nose will start to look bigger/ ears smaller than what seems "normal". The next variable is f/stop which controls flash exposure and DOF. f/8 at 8ft. creates enough DOF to keep entire subject in focus. The final variable is how much manual power is needed from 8ft. to correctly expose the face and white collar. Try different power until finding the setting which exposes the face and white shirt collar correctly. Let's say it is 1/4 power. It's will always be 1/4 power at 8ft. in a similar room (the amount of bounce spill will affect exposure.
It's similar with dual flash. Shoot from 8ft with fill on bracket. If off camera key is at 8ft key=fill (assuming both lights are identical), but key overlaps fill like this:
1:1 Even fill from bracket @ 8ft
1:0 Key overlaps creating highlights from 8ft.
2:0 The highlights created over fill reflect 2x more light.
The H:S ratio convention assumes key overlaps even fill so 2:1 is the lowest integer ratio under it. It produces shadows which are a bit lighter than average room lighting and is a good ratio for making wrinkles disappear by making the shadows they cast lighter than "normal".
We shot candids at receptions with a 3:1 ratio. With the identical flash scenario shooting from 8ft with fill on the bracket you get a 3:1 ratio with darker shadows by moving key closer to fill. How close? The inverse-square law predicts that. If fill is at 8ft, the key will be 2x it's strength when it is 5.6ft.
So in the dual flash scenario when shooting at 8ft we move the off camera flash to 5-1/2 feet. How do you measure both distances? Well in my case my arm span is 70" so I stretch my arms from flashhead to tip of nose. Then I take three measured paces backwards to 8ft.
1:1 Even fill from bracket at 8ft.
2:0 Key light at 5-1/2 ft is 2x brighter (incident where it overlaps fill)
3:1 Reflected ratio is 3:1 - highlight reflect 3x more light than shadows.
Finding exposure is identical to one light, except you change power on both lights. For example start with both flashes at 1/8 power. The ratio will be 3:1 but the result may be underexposed. If so try 1/4, 1/2, etc. until finding the power that produces correct exposure.
I shoot speedlight portraits that way with my 580ex flashes and DIY diffusers. Shoot from 8ft, key light at 5-1/2 ft., aperture f/8, power on both flashes at 1/2. The 3:1 ratio perfectly matches detail in a black suit and white shirt to the sensor of my 20D / 50D. Regardless of the race / complexion of the face in the middle of that range it is reproduced accurately in tone with highlight shadow contrast which seems unremarkably "normal", how you'd perceive them if you met them in average room lighting.
When use M and ETTL?
As mentioned above M is simple if used systematically - lights at same distance, direct not bounced, repeatable distance / f/stop / power combinations. Bounce the light off the ceiling and the distance = power does not apply.
TTL (through the lens metering) in the generic sense (not Canon flash mode) reads light at the camera and adjusts flash power dynamically until the ideal amount reached the sensor as close as the metering can guess. The Canon system introduced in 2004 uses many metering zones and can make much more sophisticated guess than one that averages the entire sensor.
When you press the shutter button Canon E-TTL II flash metering first takes a baseline ambient reading separate from the one used for setting aperture / shutter, then fires a separate preflash with each group A, B and C if in play. (the 600EX can control 5 groups). The multiple metering zones allow the flash logic to know where key overlaps fill and where both overlap the ambient by subtracting the ambient zone map from the pre-flash maps (which also include the ambient. It's a very clever way to sort out what the flash is hitting and changing when mixed with outdoor lighting.
So if you figure out how to use it ETTL mode on the flash (which is actually E-TTL II on all bodies after 2004) can do a pretty good job of sorting out exposure when flash is used in modifiers, bounced, etc.
First set A:B ratio to 1:2. Master defaults to A so set Slave to Group B. Why 1:2?
The Canon ratio of 1:2 tells the Slave to put 2x more light on the subject (incident) which when it overlaps with the Master A fill creates the same 3:1 H:S ratio described above:
1:1 Even fill from the Group A master
2:0 2x as much light (incident) from Group B slave
3:1 Reflected ratio: highlights 3x brighter (1.5 stops) than shadows
So set A:B = 1:2 at least as a starting baseline and you should get a full range of detail in suit / shirt of flash lit foreground subject. All you need to do is adjust FEC until you see the white collar start to clip (black out) in the playback then back down FEC by one click (1/3 stop) to put the shirt 1/3 under clipping.
The metering zones are relatively large on entry level / intermediate cameras so they don't "guess" as well as the pro bodies with more evaluation zones. For each new scene you'll need to adjust FEC.
For example if you were shooting 50 couples at a party with ETTL the difference in clothing reflectance will affect exposure and you'd need to adjust FEC by trial and error with several exposures for each new set of subjects.
By comparison with M mode described above you set exposure for the first couple and since there is no metering changing power the same setting will work identically for the other 49.
So use M for static lighting set-ups and ETTL when you are running around shooting and you'll getting the best of both.