Removing artifacts produced by software

Imagine people coming here who do not know the rundown of composing the ideal post. You seem to know what you need to know. If it is missing just ask for it.
I think you missed my very simple offering of a prospective. Typically, most the time at least I try to stay in 'what can we (I more specifically) might we do -now- at this point we find ourselves mode, to make 'the project happen'. To me this applies to all things, audio, to .. almost anything. Then of course, sometimes miss the mark', stray' sure no doubt about it.

The idea', stands. You could have gotten much further, much sooner.
No big deal.
 
Chill out man. We are trying to help and you are not accepting the help. This is your problem.

Sorry but you are not helping yourself here.

Best wishes...
I am not accepting the finality of one person's post. That person also seemed to be offended that I did not accept his comment as the end all be all. I am just trying to keep the thread open to help from someone that knows something more.

Then you come in and say I am being a dick.

I have repeated gratitude for help offered however a lot of what has been presented does not apply so I have tried to describe more and provide links to videos. That seems to be mistaken as resisting help when it is just an attempt to steer the ship on course.

Also, why would anyone accept the finality of a response when the people saying it can't be done are the same saying they do not understand what I am describing? How could they be so sure what cannot be done when they have yet to comprehend the problem? Yet somehow to you I am not accepting help?
 
I just posted the video link so you could see the interface and get an idea of what the software does.

Still not sure what you posted. Is it software that you have used that is inadequate, or is it expensive software that works but exceeds your budget? Or maybe its just a generic example? I can't hear anything wrong in the video, even though your OP says that "all these types of software" have the problem.

Anyway, it seems to me that software that corrects an unrealistic sound to a realistic one would be more complex and more expensive, and would be more than capable of doing what the most expensive voice alteration software does. I don't know why such a thing would be a separate plugin, and not simply a dial on the alteration software. Just my opinion, although I know you are looking for expertise.
 
Since I understand the problem, and have been doing this sort of work since before computers, though not as long as Bobbsy, you will take my word for it. A perfectly natural sounding plugin that changes a man's voice to a baby's voice doesn't exist.

As I've explained before, the reason they all start to sound like vocoder is because they essentially are vocorders. What they do is they split they split the audio up into parts (consonant, pitch and formant) and use a pitch shifter to adjust the pitch and formant. If you have any experience with pitch shifters, you will know that even the best ones can only change the pitch so much before starting to sound synthetic. That is what is happening here.

The plugin that you posted also has another control that makes the audio sound more gravelly, which sounds to me like they are chopping the audio to make it sound like that.

You are expecting too much from the plugin. The reason it sounds ok in the video is because you are hearing the audio through the camera mic. I'll bet that if you heard it direct, it would have the problems you have found doing it yourself.
 
I have control over everything. I get the same results whether using studio produced voice tracks or those I created on my own. The artifacts are from the software not the recording process but I understand that I need a pristine source to get the best results. It is obvious to me that the source of this artifact is in the software manipulation stage not the creation of sound file stage.

I think that you should take a bit more care with your descriptions.

I am guessing that the "creation of sound file stage" refers to the input file (before transformation), not the output file (after transformation). And "the software" you say is the source of the artifacts refers just to the sound authoring package.

(I was also not always sure when you were referring to software as artifact removal software, and when you were referring to the software that was creating the artifacts. Sometimes, you just said, "the software".)

Algorithms for transforming sound exist on sound authoring software, on sound card interfaces and on sound card on-board processors (hardware). The authoring software can query a software interface driver, and if it has the capability, ask it to perform elements of the algorithm. Even though the requested function may be defined in some international standard (guaranteeing a well-defined outcome), there are various standards to choose from. So, the processing can depend on which standard the driver conforms with. Similarly, the driver can ask the sound card's on-board processor to do the task. So, the algorithm can end up being shared among different hardware and software components on your system, and be hardware and configuration dependent.

If your processing is well-defined by international standards, then you can be guaranteed system independence for basic transformations (like converting file format), or high fidelity for complex ones. This is because, even if the requirement is subjective or otherwise imprecise, hardware manufacturers and software developers are competing to one standard and subjectively trying to make their product sound good. So, you can take the standardized output file from one system and, although it may sound slightly different on another system, it will be as good as the producer of that system intended.

On the other hand, if you are dealing with transformations that have no standards, one manufacturer/software developer cannot guarantee that what sounds natural on his test system won't sound "mechanical" on another. I do not see any way that you can prevent your transformation software from being completely system dependent. There is no standard for stakeholders to test against to ensure that their product conforms with any other.

Researchers who pioneer new transformation technology will have little choice but to pick one system and make it work well on that system. If you change system, all you can do is tweak everything from interface software to hardware, to try to optimize the technology.

In summary, unless you know for sure how the software works, it may not be so obvious where sound artifacts are coming from.

Perhaps you can get a free trial of that $3000 software and try it out on your computer, and see if it is still as good as advertised.
 
Back
Top