I'm honestly sorry to offend- with what I choose to try to do, and the way I do it, it puts me in awkward situations at times, and I've learned over the years that I must apologize for my manner, and to tread lightly, leaving it to others to perceive (or not) the merit of what I've tried to do.

I stand by my claim, but I hope I am doing so respectfully. This is why blind tests and such exist- to find the reality when some poor socially-inept fool (like me) manages to accomplish the surprising- and this is why I was so thrilled to discover this file-experiment going on.

I can do little enough really well- and surely interacting with people is one of those weak points- but recordings strip away everything but the sound of things, blind comparisons even more so, and given the chance I've always been able to do well with that.

Apologies for getting so excited, and I'm still looking forward to people listening- especially if anyone gets together a blind comparison using these files. I fear I've really put my foot in it again. I was just thrilled I was able to get it so close, having designed the product simply to be a generic, tune-able 2012 tape sim...

Could you listen to and comment on the samples, rather than talking about belief? Very interested to hear your and other peoples opinions, since there's actually something concrete here rather than just an empty claim.

Whether or not I get to hear listener opinions, I'd sure be excited if people can continue to update this thread with new products as they come out. It just seems like such a phenomenal third party resource, and the beauty of it is that since the reference files are publically available, with a bit of care in settings and level matching Gearslutz can have not only reference examples for what's out there but also the ability to do nulls and analyze what's being done in each product. That's a huge deal, and only possible thanks to Justin's efforts.

The fly in the ointment is that no two analog devices, much less tape machines, of the same model sound exactly alike and many sound different when driving different loads or being driven by different sources or used with different converters.

It's certainly legitimate to prefer the sound of one emulation over another but I think to suggest that any one is more or less accurate without having the original machine, tape batch and signal path that was modeled for comparison is extremely misleading.

This is definitely true. It's also why, at some point, I think I'm increasingly beginning to not care if a tape emulator is %100 accurate. As long as it's pretty close (90% to 95% like the Anamod or UAD) and does all of the things that I like about tape (freq. resp., saturation, compression), I'm a happy camper.

If the net result is that I get a sound that I really like, even if it is only 90%-95% close to tape, I think I'm beginning to really become okay with that. That's close enough for me. Hell, I'd say that there is at least a 5% to 10% difference between the sounds of different models of machines so that puts you right in the same ballpark as the difference between one particular machine and it's emulation.

And we're talking about the differences between the same machine. This doesn't even take into account the difference between different makes and models of machines.

Now if someone would just model the Ampex MM1200...

Agreed. Accuracy isn't really the main point of this thread. Just comparison.

And I think we have been able to show some specific ways that some of the tape sims succeed and/or fail to recreate certain aspects of what real analog tape recording/reproducing does no matter what machine/tape batch is used. One example - Brad's graphs which show some of the sims failing to recreate the HF dynamic behavior of tape.

But ultimately yes, it comes down to sound and work flow. Personally, I use Waves KMT, UAD ATR 102/A800, DUY Tape, and my Ampex ATR 102 whenever I feel they are called for, and hope to own an ATS-1 in the future.

I'd like to add that even though the various sims may model certain behaviors "correctly" and other incorrectly, the end result is what matters. I think in all cases you can use the products to achieve aesthetically pleasing and useful creative results. I utterly adore my ATS-1. Like I said I don't think I will ever send another track into the darn computer (or Radar) without it passing through that box first.

I still intend to do some analog realm tests with the ATS-1...just need to find some time. Still interested, Justin?

Brad

Yes, but I need to find some time as well!

I listened to the Anamod 351 and the KMT MPX files as the 351 is the card I'm always using on the Anamod (and I've also got the MPX). I'm quite surprised by the Anamod 351 file sound here. My Anamod 351 doesn't sound as dark as this one but much more similar to the uneffected sound. When I bypass the Anamod the difference isn't as obvious as here and there always seems to be an increase of high harmonics whereas here it's the opposite.

Interesting...what version is your 351 card? There should be a sticker or marking on the card I think.

Feel free to take Justin's dry sample and run it through your ATS-1 with all knobs at 0 and post up the clip for everyone to hear.

Brad

Use the "DA/AD Only" file for a truer comparison.

Bump for pre-echo

Without it, there would be a lot of phase shift at 20 kHz using the sampling rates currently in use. The tradeoff is perfect phase response for an anomaly in the time domain.

Sorry... I've been away... NAMM and stuff. Just getting caught up now on this thread.

Thank you M

So then what does pre echo sound like? Would you hear those higher, "pre echoed" frequencies earlier in the time domain than you would if you were hearing them naturally?

I guess, because of the more relaxed filters used for the higher sampling rates, you hear less of the pre echo when you use higher sample rates due to pre echoing also taking place way up higher in the frequency range? If so, am I reading you correctly that you seem to imply that there might be a high enough sample rate, even if currently not available, that could eliminate or at least sufficiently negate the effects of pre echo?

I don't really know anything about pre echo so I'm interested to hear your thoughts on how it relates to sample rate. I exclusively use 96k and definitely hear a difference for the better over lower sample rates so I'm curious how pre echo fits into all of that.

Yes... That's right except it will never go away entirely. It just gets shorter as the sample rate increases. Theoretically with a high enough sampling rate and some allowance of group delay distortion in the high frequencies, you could use a causal filter which has no pre-echo.

So you will hear it within any given frequency range, regardless of sample rate? The only difference will be how short it is within any of those given frequency ranges?

If so, I guess you're saying that sample rate will only change how long the pre echo is and not what frequency(s) it is present in?

It's not a matter of frequency range but it affects things like the perception of transients for instance.

Yes, the length.

What do you think a high enough sample rate would be to implement such a design and do you think such a design would be superior to the current, more common designs currently out there?

As for the current designs out there that do use pre echo, how well do you feel 96k does at negating the negative effects of pre echo? What about 192k? Are any "acceptable"?

Yes, that's what I was attempting to say but you stated it much more clearly. High frequency transients is where you would notice it most even though it is taking place across the entire frequency spectrum. Right?

This is interesting stuff. I'm always looking to better understand where the deficiencies in digital conversion occur and what causes them.

How long a pre echo are we talking about at 44.1kHz sample rate? At 96kHz? Thanks

Hey, I just found some pretty good info on pre echo in this thread:

Why Do We Love Tape So Much?

I'm sure there's more discussion about it on pages prior to and after page five but it was pretty interesting. Also, Dave chimes in quite a bit there as well.

Still, I'd like to hear some more about it...

I can't stand it which is why I only use tape. The sampling rate to have only 45 degrees of phase shift (like a good tape machine) at 20 kHz is about 1.23 MSPS if I've done the quick calculation correctly.

page 6 of this Intersil app note describes their decimation filter. I'm not sure they've given all the info to calculate the pre-echo length. Sometimes you can estimate it by the converter latency, allowing for a sample or two for overhead, the rest is probably the filters.

So we're talking about 44,100 or 96,000 samples vs 1,230,000 samples. That's quite the difference. I guess my computer could handle two or three tracks of that... heh

Yeah... It's doable. At least 8 tracks.

Maybe the newer computers coupled with SSD drives and LOTS of memory. I know mine certainly couldn't. How did you calculate that number by the way?

By the way, if the pre echo problem could be done away with, through whatever hypothetical means it takes, do you feel that digital audio would otherwise be up to task or are there other areas that you feel conversion is still inferior?

I suspect I'll abandon tape at that point.

Probably wouldn't need a computer... I'm envisioning a RADAR type system.

Get the tape deck. You'll have more fun with it.