Rain, 
 
I know you don't use Nebula but you might be interested looking at this Alex B manual as a reference with more frequency and THD responses.

Although I haven't tried the AC202 Nebula 3rd party tape/ sat programs are by far the better tape emus out there. 

http://www.alessandrobosc...turators_FX_Manual.pdf

I'm actually re-demoing Nebula these days - I may very well end up buying it.

Thanks for the link!

Good idea, Steve, testing for aliasing. I ran a wide sweep and then a 10KHz sine against each of my test plugins. There is indeed significant aliasing with FX Tape Sim,  ReelBus  and TesslaPRO (in that order, worst to best). Ozone 3 and 4 had no aliasing.

Bit,
How could I do something similar with the test setup I used for the video?
As I now have the Kramer it would be interesting to see exactly what you experienced?

BTW: What metering program are you using with this?

Thanks,

I'm using SPAN. I just duplicated your test, but with a 10KHz sinewave. 

Given that these effects introduce mostly odd-order harmonic distortion, and the 3rd harmonic for 10KHz is 30KHz, which is illegal at 44.1KHz or 48KHz, we'd expect a well-designed tape sim to have no visible effect on a 10KHz signal lest they introduce an illegal frequency component and resultant aliasing. 

I would expect that to be accomplished via a combination of oversampling and filtering. Each of the plugins I tested do appear to be applying filtering. ReelBus does use oversampling, apparently fixed at 2x.

10KHz is probably not the best test frequency to see aliasing. Something like 11KHz might show the effect better.

I re-ran the tests with a 12KHz test tone, which does show the aliasing effects better, although I suspect an odd frequency like 11.7KHz or something might show it even better.

FX Tape Sim shows really nasty aliasing. Seeing this, I will never use this plugin without a LPF before it, if ever. ReelBus does much better, with the alias components 60db down and probably not audible in a real-world application. TesslaPRO was far and away the best performer of the three. Once again, Ozone 3 and 4 exhibited zero aliasing. Those guys know what they're doing!





I am using the "old and rusty" preset on ReelBus; aliasing was lowest with the "Hybrid" preset. 

I could effectively eliminate aliasing in TesslaPRO by lowering the Drive parameter. The distortion effect remained sufficiently audible even after reducing the drive. There seems to be a sweet spot where aliasing kicks in. The same trick doesn't work as well with ReelBus; lowering the drive until the aliasing goes away also kills the desirable harmonic distortion.

I'd love to compare these free and cheap tape sims with the more expensive products such as the Kramer, UAD and Slate VTM.  And I'm also working on a way to compare transient responses. Any ideas?

Bit you really have got me going here ;^)

Stay tuned to this station for further developments.

I've downloaded this one for fun: Tapehell2. It's purportedly a cassette tape emulator. 

I've heard of people running full mixes out to a cassette recorder and back in. I've mixed to many a cassette back in the day. It does make mixing easier because the tape almost mixes itself. Might be fun. What's not to like about a plugin that offers "Soft", "Hot" and "Hell" options?


On another note, I've been reading Jeroen Breebaart's blog where he talks about the development of ReelBus. To understand tape sims you have to start with an exploration of tape limitations, and these posts have a lot of good information about tape artifacts. 

One paragraph was of particular interest in the context of recent posts here, regarding intermodulation distortion with the bias oscillator, something I'd never thought about! Turns out that one of the spikes I read as aliasing isn't really aliasing, but rather part of the simulation. Not only that, it can be altered using the "Overbias" control. Try rotating that control while watching the spectrum with a high-frequency test signal.

Bit,
When you were trying Ozone were you using the Exciter module?

Yes. One band, "Tape" mode.