Recording low level faint sounds

[JessicaBella0187]

7 Yếu Tố Xếp Hạng SEO 2018 – Bạn Đã Biết Chưa

Hầu hết qua mỗi năm Google lại đưa ra những yếu tố để xếp hạng site mới. Và trong năm 2018 này những yếu tố xếp hạng nào sẽ lên ngôi. Hãy cùng Dao tao SEO - Khoa hoc SEO dang cap the gioi tai VN tìm hiểu 7 yếu tố xếp hạng SEO 2018 dưới đây nhé.

1.Nội dung chất lượng vẫn là tiêu chí quan trọng

Tôi dám cá rằng chẳng một ai làm SEO lại dám vỗ ngực rằng: “mình không cần có nội dung chất lượng mà vẫn có thể xếp hạng cao bền vững”. Đến nay chất lượng nội dung vẫn là tiêu chí quan trọng đối với xếp hạng của website cũng như trải nghiệm người dùng. Khi nội dung được tối ưu hóa với từ khóa chuẩn thì sức mạnh của SEO mang lại là rất lớn. Trong khi đó nếu xây dựng nội dung mà không có giá trị chắc chắn sẽ gây ra những bất lợi đối với xếp hạng cũng như khi đứng trước các thuật toán của công cụ tìm kiếm.

Xuất bản nội dung chất lượng

Xuất bản nội dung có giá trị và chất lượng sẽ làm gia tăng tỷ lệ truy cập trang, gia tăng thời gian ở lại trang và giảm tỷ lệ thoát trang. Điều này có tầm quan trọng để Google xếp hạng, đánh giá site.

2.Yếu tố bảo mật của trang web

Yếu tố bảo mật của trang web cũng đóng vai trò quan trọng trong tiêu chí xếp hạng của Google. Để gia tăng tính bảo mật của trang web và giúp trang web có cơ hội được xếp hạng tốt thì bạn hãy chuyển trang web sang HTTPS nhé. Tuy nhiên bạn hãy đảm bảo việc chuyển trang web sang HTTPS phải được thực hiện bởi những người có chuyên môn và kinh nghiệm để không gây ra những thiệt hại gì cho SEO cũng như các yếu tố trên website nhé.

3.Hãy tăng cải thiện trải nghiệm người dùng

Trải nghiệm người dùng là yếu tố vô cùng quan trọng chứng minh chất lượng, nội dung của trang web là tốt đối với công cụ tìm kiếm. Còn nếu như trải nghiệm của người dùng không tốt điều đó cho thấy bố cục trang web, thiết kế, giao diện, nội dung cùng nhiều yếu tố khác không đảm bảo, không đạt tiêu chuẩn.

Theo đó Chuyên gia đào tạo SEO Online Marketing Eric Doan khuyên bạn rằng yếu tố quan trọng nhằm tăng cải thiện trải nghiệm người dùng chính là cấu trúc trang web. Hãy đảm bảo rằng trang web dễ dàng sử dụng. Mọi trang đều có thể điều hướng được mà không bị lỗi gì. Cấu trúc trang web rõ ràng, dễ nhìn nhưng cũng đảm bảo bắt mắt và thu hút để chinh phục người dùng ngay từ cái nhìn đầu tiên.

4.Website đảm bảo thân thiện với thiết bị di động

Với xu hướng người dùng ngày càng sử dụng những chiếc điện thoại thông minh nhỏ xinh để truy cập vào website, tìm kiếm thông tin trên mạng thì công cụ tìm kiếm đã đưa ra tiêu chí xếp hạng cho SEO website là yếu tố phải đảm bảo thân thiện với thiết bị di động. Do đó bạn hãy đảm bảo rằng website của mình được thiết kế thân thiện với mobile nhằm giúp người dùng có trải nghiệm tốt nhất ở bất cứ nơi đâu, bất cứ khi nào. Đồng thời gia tăng thêm nhiều lượt truy cập mới cho website.

Website đảm bảo thân thiện với thiết bị di động

Tuy nhiên khi thiết kế website thân thiện với cả thiết bị di động và desktop thì phải đảm bảo nội dung hoạt động đồng nhất trên cả những thiết bị này nhé.

5.Thực hiện tối ưu hóa trên trang

Tối ưu hóa trên trang hay còn được gọi phổ biến là tối ưu hóa SEO onpage. Việc tối ưu hóa SEO onpage đã có từ rất lâu và đến nay vẫn đóng vai trò cực kỳ quan trọng đối với việc xếp hạng trên công cụ tìm kiếm.

Thực hiện tối ưu SEO onpage có rất nhiều yếu tố mà chúng ta phải thực hiện, có thể kể đến như:

Tôi ưu thẻ mô tả, tiêu đề, đường dẫn URL: đảm bảo viết ngắn gọn, súc tích và có chứa từ khóa chính cần SEO.

Tối ưu hóa hình ảnh bằng cách điều chỉnh dung lượng, kích thước, đặt tên ảnh, thuộc tính ALT.

Sử dụng hệ thống thẻ heading trong bài, phân bố từ khóa chính phụ hợp lý và dàn đều trong bài.

Xây dựng liên kết nội bộ hợp lý và khoa học.

6.Tốc độ tải trang phải được tối ưu

Tốc độ tải trang là 1 tiêu chí đánh giá và xếp hạng SEO website trong năm 2018 này. Tốc độ tải trang có tác động đến trải nghiệm người dùng cũng như lập chỉ mục của công cụ tìm kiếm. Vì thế hãy đảm bảo tốc độ tải trang phải được tối ưu. Theo đó bạn hãy kiểm tra các yếu tố gây ảnh hưởng đến tốc độ tải trang để tối ưu tốt nhất. Với thiết bị di động tốc độ tải trang nên dưới 1 giây, còn desktop tốc độ tải trang nên dưới 3 giây.

7.Yếu tố backlink

Đào tạo SEO danh tiếng thế giới Eric Doan muốn chia sẻ với các bạn yếu tố cuối cùng trong 7 tiêu chí xếp hạng SEO website của Google năm nay chính là backlink. Những backlink chất lượng, được xây dựng tự nhiên sẽ mang đến những giá trị tích cực cho SEO và hỗ trợ cải thiện xếp hạng cho trang. Đồng thời hãy nói không với việc spam link và xây dựng link kém chất lượng nhé.

Hi vọng những chia sẻ trên đây sẽ giúp các bạn tối ưu tốt các tiêu chí xếp hạng của Google nhằm đưa site lên top thành công, bền vững và hiệu quả nhất.

[pandy]

Probably you will be asked to provide more data - what kind of microphone, what kind of audio interface, how microphone i mounted, placed etc. etc. One of simplest tricks to improve SNR by approx 3dB is using mono microphone connected to stereo inputs (i.e. to Left and Right input - later by adding two channels in digital domain SNR can be better by approx 3dB)

[JVRaines]

Use a microphone with low enough frequency response to pick up the sounds. Put it as close as possible to the source. Get a better preamplifier with lower noise floor and use balanced mic cable.

[Cornucopia]

...One of simplest tricks to improve SNR by approx 3dB is using mono microphone connected to stereo inputs (i.e. to Left and Right input - later by adding two channels in digital domain SNR can be better by approx 3dB)

Don't give her false hope. You and I both know that that only works when the noise is decorrellated from the signal and thus is random compared between channels (constructive interference adds signal fully 2x but noise has both constuctive & destructive). This occurs with analog storage/transmission, but digital usually has interchannel noise that is correllated amongst itself, so cannot take advantage of that trick unless it is of very high audiophile A->D type that includes separate gaussian dither generators (and their clocks) for each channel.

Otherwise, I am in agreement w jvraines' assessment, though the term "low frq. response" isn't accurate for what you're trying to say. Broad, wide, sensitive maybe, etc. Low connotes no high (freq) end.

Scott

[pandy]

...One of simplest tricks to improve SNR by approx 3dB is using mono microphone connected to stereo inputs (i.e. to Left and Right input - later by adding two channels in digital domain SNR can be better by approx 3dB)

Don't give her false hope. You and I both know that that only works when the noise is decorrellated from the signal and thus is random compared between channels (constructive interference adds signal fully 2x but noise has both constuctive & destructive). This occurs with analog storage/transmission, but digital usually has interchannel noise that is correllated amongst itself, so cannot take advantage of that trick unless it is of very high audiophile A->D type that includes separate gaussian dither generators (and their clocks) for each channel.

Otherwise, I am in agreement w jvraines' assessment, though the term "low frq. response" isn't accurate for what you're trying to say. Broad, wide, sensitive maybe, etc. Low connotes no high (freq) end.

Scott

Nah - every electron counts, amplifier etc - many noise sources - averaging works fine and by principle this very efficient universal trick (uncorrelated reduced by 3dB, correlated increased by 3dB but usually easy to be removed trough statistical processing in frequency domain) - i didn't wanted to provide feedback as such (low noise amp, good mic) as there was no information about current setup (perhaps it is already best equipment yet no results).
Would only add that proper placement may help (like improving directivity by proper place and orientation, using directive mic etc) - once again we don't know.

btw gaussian dither may be and usually it is suboptimal in audio also decent ADC can (should) exploit by principle idea of subtractive dither but usually 99.99% of audio ADC is delta-sigma thus dither is noise shaped and pushed way above useful range.

[netmask56]

Get the microphone as close to the source sound you want and don't wind up the recording level much above the normal setting you would use. Another view is to lift the recording level up but not so much that you get overload, then in post you can lower the level to required aural effect. If you are outdoors and wind is a problem then a wind gag on the microphone would help.

[johnmeyer]

As everyone else says: get the mic close. Also, the quality of mic makes a huge difference. Unfortunately, there are no good cheap mics. Expect to pay several hundred dollars for a mic that will have not only the sensitivity you need, but also all the other characteristics that result in good signal.

Finally, you might be able to recover a little signal from the noise using noise reduction software, after you've recorded. If you go that route, and if the thing you are trying to record is mostly low frequency or mostly high frequency, you could use an equalizer during recording to boost the volume of the frequencies not contained in the thing you're trying to record and then applying the opposite response in post production. Mr. Dolby made a fortune selling systems that did this for tape and, of course, 78, 45, and 33 1/3 records were all cut with some sort of pre-emphasis followed by equalization, although it was done not only for noise, but also to keep the needle from wandering into the adjacent groove.

[amaipaipai]

1. Get Audacity
2. You need to generate a noise profile, this will remove any noise you don't want from what you want, in Audacity record a few seconds of "nothing". Make sure the mic is away from the the source you want to record.
3. Do CTRL+A to select everything, go to Effect > Noise Reduction > Get noise profile
4. Now record what you want using the same recording levels.
5. At the end, select all recording, go to Effect > Noise Reduction > Ok

This should extract the noise and leave just what you want, you can see more of that with YouTube videos:
https://www.youtube.com/watch?v=10FFKl_0GSA

After removing the noise you can go to Effect > Amplify to make what you want louder without clipping, if some of the noise get amplified you need to do another noise profile and re-apply it, or YouTube to the rescue:
https://www.youtube.com/watch?v=O5H7xRzjVkw

[Cornucopia]

Oh, so just as was attributed to Michelango: "you just chip away everything that doesn't look like David"?

In a perfect world, subtractive boolean mixing would give you a perfect copy of what you want. But this isn't a perfect world, noise profiles aren't even close to what you would ideally want or need (which you would know if you had ever worked frequently with them in detail), and so you won't get close to what you want that way.

In fact, using subtractive methods are inherently limited by the self-noise floor and by increasing entropy in processing.

The only way the OP can get true good-quality signal is by recording it optimally.
That means:
Optimal placement
Opt. Pickup pattern
Opt. Freq. Response (to the subject)
Opt. Sensitivity overall (w/ minimal self-noise)
Optimal gain structure
Optimal, lowest noise & distortion mic pre
Optimal, lowest noise, FIR phase-linear, oversampling A->D

It's no coincidence that a matched pair of Schoeps reference/testing/scientific mics can cost from $3500 to over $5k. You don't get that clear, sensitive sound capability without putting in the extra workmanship, attention to detail, and super high quality parts to get to that level.

Of course, we haven't even heard from the OP since the beginning, so for all we know this could be a pseudo-ghost chaser effort.

Scott

[amaipaipai]

We all know that if the OP had the knowledge I'm sure he or she would know what to do, what mic to get, etc.

We are not talking about a high production recording for a studio, but some one that need help to get through a task that is trivial to many of us.

I believe the given solution is good enough to get started.

[pandy]

We are not talking about a high production recording for a studio, but some one that need help to get through a task that is trivial to many of us.
I believe the given solution is good enough to get started.

All you wrote ignoring some scientific aspects - processing in frequency domain you must deal with frequency vs time resolution, spectral leak, periodic vs other signals - FFT is not magical answer for problems. Also you can't ignore limited ADC accuracy and resolution - many factors involved and this is not obvious knowledge - it is not intuitive, non trivial - situation is way easier if your signal is periodical then some other tricks can be used but in case speech they are not usable - speech or other signals common to our world are rarely periodical...
Best way for good signal is good signal acquisition - the worse signal acquired the more difficult to do and less spectacular effects of signal processing.

And no, i'm not exaggerating anything - one of choices you facing when FFT processing is involved is selection of optimal for your signal https://en.wikipedia.org/wiki/Window_function and this is one of many choices you must do before starting signal processing. Which window is OK for your signal?

[amaipaipai]

We are not talking about a high production recording for a studio, but some one that need help to get through a task that is trivial to many of us.
I believe the given solution is good enough to get started.

All you wrote ignoring some scientific aspects - processing in frequency domain you must deal with frequency vs time resolution, spectral leak, periodic vs other signals - FFT is not magical answer for problems. Also you can't ignore limited ADC accuracy and resolution - many factors involved and this is not obvious knowledge - it is not intuitive, non trivial - situation is way easier if your signal is periodical then some other tricks can be used but in case speech they are not usable - speech or other signals common to our world are rarely periodical...
Best way for good signal is good signal acquisition - the worse signal acquired the more difficult to do and less spectacular effects of signal processing.

And no, i'm not exaggerating anything - one of choices you facing when FFT processing is involved is selection of optimal for your signal https://en.wikipedia.org/wiki/Window_function and this is one of many choices you must do before starting signal processing. Which window is OK for your signal?

You don't buy a first class plane ticket to eat peanuts, there is more efficient ways to do it you know.
The OP already told us that he/she has "0 knowledge" about it, let he/she handle it the way he/she see fit one step at a time.

You sound like Jeff Foxworthy trying to explain the medicine side effects for a person that is just asking for a aspirin.

[pandy]

You don't buy a first class plane ticket to eat peanuts, there is more efficient ways to do it you know.
The OP already told us that he/she has "0 knowledge" about it, let he/she handle it the way he/she see fit one step at a time.

You sound like Jeff Foxworthy trying to explain the medicine side effects for a person that is just asking for a aspirin.

Ok, then i will put this simple - i would never recommend FFT denoising as alternative to good signal acquisition.

[amaipaipai]

Ok, then i will put this simple - i would never recommend FFT denoising as alternative to good signal acquisition.

Everybody knows that, but this is the best he/she will get... for free and for now. If he/she need something better, I'm sure he/she will ask.

[pandy]

Ok, then i will put this simple - i would never recommend FFT denoising as alternative to good signal acquisition.

Everybody knows that, but this is the best he/she will get... for free and for now. If he/she need something better, I'm sure he/she will ask.

What i can say - as electronics engineer i was educated to use ADC in particular way - first and most important is proper signal level that allow to fully utilize ADC dynamic range - this means that signal peak should be around -3 ... -6 dBFS - this is only way to not loose anything... so my point was - FFT denoising is unable to replace proper signal level.

[amaipaipai]

xxxx

[pandy]

Are you really gonna give me the attitude, boy?

lol

Have a nice day.

[redwudz]

To all posters in this thread. Keep the discussion civil and without name calling or the thread will be closed.

Moderator redwudz

[netmask56]

Iam trying to record some sound disturbance using condenser mic and audio interface. Having absolutely 0 knowledge on this subject i quickly found that cranking mic gain to the max, in hope to catch everything, will make it even worse due to all the introduced noise. Sometimes i get better results with lower mic gain and some post processing after.

Could someone steer me in the right direction, maybe advise a correct workflow for recording low freq. sounds, or some plugins that i need to investigate?

There have been all manner of suggestions so far some more complicated than probably is necessary. Often the simple approach works out the best for ones needs. In your post you describe the sound you are trying to record as "some sound disturbance". Could you elaborate? Low frequency sounds - what is the source if you know? Give more details of the nature of the disturbance? Might help in tracking down the best approach. However the fundamental thing is get the microphone as close to the source that is practical, use a microphone with good low frequency response , but then that introduces a set of problems like handling noise, wind and the like.

[jagabo]

Also, decouple the mic from sources of low frequency noise. For example, don't put the mic on a wooden table on a wooden floor. Low frequency noise like footsteps will be transmitted through the floor and table to the mic.