View Full Version : Loss in Recorded Audio Level

Dan Barnett
04-26-2003, 07:03 PM
I have been trying to figure this one out.
I am recording from VX-2000 into VT2 through a Makie and straight through also.
Upon doing so my play back levels are lower than what I had started with in recording.
I'm recording using just the mini plugs on the Toaster card. No SX-8 yet.
Consequently any audio imported from another computer (music and what not) is just fine and usually needs to be turned down.

Is there an AGC in the VX-2000 or the VT2?

Anyone else experience this?

04-26-2003, 07:25 PM
What you need to do is play some stuff into the Makie, then let the feed go into the Toaster. Make sure that the Toaster can record that level now.

Hit the Solo buttons on your Mackie, only on the inputs that your Toaster is coming on. Tweak the audio mixer in the Toaster, until solo matches your orignal inputs (without solo).

Everything on the Makie and the Toaster should start at unity.

I hope I explained this right-
Faraz Ahmed

04-28-2003, 10:35 AM

Another thing to check is to make sure you are preserving the balanced or unbalanced chain with your audio. T2's inputs are balanced. Make sure the output from your deck is balanced and when hooking up to your Mackie board, make sure you are going into balanced inputs. Make sure the Main Outs from the Mackie are balanced too.

04-28-2003, 04:10 PM
T2's mini inputs, 3.5mm (as you are using them) are balanced. With BOB you have a selection of balanced or unbalanced inputs/outputs.

04-29-2003, 10:04 AM
Build 3890 records 6dB lower than you think.

Play back a file with peaks at -0dB Full Scale. A file ripped from a comercial pop CD or sine wave built in Soundforge works.

Note where the signal hits on the main meter.

Record a clip, hitting the same place on the meter.

Look at the waveform with Sound Forge or Cooledit.

Not that the peaks are at -6dB.

This doesn't appear to be a 'bug' but a design decision on the part of NewTek.

04-29-2003, 06:20 PM
We've found that it records 6db low also. My boss was shocked (just today actually) when he discovered that not only does it record low but the audio is unbalanced. They couldn't imagine why there was so much audio trouble and why the toaster doesn't allow third party audio cards. Editing would be much happier if newtek at least let us record at the exact level coming in. Then there are also those unused AES/EBU inputs on my SDI card looking rather forlorn. When are they going to be activated?


Joel Slack
Media Department, New Life Christian Fellowship

Dan Barnett
04-30-2003, 01:24 AM
Thank you one and all!
Now that we have solved that particular problem.... hey, what can you say.
Anyway I understand how things are working. Now, I know what direction to take this issue in.
If I come up with a solution I will let you know.

04-30-2003, 07:04 AM
I was watching this topic to see if someone else would note the 6db loss in recorded audio. I've known about this for some time and it seemed that only a few users were willing to confirm it. From past discussion on the issue, I was of the understanding that it was in fact a glitch or it was compensating for a glitch and was not an intended feature.

Whatever the reason, it is a reality that the recorded audio is lower than the audio that runs through the Toaster. This is completely unacceptable and I have been waiting patiently for a new build that returned our ability to maintain unity audio levels throughout the system.

The audio capabilities have been all over the place with the Toaster. In the beginning the actual audio levels were pretty close to unity, but none of the level meters in various applications were of any use. Then there were several attempts to make the level meters more representative of the actual signal, and I'm sure a lot of other audio related issues were addressed. Now it seems the meters are pretty funtional, but the recorded level is incorrect.

I would gladly take the audio functions we had in the first build over what we have now. I can ignore some stupid GUI audio meters that are completely incorrect, but it's much harder to ignore the low audio levels on recorded materials. Unity is just another word for accurate. Right now, things are not in unity and it needs to be fixed.

04-30-2003, 07:38 AM
I have been fighting with the same issue for over a year now. I am running the current build and have never been able to get out what I put into the toaster, as far as audio is concerned. In addition to the appx 6db loss, I am also dealing with an unexplained firewire issue in which the l/r channels are reversed upon capture. I, too, think the audio section of this fine machine needs some serious attention including a return to unity and the ability to do some basic audio processing like other applications.

There...my 2 cents worth...and worth every penny !

Blaine Holm
05-01-2003, 02:04 PM
this does answer some questions we have been having for a while now, and at least I'm not going crazy...
...at least I'm not crazy over this, jury is still out on everything else...:p :

05-01-2003, 09:42 PM
to quote:
T2's mini inputs, 3.5mm (as you are using them) are balanced. With BOB you have a selection of balanced or unbalanced inputs/outputs."

Are you certain T2's mini inputs are balanced ?
Can't see how with a stereo cable terminating in a mini-plug.

05-02-2003, 12:10 AM
Just to clarify:

The T[2]'s mini-plug audio inputs are UNBALANCED.

And the -6dB recording level was an intentional design decision.

05-02-2003, 05:38 AM
No offense Eugene, but I'd like to have that from Newtek themselves. I don't doubt the mini-pin connections are unbalanced despite numerous claims to the contrary, but the 6db loss is another issue.

If somebody intended the T2 to lose 6db of level whenever you record, it can only be due to a bigger problem that would require some awful attempt like that to compensate for it. Whatever the reason, this is wrong. If the T2 is going to route audio through the system in a switcher/mixer/live environment, then it MUST maintain the same level on its own recordings that it routes through the system. No person in their right mind would accept a recording device that shows 0db levels feeding into it, but reproduces a -6db signal upon playback.

I've never heard anything about this being intentional from Newtek. In fact they aknowledged to me that this was an issue that would be corrected many weeks ago.

10-12-2003, 01:45 AM
I experienced the same problem with audio levels through the RCA inputs; confirmed when I captured something from my Digital8 deck and it played back normal. This lead me to a workaround for anyone who has a MiniDV or Digital8 deck/transcoder. I am now running my video source (in this case, a DVD player) into the digital deck where it converts it to a digital signal. When I capture this signal, it plays back perfectly. I hope this might help someone else out there!


10-13-2003, 12:54 PM
Mini jack/plug is by nature an unbalanced connectior....

My hope is that Newtek can hunt down and steal a talented audio programmer from the Sound Forge Team who is increasingly dissatisfied with the Sony Corporate environment!

Tom Wood
10-13-2003, 02:07 PM
Just speculating here, and I'm not an audio engineer by any stretch of the imagination, but based on the research I did when shopping for a 24 bit 96 Khz audio card, I'd say the -6 dB is related to hardware issues inherent in the 16 bit rate during recording. It gives the -hardware- that much extra headroom before clipping, if I understand this all correctly.

Newtek -may- be waiting to replace/upgrade all their cards when 32 bit recording becomes the new thang, at which point there is no more need to upgrade because the sampling is effectively infinite. Pure speculation on my part. I'm doing voiceover for animation, so I'm able to use a second card with 24 bit 96 Khz sampling and sync it up later. If you can figure out a way to use the various sync/lock systems to synchronize audio to a camera, it may be possible, I don't know, I'm only using TED. Going out at 16 bit isn't an issue.

One thing that makes a dramatic difference throughout the recording process is to use a compressor on the microphone input -before- going into the computer. It smooths out all the 'hot' peaks and lets you turn things up a bit. I then use a normalizer filter in Wavelab to boost the entire waveform, going for that big broadcast sound.



EDIT: Also, the trick of using more than one copy of the audio waveform in TED does work to boost the signal, but it's all a guess as to the real level going out. I just won a bid on a Mackie 1202 mixer on eBay (yay!) so I'm going to experiment with going through that before going to the DV deck.

10-13-2003, 09:25 PM
Originally posted by slacker
Then there are also those unused AES/EBU inputs on my SDI card looking rather forlorn. When are they going to be activated?


Good News, I think... the AES is now activated for record with the first VT3 Patch. I have posted questions about it and emailed Aussie but have not heard anything. Best I can tell, when you set the preferences to record AES with SDI input, you will activate SDI input, but only when you are capturing, i.e. - it does not seem to show up as a Audio input but automatically feeds in when capturing SDI.

Lastly, Capture panel seemed to be 1/2 level while Batch Capture seemed appropriate, or vice versa, can't remember... Last lastly, I find no evidence of AES out, but may be my cables and will check it when I get time.

Dick Nelson
10-28-2003, 08:50 PM
Originally posted by bradl

Good News, I think... the AES is now activated for record with the first VT3 Patch.

What "first VT3 Patch?" What else does it do? How can one get it? We still have issues with CG lockups (despite a patch that NewTek sent me awhile back -- I gathered we were a kind of beta test for it) and DDRs skipping arbitrarily from the selected clip to the next one in the queue following an effect. We have also been battling this audio level problem, too.

10-28-2003, 09:44 PM

The full list of features is there.

Dick Nelson
10-29-2003, 07:12 AM
Originally posted by Danner

The full list of features is there.

If this is a reply toi my message, I don't understand it.

10-29-2003, 07:27 AM
He supplied you the link that answers your question.
"What "first VT3 Patch?" What else does it do? How can one get it? "

Click It.
VT3 Update Info.

Max Steel
10-30-2003, 01:35 AM
Actually.... VT[3] certainly does record many dB lower than live audio going thru.

However.. this may be a fix to some of you.
When recording.. make sure to record 'Program Out' instead of 'MAIN IN'.

I have discovered that recording the Program out yeilds virtually identical levels to whats going thru live,



11-03-2003, 12:05 PM
I always record "program Out" and still get audio playback about 6 db low. You would think that NewTek could fix this problem since in some earlier builds of T2 the audio was almost correct db levels. Of course NewTek abandoned T2 and all who paid for it, so wht should they fix it in T3 when they know we will all pay another $500 for T4 soon. No one listened to me when I said NewTek should have fixed all the problems before we paid for T3 now ew are no better off. and have sunk another $500 into the problems. And before anyone can jump on my case again and say T3 is just fine I will admit I am not happy with my T1 at $2500, my T2 at an additional $500, and now my T3 at another $500. That is $3500 for something that still doesn't work. Now go ahead and throw stones at me.

Max Steel
11-03-2003, 05:29 PM
Steve... I think another concern or weakness of T[2] and VT[3] is on the hardware side.

It's hard enough to debug the software without adding in other issues such as the apparent assortment of various harware revisions both in Toaster cards and SX-8's.

To answer your comment tho regarding Audio.
T[2] audio has always recorded perfect levels for me on both Main IN and Program out.... as soon as I installed Vt[3]... the 6dB (or slightly more for me) drop occurs, but ONLY recording Main IN.

I have a revison C SX-8.. and apparently some recall was announced last year, but I never was informed, but I had a new SX-8 anyway.

I'm certain that a number of users who have audio issues would be using original SX-8 hardware that hasn't been updated or whatever, or are using different revision Toaster card hardware.

Now.. I'm not being deliberately rude to betatesters etc.. and have utmost respect for Toaster users.. but how the smeg did such a thing make it thru all the months of intense betatesting??

ie) Did this audio weirdness just creep in in the last builds?

I reiterate.. this drop in recorded volume never occured to me in T[2] with my revision hardware.

So, I actually think this problem is a combination of factors that effect various Toaster end users in different ways.

Steve.. Does your SX-8 have a letter after the serial number?.


11-03-2003, 10:55 PM
I am using the mini plugs on the Toaster board and am getting exactly the same results as every one else. E to E play through while recording is at 0db but playback of the audio recorded through the toaster software is at -6db. Just to add, when I record using the same toaster mini plug inputs using Cool Edit I get a perfect 0db recording and playback. This proves the Toaster card is capable of normal audio recording but the T3 audio software is faulty or incorrectly written. The recordings with Cool Edit play back at about +1db in toasterEdit. I hope Paul does not ignore this post, even if he chooses not to respond.

Paul Lara
11-04-2003, 07:34 AM

Jim Capillo
11-04-2003, 07:49 AM
I know this problem has been bandied about for a while now, but we saw almost the exact same conditions when inputting an analog 0db signal into a Panasonic DV-2000 - almost an *identical* -6db loss.

Coincidence or inherent ? :confused:

Max Steel
11-04-2003, 01:43 PM

To be honest I haven't tried recording without the SX-8.
(not in ages anyway)

With T[2] I used to do 'a lot' of recording using SoundForge.

I have used T[2] with both balanced and unbalanced signals fed from a high end Fairlight audio edit system, so I know exactly what levels are going in and out.

To me.. this is the entire point of having a vitual audio mixer.. ie) So I can calibrate all my i/o.. and with T[2] this worked for me perfectly. VT[3] however is another matter completely.

I agree with you about it certainly being a software problem in respect of VT[3], but I don't have this issue with T[2], so logic would dictate, that the issue can not just be a 'pure' software only problem.
(hence my comments in the previous post above)

I would need to confirm that the same loss occurs when recording Program out on my system, but without SX-8.

If I don't get the recording drop in level..... then perhaps this
would indicate some difference between Toaster hardware revisions, as if we are runnng the same software with a pure Toaster only card solution, and we get different results, then surely this means it's a hardware difference?.

(grin) Talk about yer weird science!.