Problems transfer to Digibeta

[soundslikejoe]

I mixed a film months ago. The director is having a transfer to Digibeta. The house uses AJA Kona 3 video out from Final Cut to Dbeta. I don't know if this is AES/EBU or Analog transfer.

The house has noticed that the Dbeta "sounds squashed" on playback. It doesn't sound bad monitoring the transfer but it does sound bad on playback from the deck.

Do Digibeta machines have any limiter that I could be slamming?
The file is 20bit, 48khz and not 0fsd; peaks are -2dB for L and -.93 dB for R, RMS max is -10dB.

There were no problems when creating the AC3 file for DVD and playback on other systems has been fine.

Any other speculations?

[UnderTow]

If it is going through analogue and things aren't calibrated, the signal could be clipping at the inputs of the digibeta.

What is being monitored during the transfer? The digibeta or FCP?

It is hard to tell what is going on without knowing how everything is setup and connected together. (Including the monitoring path).

Alistair

[kk@jamsync.com]

Originally Posted by joenovice

I mixed a film months ago. The director is having a transfer to Digibeta. The house uses AJA Kona 3 video out from Final Cut to Dbeta. I don't know if this is AES/EBU or Analog transfer.

The house has noticed that the Dbeta "sounds squashed" on playback. It doesn't sound bad monitoring the transfer but it does sound bad on playback from the deck.

Do Digibeta machines have any limiter that I could be slamming?
The file is 20bit, 48khz and not 0fsd; peaks are -2dB for L and -.93 dB for R, RMS max is -10dB.

There were no problems when creating the AC3 file for DVD and playback on other systems has been fine.

Any other speculations?

Did the post house make sure that the pans were set correctly for output from FCP? The classic problem is that people leave the pans on the output mixer centered which makes the transfer *mono*.

[Frank S.]

Originally Posted by kk@jamsync.com

Did the post house make sure that the pans were set correctly for output from FCP? The classic problem is that people leave the pans on the output mixer centered which makes the transfer *mono*.

Also check to see if the input gains on the DBeta deck are pushed in. When pulled out they become variable rather than just passing the signal through. This has bitten me in the butt a time or two when someone who used a deck prior to me didn't zero it out.

[soundslikejoe]

Thanks for the tips folks.... I haven't heard back yet but am leaning toward your estimations.

The first guess from the post house was bit depth compatibility; my file is 24bit dithered to 20bit. I'm certain that Digibeta supports 20bit audio. I've been told that typical workflow (broadcast services) is 16bit.

Any reasons why people would prefer 16bit laybacks vs. 20bit?

Thanks again....

[Frank S.]

My guess is because it is more widely accepted. However, my system is 24 bit and I lay back to D-Betas all the time with no issues.

[bignate]

The Kona 3 card does have 8 digital I/O across 4 AES pairs. They should also check and make sure that the playback settings/sequence settings are set correctly in FCP for 24-bit audio. Also, the Kona card has a separate control panel software. I don't believe it has any kind of level adjustments, but it wouldn't hurt to double check the output settings.

Failing that, have they tried setting the DBC deck to the SDI inputs and trying the embedded audio? When they say that it monitors okay, are they listening to the output of the DBC or the output of FCP? If they are listening to the output of the deck, are they trying it with the Confidence button turned on or off?

Just throwing out a few options for you.

-Nate

[AndreaUB]

If the transfer is analog I would suggest you to check the VTR input impedance switches on the back panel.
Digibeta has 20 bit resolution; while SP decks had a limiter I can't remember if Digis have one too: we never use it here but I can tell you when someone engaged it by mistake the results where awful...

good luck
Andrea

[Stizz]

Originally Posted by joenovice

Do Digibeta machines have any limiter that I could be slamming?
The file is 20bit, 48khz and not 0fsd; peaks are -2dB for L and -.93 dB for R, RMS max is -10dB.

No limiter on our Sony DVW-500 - but don't these levels seem VERY hot to some of you out there? I generally shoot for peaks around the -10dbfs range, and RMS more like -20db = 0 VU

[Frank S.]

Yeah. That seems pretty hot. I usually don't peak over -8.

[AndreaUB]

Digital betacam is a digital format, it can use the full digital scale to 0 dbfs, the same as a cd.
Some broadcasters infact treat some music format tapes applying traditional music mastering and going digital to digibeta using the full scale up to 0 db with no problems, even with the old sx format.
The same is happening when you transfer a dolby E encoded feed to the vtr. It's a digital encoded feed that uses the whole digital resolution, infact you can see the meters with red light always on.
You should not have a problem using the full scale resolution, usually peaks are between -10 db and -8 db when you print masters for the broadcasters because this is the standard. Ebu in europe sets a standard for digital transmission with test tone at -18 dbfs and peak levels may vary between -10 and -8 depending on the broadcaster you are producing for.
The other thing you could check is the dolby setting. On some digibeta vtrs you have the option to enable or disable it, on some others it is always engaged and you cannot disable it: if you print with dolby and read without you have problems, and viceversa.
The problem cannot be in your input levels, as far as I know.

Andrea

[AndreaUB]

what are you listening to when transfer occurs?
vtr output or the output of the aja?

[soundslikejoe]

Originally Posted by Frank S.

Yeah. That seems pretty hot. I usually don't peak over -8.

The film is an Iraq war documentary complete with machine guns and bombs. I mixed the VO dialog to peak in the -6dB and diagetic to -8dB... I set my target levels based on a comparison to Ken Burns projects (loud VO).

The mix was also done to enhance the impact of the explosions.... the music is kind-of loud too (obviously a result of the composer mixing the film ).

Back to Dbeta...

The director hasn't heard back from the house. These people are friends of the director and our project is backseat to clients. Until I can talk to the person doing the transfer, I won't know the specifics of the routing and monitoring.

As always you're comments are insightfull.... For the curious, I will update with resolved details later.

[BrettMorello]

Originally Posted by Frank S.

My guess is because it is more widely accepted. However, my system is 24 bit and I lay back to D-Betas all the time with no issues.

interesting: Today I tried to dump a session on a betacam, and I hear a strange kind of jitter distortion, untill I realized my session was 24 bit instead of the normal 16 bit usually used at that specifical post facility.

When I created a new session at 16 bit and imported the other session, the distortion during the dump was gone. Any technical explanation how this can be?

[soundslikejoe]

thanks for the post B.... I forgot to add my resolution.

resolution to my problem - I have suspected that the problem with the bounce was most likely related to kk's suggestion; "The person didn't pan my stereo mix hard L|R and was attempting to layback summed mono."

The movie was recently admitted to the Florida International Film Festival. After recieving the offical selection the director made the decision to pay for a "real" transfer. The project was sent in the exact same format (24bit and 20bit mixes) and transfered without any problem (probably b/c this person knows what a pan control actually does).

When I created a new session at 16 bit and imported the other session, the distortion during the dump was gone. Any technical explanation how this can be?

This really depends on several factors; was transfer digital or analog, did you use FCP or Avid, etc.

The first thought that comes to mind is Digibeta doesn't support 24bit audio, 20bit is the highest bitrate.

[TVPostSound]

The person didn't pan my stereo mix hard L|R and was attempting to layback summed mono."

That is a very common issue. Especially with "Pro Tools" files as they are NOT interleaved.
I have caught many Avid editors doing my laybacks with the audio output to "stereo" which is incorrect, it must be set to "direct output" 1 to 1, 2 to 2 etc. Thus eliminating the pan factor.

[Henchman]

Originally Posted by Stizz

No limiter on our Sony DVW-500 - but don't these levels seem VERY hot to some of you out there? I generally shoot for peaks around the -10dbfs range, and RMS more like -20db = 0 VU

I agree it seems hot.
Especially the RMS at -10.
I can see peaks hitting that hot on an occasional impact etc.
But an RMS of -10 seems way to hot.
And having the VO peaking at -6 seems up there IMO.
I think having a VO that loud, would make everything else wanting to eb very loud as well, resulting in a very fatiguing experience for the viewer and the mixer.
Was your room properly calibrated at 85 when you did the mix ?

Even when I mix for film, the RMS stay's around -20, or 0VU, even if it is a War film/documentary.
I've mixed action films, and still haven't had an RMS of -10

[soundslikejoe]

The room is calibrated to 83. I spent several days listening to Ken Burns docs and measuring the output with my meter. I mixed the narrative VO to a similar position and production diagetic slightly under the VO.

So far it sounds balanced most places I've watched it; home theaters, stage where Dbeta layback took place, DAW, and small TV.

[soundslikejoe]

OK... so I'm a duchebag... I opened Wavelab and ran the analysis tool again. Under "Loudness" tab it reports....

RMS Power -
Maximum, -9.97db
Minimum, -144db
Average, -25.93db
At coursor, -144db

That puts me at ease. I wonder why they have max and min for RMS? Isn't RMS an average loudness? (root mean square)

Anyone ever wonder how or why I choose my user name? It should be obvious by now.