loading large audio files

[tester]

Another file related practical question.

This is for another concept, but just in case I ask. How to deal with loading large files? Like 30-60 minutes (standard 44.1k 16 bit stereo) per one file. No wave display is necessary. Does it reads/loads the whole file, or buffers it partially and uses as a stream? Or - anything important to pay attention to?

[Dell]

Your biggest issue would be load time. Try loading a 1 hour wav file in your DAW for the first time and you will get an idea on what to expect.

[tester]

That why I don't need to have a wave display, just streaming audio with buffer. Try to loat whatever in winamp or whatever media player. It just plays, mixing and DSP processing on the fly, no problem with looping.

I case of seeing the wave display - I guess even a small (let say 5-20 seconds long) self-redrawing (or scrolling) wave window would be nice; such partial loading so to speak.

For DAW-like solution, some peak builder would be required?

[Dell]

Peak loading in a DAW creates its own proprietary document so the audio file can load faster at a later period. However, not having any visuals present will not make the file load very quickly if the size of the audio file is big. If I recall correctly, a 1-hour wav file should be around 600 MB. Load time will not be fast under those conditions.

I've created plug-ins that offered audio files embedded in the design. The plug in takes a long time to load. This is based on the audio file’s size and it is nowhere near 1-hour much less 30 minutes.

You should be using other devices such as VST Reverbs that use impulse files to attain a particular (reverb) sound as a means of reference. Not Winamp. All professional audio devices load time will vary based on the audio file size.

[tester]

@Dell - I understand your point, but please try to understand mine. I'm focused on something a little bit different. It is in synths cattegory, because it makes sounds and processes them in ways typical to synth/modular understanding. Self generated and effects that can be mixed or applied to an audio file. I'm interested to get the full advantage of using SM, but first I must learn and recognize if the SM can do what I will need; concepts are the same in several occasions. That why I ask about partial loading (streaming/buffering possibilities).

[Dell]

Synthmaker can do anything you want from an audio perspective providing you know how to implement it.

However, you cannot overlook the side affects of creating a very complex design. What you are explaining is similar to a reverb that loads impulses. This will take time to load in addition to consume more CPU. These side affects will vary on the strength of your computer. If you add a buffer delay will come into play. The laws of physics still apply even in the digital world where there are no mechanical moving parts.

One of the things many new developers do is create a no compromise plug-in and find that the computer is operating beyond 50% when in use. You must take things one-step at a time. The longer you work with Synthmaker the better you will become. If you try to do everything at once, chances are the results will not be to your liking.

Creating a small Loop in the audio file will be more efficient than trying to load a 1-hour audio file. It may be best to implement your ideas in Synthmaker first, and pose questions if you stumble upon a dilemma. Asking a question from a hypothetical point of view with no reference off hand, will not give you the answers you are seeking.



Cheers!

[tester]

@Dell

I understand your point, but I thinkk you don't understand mine. It's not about loops. It's about processing (non-invasive on-the-fly or offline) and people who use exe files, because they are not DAW people (it does not means that they are not pro audio). It's not hypothetical, because such market is greater than DAW market. And it's about things I'm familiar with in non-SM environments.

You are saying that everything is achievable in SM. I agree, but in exchange I tell you something. I resigned from Synthedit after few days, because although everything is achievable there too - I simply did not how to make it. SM seems to be easier to use, despite the fact that some behaviours semed to be a little better in SE.

Just simple streaming. It's not like loading a file to memory, where you need to reserve that memory for the whole file or files, it's partial loading, working with buffers. I don't know how to make it, but I have a vague knowledge how it works.

I'm a task oriented person, and I generally know how much certain real-world tasks eat of the CPU and RAM. I did split my thinking into steps and this is the step I asked about. My excitement is about SM in general, but my precautions are regarding practical limitations that I can face with.



Do I need to explain that in secrecy I'm doing a warp drive for space travels to exceed the speed of light?

p.s.: my intention is not to offend you nor insist; it's just this topic is much bigger than my english.

[Dell]

Hi.

The problem is not your English. You tend to leave out vital parts when asking a question.

For example;

“I woke up this morning and, brushed my teeth.”

“After I was admitted in the Hospital I…”

This is what your questions come across as.

Everyone on this forum has some type of technical experience before investing in Synthmaker. I’ve been designing & repairing audio hardware over 20 years. When you around technical people details (in regards to what are you trying to do) are needed to give you the best possible answer.

I never had a discussion with Synthedit developers so I have no idea how their thought patterns are in terms of helping one another. However, on this forum if you are hesitant to say what you are trying to achieve, and you do not offer an OSM much less a visual image chances are you will not get an answer.

If you prefer not, to offer an OSM that is understandable. However you need to offer more details on what you are trying to achieve so those who are trying to help you can get an idea on what you are aiming for.

If you do not know how to build a particular component, it is better of saying so than leaving an impression on others that you are already familiar on how to design the component.

Cheers!

[Tzarls]

To the original question: SM loads the whole file at once. It can´t perform streaming from (or to) disk. The time it takes to load a file depends on the length of such file and my guess is that the type of data also matters, since SM has to convert the data to 32bit FP and the conversion has to take some time and CPU.

[tester]

@Dell

Okay, I see what you mean now. We just talking from different perspectives, thus we talked about different things. I will have OSM to show when I buid it. Now I'm focused on gui and oscillator things. I'm just releasing other further questions (questions for tomorrow) from my head in order to have more inner space to work (and because I'm aware that others probably have encountered similar issues and maybe solved them).

Tzarls is right, and asked the question using more proper terms and concepts, so let just focus on that.

[tester]

I'm curious if the projects are (indirect but) inter-changable between flowstone and synthmaker.

http://www.dsprobotics.com/support/view ... f=55&t=366

I see there is a primitive module audiostreamin.