TimeFactory 2 has entered the ProTooler kingdom, and it’s said to be good too. Let’s find out…

The process

About half of the TimeFactory 2 window is the “input files” section. Simply click n’ drag any amounts of files you want to preview or modify to this. There you’ll see some specs such as length, format and play time.

The action happens in the “Time/Pitch…” section though. When I first looked at it I wasn’t very impressed, you can change length (in percent), pitch, formant and choose whether the change should be applied with transpose or natural pitch. Fortunately there is also an “Expert settings” window that you can click to show. It stayed open the entire time for me since this is where you’re able to change duration, length in frames as well as BPM.

Algorithms

There’s four algorithms to choose from: MPEX2, MPEX3, DIRAC and ClearScale. I’ll give you a short explanation or each of them. The MPEX2 was the algorithm used in TimeFactory 1.x and is mainly included for compatibility purpose, in other words, that won’t get much time on the field. MPEX3 is the new algorithm said to be “used in many popular products on the market”, which makes me wonder why I should buy this product if I already have one that uses it. It’s also said to “allow a crystal clear 200% time stretch”. We’ll put that to the test later.

DIRAC offers a variety of settings and perhaps the widest range of parameters that can be tweaked to work well with a particular type of signal… Or so the manual says. it also suggests to use it on drum loops because it doesn’t alter the timing structure, and also for doubling vocals or altering the timbre or gender or a voice. ClearScale was developed by the same people who developed DIRAC but doesn’t have any quality parameters since it auto-adapts to the signal. Should be good for beginners.

The cop show theme test

The first test was made on an entire mix, a theme sounding like an old cop show or something like that. It’s 32 seconds long but should only be 30, two seconds change in duration, approximately around 6-7% of the time, shouldn’t be a problem. I used the MPEX3 algorithm and set it to poly because it’s an entire mix and not just one voice. There are several quality settings you can choose here as well. Of course it worked well, that was until I decided to import it into Pro Tools to compare waveforms. Look at the picture below, do you notice how it have introduced a short pause at the beginning?

I was very surprised about this and I’m not sure about the reason for this but I did notice one thing that might have something to do with it. In the expert settings duration is said to be an even 00:00:32:000, that is 32 seconds blank. When I preview it play time is said to be 00:00:31:904. Could it be that it is compensating the time difference with a pause? I decided to try the same thing with the DIRAC algorithm (standard setting, good quality), and guess what, no pause in the beginning! Could it be that I had done something wrong? I tried the MPEX3 setting again but the same pause appeared in the beginning.

Anyway, I continued with the test and decided to change the MPEX3 to 20 seconds which of course made it play REALLY fast, but it still sounded pretty good. The Digidesign TC/E sounded horrible and the DIRAC sounded even better than MPEX3 to my ears.

In what other way could I beat this mix to an unrecognizable mess? By stretching it back of course! Stretching back the MPEX3 revealed some artifacts, while stretching back the TC/E file brutally murdered it. When I did the stretch back with the DIRAC algorithm, I choose the ‘best’ setting which the manual says can be very slow, on my MacBook C2D it didn’t run particularly slow though. While it included some artifacts, it did sound even better than the MPEX3 stretch. But seriously, if you need to stretch an entire mix back and forth more than 30% you might want to consider other things, like rerecording or at least keeping backup copies of your original session!

The Paul Gilbert test

I was curious about the MPEX3 Transcribe setting. The manual says it can produce a 2x time stretch. I decided to put it to the ultimate test and have it do a 2x stretch on one of guitar shredder and speed freak Paul Gilberts songs. If the notes still picked (or shredded) by Paul Gilbert still have the same pitch we will declare it a winner.

So how did it fare? It took some time to process the file but the results was very good sounding. I agree with the manual on this one, it should work very well for learning solos and things like that since you can get it down to half speed. A nice feature indeed.

What else?

You can decode files, that is make a raw PCM file out of an MP3 for instance. You can also sum tracks together, for instance five mono files into one 5.1 surround file. Both of these features are nice to have within the program so you don’t have to rely on a second or third application for this. Another very useful feature is that you can simply double click on any file in the window to locate it in Finder. Very helpful. I have one more complaint though, the locator bar that shows where in the particular audio file it’s playing is somewhat annoying. You can’t just scroll to a place and then click play because it will start from the beginning anyway. What you’ll have to do is click play and then scroll to the position you want to find. Small thing, I know, but annoying anyway.

Conclusion

My opinion is that the new algorithms sound good and can be used to either time stretch a voice or an entire mix. The program itself is very easy to use and if you don’t understand it then there’s a very useful manual included. But it’s not all sunshine, the pause introduced in the beginning of the stretch when using the MPEX3 algorithm is a major letdown, and unless there is something I don’t understand here this is something Prosoniq really need to address. The DIRAC algorithm is my favorite though, and doesn’t have that problem. It sounds good and it seems to me like it really keeps the timing like the manual says it does.