Is "normal record mode" broken?

Can anyone reproduce this when recording audio? -set record to normal record mode - snap the play cursor to any bar, -Press record (loop set or no loop doesn't matter) -stop recording (doesn't matter if it's looped round and added takes) afterwards when I zoom in and look at the newly recorded audio items start point it is way of the bar's start point and seems to be different each time mostly in front but sometimes behind. I've measured up to 3ms out. (need to look at this as it seems it might be sound card related) I believe it is still in time as you can slide the items start back to the bar without pulling the whole item back suggesting that the audio is recorded in time. If this is reproducible, This is really bad because without relative snap this means when moving the item it will snap and move your item out of time. Hopefully it's just a option I have missed.. :) I will test more today, also Time selection record works as expected.

Sorry, I have no idea what might be going on there and I can't repro it here. The recorded item starts at the grid no matter what. :(

Not reproducible her, either (3.06). I recorded noise from my sound card, and the item start position is exactly on the beat where I expected it. What does the item start position (on the Media Item Properties dialog) say about your item?

Thanks for trying. Done some more tests and it seems that maybe the lynx l22 does not report it's record offset correctly. The offset is not random as i stated either its always the same just had pdc compensation on the track. With it off it starts just before the bar by maybe a few samples but with it on its 3.8ms (my output latency for 64 buffer size) to the right of the bar and it seems in "time selection record mode" it's still out but the a bit of the missing bit at the beginning is filled with silence then 3.8ms the audio starts. Will try and auto adjust it but you would of thought lynx would report their latency correctly? :(

I think there's a misunderstanding - it sounded like the item itself doesn't align to the grid. If the contents of the (otherwise aligned) item are shifted, adjusting the input offset may help you. Maybe just use the project Billoon posted in the other thread.

Try the same thing but set your tempo to say 102 or 143 (the two tempos I tried) and also from a few bars into the track I have found that the items are only BAR accurate (with PDC off and auto adjust latency for aduio driver is on) WHEN you have tempo set to 120. If reproducible then this seems to be related to the sample rounding system that reaper employs due to its re-sampling thing. If you guys can try it at a different tempo than 120 (if you haven't already) then that would be a great help as the sample position is different depending on the tempo (but always behind the bar unless PDC track compensation is on. trying his latency test now.

Sorry, no difference with 90 or 150 bpm.

Thanks for trying again. So what might be happening with me then? ;) also, I created a sine wave in a separate program that is definitely 44100hz at 24bit to use for a phase test (doing the same classic loopback test)and I found that no matter how I aligned the newly recorded file up out of phase and at even every Db level (upto 4 decimal places) still couldn't get them to cancel (I know it was output through my converters but not even the sine wave would cancel leaving the quantisation and distortion of my converters). What was more confusing is that if you look at the sine wave that was created in this different program and line it up with the samples on the time line then by the end of the item it is completely out of sync with the sample time line, It should surely be one item sample per time line sample? (is this a redraw error as tried a sine from an independent 24 bit wav test file too and got the same results) I know this sounds "anal" but it's getting frustrating :( Schwa any insights on this?

Latest findings..