Since as far back, long before X1 I have been mixing and mastering simultaneously.
 
I have been using Cakewalk since the very first MS DOS version.
 
When I had my quad core computer I mixed and mastered simultaneously. If my mastering plugins needed oversampling I would turn that on high right before rendering.
 
I would rather have less mastering plugins than have to mix down my mixes and load them in stereo to master.
 
Really I only use a very few plugins for mastering.
 
I use a multi-band compressor, LP EQ a stereo analyzer and widener with correlation trace and two expanders.  
 
These items are very processor hungry.
 
When I was using my quad core they would often bog down my system and I would have to disable them periodically to complete a mix.
 
I will first track my songs without the mastering plugins (or have them disabled) so i can get maximum latency.
 
The idea of mastering a stereo rendered mix is as foreign to me as could be.
 
When I need to adjust a vocal inflection that is too high, having to re-render my mix would be out of the question. It is an absolute must that my mix is there so I can work on them simultaneously. Cakewalk fits the bill on these multiple processes.
 
Now that I master on an Intel I9 24 thread CPU I can turn on all of the tube enhancements, console plugins and have stereo analyzers, effects galore on each track and set the oversampling of my expanders to the highest settings and only register about 2% usage on all 24 threads.
 
I can tweak my mix or the mastering plugins at will and never have to re-render a stereo mix. Since I have been doing it this way all along there has been no learning curve other than having more stuff going at once.
 
I have always been one for simplified mixes.
 
Things like parallel reverbs and delays I find only cloud up a mix and make it too complicated to mix.
 
I just put the amount or reverb and delay I want on a track and EQ the effect. The vocal will stand out enough if you compress it right and use melodyne to edit the amplitude of each note.
 
Carve out a place for the vocal in the mix with EQs and use a sidechain compressor so your vocal and its effect cuts itself out of the mix.
 
Add background vocal to this side chain also.
 
I try and get a mix into the ballpark before I turn on my mastering plugins.
 
Some people may disagree with me about parallel reverbs and delays but I find the added complexity of using and setting them to be consistently more of a detraction than an addition to my song.
 
I remove the parallel reverbs and simply add the reverb to the tracks and suddenly it fits into the mix right. I don't know if it is a phase issue or what but I prefer not to use parallel effects.
 
A simpler mix is easier to nail.
 
Thoughts, comments and criticisms are welcome.

No doubt many computers now are capable to mix/master all in the same program. I have done this both ways.
 
I see some wisdom though in having two processes, even though it might not be necessary. If a person mixes and masters all at the same time there is a tendency to play with the mix while mastering it.Some people will second guess themselves to death. I mean, if you liked it as a mix why play with it when mastering? They should be two different processes IMO. The compartmentalization is an asset in that regard. If forces you to ONLY work on the master. There are so many variables involved in trying to work both processes that it would be easy to get something more wrong than right.
 
Getting a quality export file is another reason to concentrate resources on only the master and related export file quality.  This is why Izotope and TRacks offer the option to use their mastering programs stand alone. Izotope particularly have a way to hear the file as an mp3 and make the master best for that format. Even a more powerful computer has less to do if all resources are dedicated to only the plugins used in mastering so potential problems are avoided.

They are two different processes.  What you have described is really mixing.  Those plug-ins can be used for mixing as well as mastering.  Just because you put something on your master bus or your 2 bus, does not make it mastering.
 
The main purpose of mastering is to prep a song for release.  Slight adjustments to EQ and compression to account for what might happen when the song is rendered to MP3.  It is also used to set loudness and match loudness if you are mastering a collection of tracks for distribution.
 
Regarding "parallel" reverb and delay, I am not sure why they don't sit in the mix for you.  The first thing I would check is to make sure that the mix knob on the plug in is set to 100% wet.  Using these FX on a bus rather than on each individual track saves resources (may or may not be an issue with your system) and allows you to create a cohesive reverb and/or delay effect and save time when adjusting them.
 
JMHO.  If your process and workflow is working for you, there is no reason to change.
 
 

I, too, combine the mastering and mixing processes. That's because mastering changes the mix and changes to the mix affect the mastering.
 
However, I adhere to the philosophy that the less mastering the better. A good mix requires very little post-mix massaging. That means waiting until the mix is 95% done before applying any effects to the master bus. Now you can make your final mix tweaks, informed by and in the context of what the mastering is doing to it.
 
Applying master bus EQ, compression and limiting too early in the process is a recipe for over-processing. If you're applying more than 1-2 dB of EQ on the master, then you've started too soon.

I think a lot of agile programming principles apply to mixing and mastering.

Fast cycles including everyone helps. You may be one person, but I am suggesting involving the band early and often.

More ears are often helpful. The sooner you discover what is wanted that you didn't expect the happier you will be.

Final bounces no one wants are expensive and avoidable.

I prefer not to do the two things at once.  There have been some interesting articles about some famous albums from the past where they mastered straight after mixing and regretted it after.
 
I see the two process as quite different and it is healthy to have a little time between them.  Especially if you are mastering your own mixes.  And despite what they say on this, you can master your own mixes.  It's great to live with a mix for even a few days.  It can be remarkable what you will hear and change for the better.  The time between mix and mastering allows that experience of hearing your music fresh for the first time.  Very  important. In the morning after a good rest.  EQ decisions for example a week later will be much more subtle as compared to making those same EQ decisions after a long 10 hour mix with slightly fried ears. 
 
VU meters are very helpful and essential in the mastering process. Real or imaginary.
 
Some gentle compression on the master buss is not quite the same type of thing as the compressor in mastering chain a week later.  It's OK as well to use some stereo buss compression.  Compression ratios multiply though.  So if you are wanting an overall mastered compression to feel like say 2:1 then if you used 1.5:1 on the stereo mix down buss, then only 1.3:1 later in mastering will create a total overall compression ratio of 2:1. 
 
Aiming for a stellar mix at a ref rms even of say -14, or -18 or -20 will suffice.  Pre mastering.  This allows you to nicely do an overall EQ over the whole mix.  This will not be drastic.  There are a few things the mastering EQ will do however and these are:
 
 
cleaning up bottom end
shaping the top end air eq slope preferably downward
pulling out some slight thickness around 300 Hz
nudging ever so slightly the overall mid range oh so very slightly
add some gentle air crispness to the whole mix.  
A little bit of overall rms gain can be added in the EQ as well.
 
The mastering compressor when used well just adds a professional dynamics control sheen to the mix.  Plenty of punch still but taming things and reigning things in here and there.  Makeup gain should be set for the average amount go GR that is taking place.  These compressors need to dance lightly over a mix.  As well as makeup gain, another 2 or 3 db in overall rms gain can be added here as well.
 
Note: if we started off with a -18 dB rms master we have just lifted the overall rms gain now by 2 dB say in the EQ and 3 dB in the compressor.  Neither of which was working hard at that aspect. i.e. just adding some gain. -18 is now sitting around -13 dB rms.
 
The final limiter just adds as much rms gain as you want to take its to final level.  Another 3 dB of limiting makeup is going to put the final mastered product at -10 dB rms.  It will be loud, punchy, transient and basically kick ass by about this point.  The final level a mix reaches now is more dependant on the streaming sites this music may be coming from.  Adjust accordingly and use a LUFS while doing it too in the final stages.

Great points all. Lot's of wisdom in waiting a bit before attempting a master, mastering seperately and in looking at rms. 

Jeff Evans
I prefer not to do the two things at once.  There have been some interesting articles about some famous albums from the past where they mastered straight after mixing and regretted it after.
 
I see the two process as quite different and it is healthy to have a little time between them.  Especially if you are mastering your own mixes.  And despite what they say on this, you can master your own mixes.  It's great to live with a mix for even a few days.  It can be remarkable what you will hear and change for the better.  The time between mix and mastering allows that experience of hearing your music fresh for the first time.  Very  important. In the morning after a good rest.  EQ decisions for example a week later will be much more subtle as compared to making those same EQ decisions after a long 10 hour mix with slightly fried ears. 
 
VU meters are very helpful and essential in the mastering process. Real or imaginary.
 
Some gentle compression on the master buss is not quite the same type of thing as the compressor in mastering chain a week later.  It's OK as well to use some stereo buss compression.  Compression ratios multiply though.  So if you are wanting an overall mastered compression to feel like say 2:1 then if you used 1.5:1 on the stereo mix down buss, then only 1.3:1 later in mastering will create a total overall compression ratio of 2:1. 
 
Aiming for a stellar mix at a ref rms even of say -14, or -18 or -20 will suffice.  Pre mastering.  This allows you to nicely do an overall EQ over the whole mix.  This will not be drastic.  There are a few things the mastering EQ will do however and these are:
 
 
cleaning up bottom end
shaping the top end air eq slope preferably downward
pulling out some slight thickness around 300 Hz
nudging ever so slightly the overall mid range oh so very slightly
add some gentle air crispness to the whole mix.  
A little bit of overall rms gain can be added in the EQ as well.
 
The mastering compressor when used well just adds a professional dynamics control sheen to the mix.  Plenty of punch still but taming things and reigning things in here and there.  Makeup gain should be set for the average amount go GR that is taking place.  These compressors need to dance lightly over a mix.  As well as makeup gain, another 2 or 3 db in overall rms gain can be added here as well.
 
Note: if we started off with a -18 dB rms master we have just lifted the overall rms gain now by 2 dB say in the EQ and 3 dB in the compressor.  Neither of which was working hard at that aspect. i.e. just adding some gain. -18 is now sitting around -13 dB rms.
 
The final limiter just adds as much rms gain as you want to take its to final level.  Another 3 dB of limiting makeup is going to put the final mastered product at -10 dB rms.  It will be loud, punchy, transient and basically kick ass by about this point.  The final level a mix reaches now is more dependant on the streaming sites this music may be coming from.  Adjust accordingly and use a LUFS while doing it too in the final stages.

 
You kind of bolster my point....
 
Not just a week later but a year or more later I can go in and edit a mix because it is not frozen into a stereo track that cannot be readily edited.
 
When mastering plugins are simply added over an in progress mix then one is free to edit the mix long after the mastering process has begun. So it gives a much larger window of time for the ears to cool down and perceive the mix with rested hearing.
 
The mastering bus can always be bypassed.
 
Once a mix has been rendered the process of re-rendering the mix is just a pain.
 
Having the luxury of tweaking the mix and lowering a vocal line or increasing the bass guitar etc... negates a lot of mastering trickery that is often crude compared to just editing the source material.
 
The objective is getting the best final product as possible, using crude mastering techniques, notch filters, bass boost or cuts, deessers versus going into the mix and correcting mixing problems insures a better/cleaner final product.
 
Going into the mix and tweaking various problems makes the mastering job less cluttered with effects that affect the mix globally. It is the lipstick on a pig problem, why not go back in and change the animal than try and dress up one that is not as pleasing?
 
As you say, it can take a week sometimes to realize a mix has problems. well it can take longer than a week to realize a song needs more work. To go back and add more instruments then do a re-rendering and have to apply all of the mastering tools once again is simply not the way to go.
 
A much better approach is to have the mix there and ready for any surgical procedures while leaving the mastering in place. Future edit-ability is the whole point of working in the digital domain across an entire project.
 
Still it is a to each their own methodology but, I personally would never even consider working with a stereo rendered track over working on top of a live mix.
 
If you need "air", add it to instruments in the mix, then you are not adding air to other tracks that become too sibilant when EQ high end boost is added to them. 
 
As for adding overall compression, that can be added to a live mix just as easily as to a rendered mix... One difference is if the compressor colors a certain track by expanding it too much you can go into the mix and lower that one element or reverse EQ that track to compensate for the coloration of the mastering compressor. 
 
:)

Its your workflow and if it works for you then great.  But it sounds to me like you have trouble finishing the mix process.
 
I mix as if someone else will master (even if I do it myself), which means I try to fix all problems in the mix before I start mastering.

RexRed
 
 
You kind of bolster my point....
 
Not just a week later but a year or more later I can go in and edit a mix because it is not frozen into a stereo track that cannot be readily edited.
 
 

the issue here is not being able to make a decision and roll with it.
 
once the mix if finished, it is finished.
finalize your decision, and move on.