Recording via optical-in

[whatboutbob]

Precompiled build with optical recording patch:

http://s41.yousendit.com/d.aspx?id=0XCI8CJMJU58F0EGPLQZ01OCE4

Please be aware that you may need to reset settings and enable optical out to get 'digital' recording working.

There's a reason the patch hasn't been commited yet...its still buggy.  Have a toothpick handy while testing because if you attempt to record while there's no signal (or if it loses signal mid-way thru) the recording will hang. You'll need to reset the unit.

Good luck, and report back your results.

[dwonk]

This is what I found regarding the SPDIF being properly recognized/captured:

You need to be feeding it a signal from your AD when you choose the source (digital) and then go to the recording screen.  If you don't, there will be no signal.  Also as whataboutbob has pointed out, the optical out needs to be turned on.

I was able to grab 3 minutes but the file header was all messed up.  there was a 29MB file but I couldn't open it.

Another thing, if the next improvements are to be made, the recording setting should be saved somehow.  Also, if the optical out has to be turned on everytime, it will be an extra step to engage the recording.  Maybe implement the optical out being on when the in is selected for the source.

[dgodwin]

I used the precompiled build from whataboutbob.  I was able to transfer a 1.5hr DAT tape to the iRiver with no apparent problems.  Again, the only overall issue I encountered was needing to have the dat player going when I went into the recording screen. 

[hellokitty]

Great !

Were you able to listen to the generated file ?
What settings did you use (khz,...) ?
Is there a recording limit ?
Where can I get whataboutbob's precompiled build ?

[MU4L]

Top o' the page  , Kitty ...

[dgodwin]

DAT was originally recorded at 44.1k  Played on a Fostex D5.  Rockbox directions were followed after installing the build:
reset cfg
optical output = on
recording source = digital
frequency = 44.1

Started the DAT

went to recording screen, and hit record.  Levels were bouncing, and everything appears fine.  I had to stop the DAT, and do a switch (the encore of the show was on another DAT) so I left the iRiver recording, to see what would happen.  It kept recording, and when I got the DAT going again, it didn't show any problems. 

I transfered the recorded audio onto my computer, and opened it in adobe audition, and it played fine, and looks normal. 

[dwonk]

Switched to the 4-24-2006 - 1600 build.

I am still getting issues having my H120 flushing the buffer to the HD.  The HD red light is on and the peak signal dies when this happens, then it locks up.  I tried this tim feeding a signal from my CD player.

[manno]

Ok, so I was able to test it out.... So far it seemed to work OK. I did a couple of short test files, one about 70mb in length and one about 40mb in length. They both came out ok EXCEPT that I fed it a 24-bit/48khz signal and it downsampled it to 44.1khz and trucated it to 16-bit. But it did flush the buffer to the HD ok. What's the possibility of keeping the sample rates (not to exceed 48khz) and doing at least 20-bit at the hardware max?

It's looking good here!

BTW, on my end when it flushes, the HD led stays on and it appears to lock but after a few seconds, it then picks up again.

[dgodwin]

Ok, so I was able to test it out.... So far it seemed to work OK. I did a couple of short test files, one about 70mb in length and one about 40mb in length. They both came out ok EXCEPT that I fed it a 24-bit/48khz signal and it downsampled it to 44.1khz and trucated it to 16-bit. But it did flush the buffer to the HD ok. What's the possibility of keeping the sample rates (not to exceed 48khz) and doing at least 20-bit at the hardware max?

It's looking good here!

BTW, on my end when it flushes, the HD led stays on and it appears to lock but after a few seconds, it then picks up again.

This pretty much describes my experience, but I've only tried 44.1k, so I can't confirm the downsampling.  I'm betting most would prefer no resampling regardless of bit rate (at least 48k or less) 

[manno]

Another test..... recorded about 53min of 24/48 stream. It still captured at 16/44.1 but after transferring the file and upsampling it to 48khz, the file was perfect (or at least no errors that I can see or hear). One cool thing though, even though it was downsampling to the HD, I was using headphones to monitor the stream and it was perfect!

[LinusN]

They both came out ok EXCEPT that I fed it a 24-bit/48khz signal and it downsampled it to 44.1khz and trucated it to 16-bit.

Using more than 16 bits will put a substantial strain on the CPU, since we can't use DMA in that case.

Rockbox does not downsample to 44.1kHz when recording. However, it always downsamples to 44.1kHz when playing back the file.

[mborus]

Also did some more tests yesterday.

I recorded one hour from a CD twice and compared the result.
As with the shorted test both files were identical. So not a single
glitch.

The only "problem" I had was starting the recording and the CD
at the same time, without missing the beginning.

BTW, the optical output is not working during the recording.

[manno]

They both came out ok EXCEPT that I fed it a 24-bit/48khz signal and it downsampled it to 44.1khz and trucated it to 16-bit.

Using more than 16 bits will put a substantial strain on the CPU, since we can't use DMA in that case.

Rockbox does not downsample to 44.1kHz when recording. However, it always downsamples to 44.1kHz when playing back the file.

I fed it a 48khz signal and the resulting file was at 44.1khz. I'll double check my settings again....

Isn't writing out 20 bits of data more of an I/O strain than a CPU strain? Is there a possibility of maybe testing it out?

UPDATE- I checked all my setting and made sure that it was being sent a 48khz signal and it is but still downsampling to 44.1khz no matter what I do. Isn't it hardcoded into the patch to only do 44.1 at this time?

[pabouk]

I checked all my setting and made sure that it was being sent a 48khz signal and it is but still downsampling to 44.1khz no matter what I do. Isn't it hardcoded into the patch to only do 44.1 at this time?

Is not the resulting wav in fact 48 kHz PCM data with just a 44.1 header? Is not the recored wav time length about 9% bigger and does not the sound have a lower pitch?

[mborus]

Is not the resulting wav in fact 48 kHz PCM data with just a 44.1 header? Is not the recored wav time length about 9% bigger and does not the sound have a lower pitch?

That's what I got when I tried 48 kHz.

I opened the file in Audacity and selected the correct sample rate (48 kHz) and
afterwards it played OK.