Getting Hit by VB QC - Suggestions?
I've been with VB for several months, completed over 100 projects with only getting hit once from VB QC for one word that wasn't spoken clearly.
This clip got rejected because "the overall volume of your audio is too loud and some parts are distorting......The issue is clearly audible throughout your recording, such as at 00:04."
My recording process hasn't changed since the beginning. I record through a Rode NT1 to Scarlet Solo interface to mac mini. My DAW is Twistedwave, Gain set to middle or just above, record 6-7" away from the mic, hitting peaks between -12 - -6. I normalize to -3, run a effects stack I had made professionally from Edge Studio, which runs a High pass, a mild dynamics processor, a little graphic eq, then a peak limiter. I normalize back to -3 and send it in. Everyone has been super happy with my recordings including VB QC until recently.
Any opinions, suggestions, criticisms?
Here's the audition....the client asked for a upbeat, fun, and energetic tone.
https://soundcloud.com/user-612592032-893732381/vb-mhealthy-incentfit-v2
Please sign in to leave a comment.
Hi Brian
I was checking the file and it does sound over compressed. I don't pickup the distortion, but it does sound with very little headroom to me and a bit over processed. Maybe the double normalizing and the limiter are not working as expected. Can you please share a file without the limieter and normalizing just once to see if it does the trick?
Let me know!
Hi Angela, Thank you so much for taking the time to listen to my clip. At first I was very frustrated that I was having some issues with VB QC, especially after having no issues with sound quality all my history working with you, but I had to step back and look at it from another angle. You guys are pushing me to be better and improve. I really appreciate that. Thanks for taking the time to give me some direction.
Here's another audio file. The same time I had that audition get rejected, I had a big job that I did that got caught by VB QC for the same issue, distortion, over processed/compressed. In the clip, the first section 1-17:00 is the first submission that got rejected with the same processing I listed earlier, 17:00-34:00 is my second submission that got accepted. I rerecorded the whole project, just cleaned up the mouth noises/breaths manually, ran a high pass filter at 0db, 93hz, a little graphic eq, then normalize to -3db. Finally, if you have stuck around this long, :-) at 34:00-50:00 is the RAW recording, breaths, mouth noises and all with no adjustments.
https://soundcloud.com/user-612592032-893732381/cato-job-split-test
What are your thoughts? Thanks again for your time.
One last thing, if your interested in seeing it, here's the effects stack I had made that I was originally using.
Thanks Again!
Hi Brian!
Thanks for sharing this in such a detailed manner!
From the 3 recordings you shared, the second section (00:17) sounds the best EQ wise. I would add a de-esser though as it sounds a bit sibilant still. The levels are still a bit high but, level wise, the last one was the best.
I believe that if you lower the gain output of the limiter and bypass the dynamics processor, you'll get the best levels! Also, I would say you are cutting on the high frequencies but I don't think you're targetting the sibilance correctly. What I do to find the problem frequency is that I increase the gain a lot on my EQ and start a sweep from 4K onwards. This will help you feel exactly where the sibilance is located and then you can lower the gain of that frequency range little by little until you feel the sound is more balanced.
Hope this helped!
Hi Brian!
Thanks for your audio and the screenshot. In general terms I think the quality is good. I do think that you may be pushing the dynamics processors a bit hard, how much gain reduction are you getting from the compressor and then from the limiter? It'd be interesting to hear a raw take to listen to your setup and acoustics.
There are also some mouth clicks here and there that distract from the performance. Try to reduce them a bit by recording a bit farther away from the mic, maybe you can also move the pop filter a bit farther to help you control that.
Also, I think you don't need that boost in the highs, as they are creating the slight sibilance which is hard on the ear.
Let me know what you think!
Best,
Juan.
Thanks Angela for your great suggestions. I will try to put those into practice. I really appreciate all your help.
Thanks Juan for your thoughts and help too. In this clip https://soundcloud.com/user-612592032-893732381/cato-job-split-test from 34:00-50:00, the recording is completely RAW. That should give you an idea of my setup and acoustics. As you can tell, I struggle with a clicky mouth and have to edit quite a bit out when I record. I have yet to try the green apple/juice method yet.
I've been relying on this effects stack that I had made from Edge Studio, but yeah, I need to train my ear better and learn to tweak things more myself. I had the stack made and I run it all at once, I've never actually run it one by one or tweaked any part of it to see what it does. I figured it was made by a professional so they know better than I do. But here's a break down if this is what you're looking for.....
Raw the levels are Avg. RMS -24.04, peak -5.90
I was told to normalize it to -3 before applying the stack. which brings it to Avg. RMS -21.14 and peak -3
After the High pass, and dynamics processor Avg. RMS -22.58, peak -8.87
After the EQ Avg. RMS -22.62., peak -6.91
After the limiter Avg. RMS -14.63, peak -0.01
And then I was told to normalize to -3 after everything, it finishes out at Avg. RMS -17.62 and peak -3
I can definitely improve on my technique as well. I will try backing off the mic some more for the mouth clicks. As you can see in my pic, my mic has a built in pop filter, but I'm getting a new gooseneck one to help better with plosives and like you said, keep the mic away from my clicky mouth. So with backing off some, I'll have to boost the input gain to compensate, what avg recording levels do you aim for in the raw recording?
Here's a picture of my recording space. A 3x5 room I built in the corner of a room. Lined with carpet on every wall, floor and ceiling, acoustic tiles, homemade bass traps floor to ceiling in two corners, filled with rock wool and wrapped with burlap, and a suspended cloud ceiling same as the bass traps.
Hi Brian!
I think the raw file is perfect. Just by looking at the wave you can clearly see
a. The overall loudness achieved with the processors in the first take
b. the edits in the first region compared to the second.
Now, we do have to find a way to reduce the mouth clicks without having to edit them out. For this, I suggest you re position your mic. Try placing the mic higher and lowering the angle so that it faces your mouth, but you you don't have to speak directly at the capsule. This may help us minimize the undesired noises.
Adding a second or even a thicker pop filter may help you too. You can hear a plosive at 0:05.
Regarding the processing, I wouldn't approach it that way. In your response you didn't mention anything about the gain reduction levels of the compressor and limiter. But it's interesting that the side chain is doing more than necessary.
First, for recording levels I think a peak -6 dbs is good enough. That should keep the noise floor low, and give you enough headroom for you to play with the dynamics of your performance without having distortion. That's pretty much how you have it right now, and if you listen to the second take you can hear that the volume is consistent throughout the read. So there's not really the need for a compressor.
Then after you normalize, the signal coming into the compressor is really hot. The compressor can be be compressing too much, and then it lowers the output, so that's why you have to apply a second normalization and finally a limiter that's peak at -0.01, which means it's as loud as it can be without distorting the master.
When it comes to processing it's very important to keep an eye on the inputs and outputs so you maintain a healthy balance in order to do as much as you can without degrading the quality of the recording.
We ask for well-edited, unprocessed, professional ready-to-use files, so if you manage to reduce the mouth noises from the recording stage, you would only have to normalize to -3 db peak, and voila! :)
I hope this helps.
Best,
Juan.