Rendering in T is not as perfect as I thought

[AndrewSimon]

Try this.
Hit a fast note on the piano.
Put a very long reverb tail on it and wait until the tail dies completely.
Sounds great.
Now render the track and listen again.

What I hear at the end of the tail is almost like a noise gate closing up. Rush of hiss and gate closing.

Now I use a recorder plug like "TapeIt" and let it record the same note while T is playing.
The result is perfect.

Any ideas?
Do you hear it as well?
I am sure I did everything correctly.
I render at x1, both "TapeIt" and T are rendering at 16bit 44.1Khz .... so why the difference?

[.jon]

Ok, this is interesting...

Can you tell more about this test:

- what is the piano?
- if you recorded it, what were the sample&bit rates?
- are you rendering to a lower bit rate, ie what's your setting when playing?
- what is the reverb you are using?

[AndrewSimon]

OK I recorded a little guitar sample.

Here is the render (at x1):

http://www.andras-shimon.com/MUSIC/Render.mp3

Here is the live capture with TapeIt:

http://www.andras-shimon.com/MUSIC/TapeIt.mp3

[AndrewSimon]

I changed the piano to guitar , same results.
I changed the reverb, same results.

The recording is 24bit 44.1

Both the "render" and "TapeIt" is 16bit 44.1

I also experimented with settings like:
"add low level noise to avoid demor..."
Changed latency in ASIO
but same results.

[AD80]

I actually dont hear it. But I opened both files in AA and the first thing I noticed was that the TapeIt file had about 4 extra seconds at the end compared to the Render. Maybe thats whats cutting off your reverb tail. Also the rendered file has a noise floor of -90db's, and the Tape it version -infinity. So there might be some bit depth difference between the two.

[Warmonger]

The "add low level noise" isn't the same as dithering. You should check it again, once with dithering enabled, and once without. Also, you might want to try exporting at 24 bit, and comparing then. I have to admit, I also can't hear the difference, but I don't have great ears, nor do I have a good listening setup.

[sonicsmurf]

Cannot hear any difference. Compared the samples in Audacity, no big difference visually.

I doubt anybody can hear a difference?? (I played it so loud my neighbours must have fallen out of their beds

[FoxV]

(I played it so loud my neighbours must have fallen out of their beds

...that got a good chuckle out of me! Hey...you gotta do what you gotta do! (The neighbors will surely understand!)

[zeoy]

Did you check the "remove silence ..." setting in Tracktion's export?

[semiquaver]

Cmon folks - it is plainly audible (use a pair of headphones). And it is plainly wrong. Yikes.

[AndrewSimon]

Wow I am supprised some of you don't hear the difference.
Try headphones with high volume.

Did you check the "remove silence ..." setting in Tracktion's export?

Nothing is checked except "render at x1"

Anyway I am sure it has to do something with the 24 bit going to 16 bit. It's no biggie just one of those things that make you wonder.

TapeIt file had about 4 extra seconds at the end compared to the Render.

Well it's "TapeIt" so I have to hit "Record" and "Stop" manually while tracktion is playing and I am a little slow.

for more info on TapeIt
http://www.kvraudio.com/get/758.html

[rpc9943]

I don't hear any difference... I must be retarded

RonC

[semiquaver]

what about dither? can you try it *with* dither and see if it solves the problem?

[semiquaver]

Folks - cut the top off and then crank up the volume - it is plainly there. If you have a monitoring setup that doesn't let you hear this you should be worried!

I will do some experiments later today myself but in the meantime Andrew please try dither and 24 bit export and lets see what we can learn. It could be that a truncated tail sounds just that bad without dither - what surprises me though is the noise - I would expect a kind of distortion or buzziness but...

[rpc9943]

my monitoring setup is fine. I simply am not hearing a difference, if there is, its subtle... At least for me subtle enough to not flip out about it.

Besides, T2 probably fixes all this

RonC