-
-
Does increasing the Preview Buffer size change that delay? It is one of the preferences.
Craig"Who put orange juice in my orange juice?" - - - William Claude Dukenfield -
No effect
MarcComment
-
Does sample rate and/or bit depth change the delay?
Craig
"Who put orange juice in my orange juice?" - - - William Claude DukenfieldComment
-
We are wondering what is different between your system / setup / test and Dan's. Obviously, it's a bug, but where do we look is the question. I suspect it has something to do with buffer size internally with the CNF. Does fft size impact the delay? Larger fft size (in theory) would require longer buffers.
Craig"Who put orange juice in my orange juice?" - - - William Claude DukenfieldComment
-
Marc - Something's not quite right with the preset. When I loaded it (It's called 78 Electric), The file conversion filter is set to Mono (L+R). The rest of the multifilter looks the same as your image. But when I run the multifilter on a 15 minute long stereo file, I end up with a mono file, with no delay apparent as an echo. Is this the right multifilter?
Both files end at the same length of time, and have peaks at the same spot, so I don't seem to be getting the delay. This is with Forensics 10.02.
I get my presets mixed up a lot.Last edited by Dan McDonald; 12-12-2016, 12:12 PM.Dan McDonaldComment
-
Dan,
Thanks for trying out the multifilter. This filter should produce a mono file as its for 78 records. I use a stereo pickup and remove the vertical component with the L+R.
When I use the filter on my system, it's in classic edit mode. So, the destination will have some noise event left over from the multifilter. When I compare the noise event in the source with the destination, I do not see the two directly under each other. Thus, when I try and manually fix the noise in the source, I often can not find it in the destination window with this multifilter.
In past versions of DC products, I could see the original noise in the source and then directly underneath the result I would see the noise in the destination.
On my system now, the two events are not occurring at the same time.
Also, if I remove the three CNF blocks, the events are now related in time. This could be related to my computer system as the processing time for the CNF blocks is long.
MarcComment
-
Just a quick note to readers - the L+R will remove the vertical component, but will remove almost everything on Edison vertical cut records (or any hill and dale vertical cut). So, that setting is useful only on lateral cut 78s. If you put just one CNF in and leave the other out, does it still loose time alignment? If you then reduce the fft size of the remaining CNF, does that change the situation?
CraigLast edited by Craig Maier; 01-20-2018, 01:49 PM."Who put orange juice in my orange juice?" - - - William Claude DukenfieldComment
-
Craig,
No...having the CNF in the filter causes a delay.
MarcComment
-
No matter what its settings, mode or whatever? Nothing more specific? Sounds like (for your system anyway) the CNF needs more global buffer space, but I am really not sure - just an educated guess.
CraigLast edited by Craig Maier; 12-13-2016, 08:11 PM."Who put orange juice in my orange juice?" - - - William Claude DukenfieldComment
-
Craig,Originally posted by Craig Maier View Post
I did try some permutations on the experiment, but couldn't find anything conclusive. If I remove the CNF blocks, the time offset, if present, is very small. If I add any CNF blocks, then the time offset is more than I can use for my work.
MarcComment
-
Does the greater the number of cnfs result in ever increasing delay?
CraigLast edited by Craig Maier; 01-20-2018, 01:50 PM."Who put orange juice in my orange juice?" - - - William Claude DukenfieldComment
Tweet
Comment