How we make video content in Puzzle English: from the chromakey on the wall to a full-fledged video recording studio
We in Puzzle English always focus on the usefulness and quality of our content. Technology is only a tool that allows you to put all our ideas into practice. Today we want to share a case study, how we organize streaming creation of video and audio content on our website: every day the site is updated with fresh material, and not a hundred people do it all. And we will do a little “excursion into history” and show how we grew and improved the quality of the produced content.
The main goal from the very beginning was to create a large body of information about the English language. The first few dozen lessons were removed by the company’s founder, Alexander Antonov. The process was organized very simply: the shooting was carried out at home, on the wall - hromakey, a reflex camera on a tripod and the simplest wire buttonhole that was inserted directly into the camera. A freelancer was hired for the installation (which has now grown to the position of project manager), the work was carried out in Adobe Premiere Pro . The replacement of the background was made through the Ultra Key and Color Key filters, that was enough.
A little later, Alexander Komarov joined Alexander, who recorded shorter videos like piping for exercises. The shooting is the same: a chroma key on the wall, a buttonhole and one light source. The quality was, frankly, not the best, but at that time the project was still more amateur than commercial, and it was not clear how much we can invest in the upgrade.
When the first dozens of paid users appeared, and the videos began to be released every week, it was time to think about upgrading the equipment. They bought a newer SLR (then it was a Canon 550D ), added another source of light, began to better illuminate the chroma key, took a buttonhole of better quality.
')
So we shot about 300 clips. By that time, we had an office (before that we did everything remotely: we sent the source code via ftp, kept reports in Excel and communicated via Skype), and it was possible to transfer the recording there.
Then we opened the “Tichera Method” section, for which it was necessary to capture a lot of good quality material. The first attempts to shoot at the chromakey failed: the quality was not high enough, so we decided to buy a white background and how to light it. On the montage, the brightness of the image is twisted so that the standard white background (#ffffff) merges with the video image. Videos for Youtube were also shot on a white background:
Outdated DSLRs didn’t fit, and “Ticher” recorded for an hour, sometimes longer, so we bought a Canon HF-G25 Legria camcorder , put it on a tripod, took three sources of light for a high-quality picture and added a ZOOM H1 recorder to the buttonholes. so the sound is cleaner. By the way, we have a lot of buttonholes, and all from different manufacturers. Who works with sound, he knows how often they break at the most inopportune moment: the matter is in the fragility of the structure. How we process a sound, we will tell a bit later.
Since there is a lot of text in “Tichera”, which needs to be recited continuously for about an hour, they decided that it is impossible to do without a teleprompter . Ordered from the USA, put an ordinary iPad in it, installed an application that generates mirror text. The scroll speed is controlled by the presenter through your phone.
At first, an echo was heard, and for a long time we could not do anything with it. They put more furniture, hung clothes next to them - that was a little, but it helped. Later, when we moved to the current office (we are currently renting several offices, and one of them is completely set aside for recording video and podcasts), sound-absorbing panels were hung on the walls:
Now about the installation. The video in the “Tichera Method” is cut into “micro-knowledge” - small fragments of 20-30 seconds from one-hour, sometimes two-hour video. All they need to batch process, and then also in a batch render and fill the party. Splitters did not work - they always cut several frames more or less, none of the tested ones worked clearly on the specified frames.
I had to do it manually: for the admin panel they wrote a script that generates the names of the fragments in the form of “block, block_1, block_2”. Installation was carried out in Premiere Pro: color correction, adding a white background, equalizing sound and removing unnecessary pieces; “Sausage” was cut into necessary fragments. Then selection via hotkey X (so that no extra frame is captured, what happens if you just put io tags) and send to the queue in the Media Encoder . The latter generates the names exactly in the form we wrote above (for this, the script was needed), it was necessary only to name the sequence as the necessary word (for example, block_96, then the files were called block_96, block_96_1, and so on). Sending to the queue, launch, and after 10-20 minutes the party is ready. Filled in the admin panel, fragments rose exactly on those places in the script, the names of which corresponded to the file name.
Before editing, the sound was processed: it is recorded separately, so you can immediately throw it into Adobe Audition . We clean the background noise with Noise Reduction : do Noise Print with silence and apply to the entire file. Then we perform normalization through Normalize and save the file separately (all sources are still stored unchanged). After that, Premiere very quickly synchronizes the audio track using Merge Clips , automatically removes the track of the video file and inserts the edited recording from the voice recorder. The finished file is mounted, and then you already know.
For a long time, all these lessons were mounted by one person, but now, when we have 2 channels on Youtube and at least 3 videos per week, as well as regular filming and recording of podcasts in the studio, the lineup has changed. We hired Maxim, who is now responsible for on-site shooting and editing, and Anton Komarov is responsible for on-site shooting and editing. For the latter, there are Canon 7D Mark II and Sony rx100m3 - they are convenient to carry around , and the picture quality suits us. Plus, we have Oleg Vegan, who makes video footage of popular songs, mounts them independently. Anton works in Final Cut X and Motion (for screensavers), Oleg and Maxim work in After Effects .
A couple of words about podcasts . For them, we have two Octava MK-105 and AT2020 condenser microphones , one ISK BM-800 studio and Behringer UMC202HD external USB audio interface . Behringer replaces a professional sound card, we connect microphones to it.
Sound processing, again, is done in Adobe Audition: equalization, compression, limiter, editing and adding music perebivok there. With the help of the Wordpress plugin we upload to the site, via FeedBurner we set up RSS.
Now the process is debugged: the presenters write the scripts themselves, send them to the editors for review, then remove and edit the schedule in Google Docs. Everyone is responsible for his section, independently uploads (on Youtube or directly to the admin area), then the editor fills in the description, and the SMM manager plans to publish to groups. Due to this, there is time for experiments and non-standard moves, and, of course, now we are trying to make not only useful, but also visually high-quality content.
It may seem that everything is simple, however, we went to this for a long time, and to fill the site with this amount of material (almost 700 grammar lessons, and the video in the “Tichera Method” for 650 hours) we needed more than one and not even two years.
We are not going to stop, new projects for the Puzzle English Live channel and a lot of interesting podcasts are next in line.
Readers of the blog give a coupon for 500 rubles for the purchase of a subscription, which includes 8 types of training and weekly mailings about English grammar and vocabulary - "Vitaminki" and "Buns".
And for unlimited and perpetual access to all features of the site there is an “All Inclusive” tariff (the discount does not apply).
The main goal from the very beginning was to create a large body of information about the English language. The first few dozen lessons were removed by the company’s founder, Alexander Antonov. The process was organized very simply: the shooting was carried out at home, on the wall - hromakey, a reflex camera on a tripod and the simplest wire buttonhole that was inserted directly into the camera. A freelancer was hired for the installation (which has now grown to the position of project manager), the work was carried out in Adobe Premiere Pro . The replacement of the background was made through the Ultra Key and Color Key filters, that was enough.
A little later, Alexander Komarov joined Alexander, who recorded shorter videos like piping for exercises. The shooting is the same: a chroma key on the wall, a buttonhole and one light source. The quality was, frankly, not the best, but at that time the project was still more amateur than commercial, and it was not clear how much we can invest in the upgrade.
When the first dozens of paid users appeared, and the videos began to be released every week, it was time to think about upgrading the equipment. They bought a newer SLR (then it was a Canon 550D ), added another source of light, began to better illuminate the chroma key, took a buttonhole of better quality.
')
So we shot about 300 clips. By that time, we had an office (before that we did everything remotely: we sent the source code via ftp, kept reports in Excel and communicated via Skype), and it was possible to transfer the recording there.
Then we opened the “Tichera Method” section, for which it was necessary to capture a lot of good quality material. The first attempts to shoot at the chromakey failed: the quality was not high enough, so we decided to buy a white background and how to light it. On the montage, the brightness of the image is twisted so that the standard white background (#ffffff) merges with the video image. Videos for Youtube were also shot on a white background:
Outdated DSLRs didn’t fit, and “Ticher” recorded for an hour, sometimes longer, so we bought a Canon HF-G25 Legria camcorder , put it on a tripod, took three sources of light for a high-quality picture and added a ZOOM H1 recorder to the buttonholes. so the sound is cleaner. By the way, we have a lot of buttonholes, and all from different manufacturers. Who works with sound, he knows how often they break at the most inopportune moment: the matter is in the fragility of the structure. How we process a sound, we will tell a bit later.
Since there is a lot of text in “Tichera”, which needs to be recited continuously for about an hour, they decided that it is impossible to do without a teleprompter . Ordered from the USA, put an ordinary iPad in it, installed an application that generates mirror text. The scroll speed is controlled by the presenter through your phone.
At first, an echo was heard, and for a long time we could not do anything with it. They put more furniture, hung clothes next to them - that was a little, but it helped. Later, when we moved to the current office (we are currently renting several offices, and one of them is completely set aside for recording video and podcasts), sound-absorbing panels were hung on the walls:
Now about the installation. The video in the “Tichera Method” is cut into “micro-knowledge” - small fragments of 20-30 seconds from one-hour, sometimes two-hour video. All they need to batch process, and then also in a batch render and fill the party. Splitters did not work - they always cut several frames more or less, none of the tested ones worked clearly on the specified frames.
I had to do it manually: for the admin panel they wrote a script that generates the names of the fragments in the form of “block, block_1, block_2”. Installation was carried out in Premiere Pro: color correction, adding a white background, equalizing sound and removing unnecessary pieces; “Sausage” was cut into necessary fragments. Then selection via hotkey X (so that no extra frame is captured, what happens if you just put io tags) and send to the queue in the Media Encoder . The latter generates the names exactly in the form we wrote above (for this, the script was needed), it was necessary only to name the sequence as the necessary word (for example, block_96, then the files were called block_96, block_96_1, and so on). Sending to the queue, launch, and after 10-20 minutes the party is ready. Filled in the admin panel, fragments rose exactly on those places in the script, the names of which corresponded to the file name.
Before editing, the sound was processed: it is recorded separately, so you can immediately throw it into Adobe Audition . We clean the background noise with Noise Reduction : do Noise Print with silence and apply to the entire file. Then we perform normalization through Normalize and save the file separately (all sources are still stored unchanged). After that, Premiere very quickly synchronizes the audio track using Merge Clips , automatically removes the track of the video file and inserts the edited recording from the voice recorder. The finished file is mounted, and then you already know.
For a long time, all these lessons were mounted by one person, but now, when we have 2 channels on Youtube and at least 3 videos per week, as well as regular filming and recording of podcasts in the studio, the lineup has changed. We hired Maxim, who is now responsible for on-site shooting and editing, and Anton Komarov is responsible for on-site shooting and editing. For the latter, there are Canon 7D Mark II and Sony rx100m3 - they are convenient to carry around , and the picture quality suits us. Plus, we have Oleg Vegan, who makes video footage of popular songs, mounts them independently. Anton works in Final Cut X and Motion (for screensavers), Oleg and Maxim work in After Effects .
A couple of words about podcasts . For them, we have two Octava MK-105 and AT2020 condenser microphones , one ISK BM-800 studio and Behringer UMC202HD external USB audio interface . Behringer replaces a professional sound card, we connect microphones to it.
Sound processing, again, is done in Adobe Audition: equalization, compression, limiter, editing and adding music perebivok there. With the help of the Wordpress plugin we upload to the site, via FeedBurner we set up RSS.
Now the process is debugged: the presenters write the scripts themselves, send them to the editors for review, then remove and edit the schedule in Google Docs. Everyone is responsible for his section, independently uploads (on Youtube or directly to the admin area), then the editor fills in the description, and the SMM manager plans to publish to groups. Due to this, there is time for experiments and non-standard moves, and, of course, now we are trying to make not only useful, but also visually high-quality content.
It may seem that everything is simple, however, we went to this for a long time, and to fill the site with this amount of material (almost 700 grammar lessons, and the video in the “Tichera Method” for 650 hours) we needed more than one and not even two years.
We are not going to stop, new projects for the Puzzle English Live channel and a lot of interesting podcasts are next in line.
For those who want to pump English and check out our content
Readers of the blog give a coupon for 500 rubles for the purchase of a subscription, which includes 8 types of training and weekly mailings about English grammar and vocabulary - "Vitaminki" and "Buns".
And for unlimited and perpetual access to all features of the site there is an “All Inclusive” tariff (the discount does not apply).
Source: https://habr.com/ru/post/348644/
All Articles